ETC 82815EM

R
Intel® 815EM Chipset: 82815EM
Graphics and Memory Controller
Hub (GMCH2-M)
Datasheet
October 2000
Document Reference Number: 290689-001
®
Intel 82815EM GMCH
R
Information in this document is provided in connection with Intel products. No license, express or implied, by estoppel or otherwise, to any intellectual
property rights is granted by this document. Except as provided in Intel’s Terms and Conditions of Sale for such products, Intel assumes no liability
whatsoever, and Intel disclaims any express or implied warranty, relating to sale and/or use of Intel products including liability or warranties relating to fitness
for a particular purpose, merchantability, or infringement of any patent, copyright or other intellectual property right. Intel products are not intended for use in
medical, life saving, or life sustaining applications.
Intel may make changes to specifications and product descriptions at any time, without notice.
Designers must not rely on the absence or characteristics of any features or instructions marked "reserved" or "undefined." Intel reserves these for future
definition and shall have no responsibility whatsoever for conflicts or incompatibilities arising from future changes to them.
The Intel® 815EM chipset may contain design defects or errors known as errata which may cause the product to deviate from published specifications.
Current characterized errata are available on request.
Contact your local Intel sales office or your distributor to obtain the latest specifications and before placing your product order.
2
2
I C is a 2-wire communications bus/protocol developed by Philips. SMBus is a subset of the I C bus/protocol and was developed by Intel. Implementations
of the I2C bus/protocol may require licenses from various entities, including Philips Electronics N.V. and North American Philips Corporation.
Alert on LAN is a result of the Intel-IBM Advanced Manageability Alliance and a trademark of IBM
Copies of documents which have an ordering number and are referenced in this document, or other Intel literature, may be obtained from:
Intel Corporation
www.intel.com
or call 1-800-548-4725
*Third-party brands and names are the property of their respective owners.
Copyright © Intel Corporation 2000
2
Datasheet
®
Intel 82815EM GMCH
R
Contents
1.
Overview .................................................................................................................................... 12
1.1.
1.2.
1.3.
1.4.
1.5.
1.6.
1.7.
1.8.
1.9.
1.10.
1.11.
2.
Signal Description ...................................................................................................................... 19
2.1.
2.2.
2.3.
2.4.
2.5.
2.6.
2.7.
2.8.
2.9.
2.10.
2.11.
2.12.
3.
Host Interface Signals ................................................................................................... 20
System Memory Interface Signals................................................................................. 22
AGP Interface Signals ................................................................................................... 22
2.3.1.
AGP Addressing Signals ............................................................................. 23
2.3.2.
AGP Flow Control Signals ........................................................................... 24
2.3.3.
AGP Status Signals..................................................................................... 24
2.3.4.
AGP Clocking Signals - Strobes.................................................................. 25
2.3.5.
AGP FRAME# Signals ................................................................................ 26
2.3.6.
AGP C3 support Signals ............................................................................. 28
Display Cache Interface Signals.................................................................................... 29
Hub Interface Signals .................................................................................................... 30
Display Interface Signals ............................................................................................... 30
Digital Video Output Signals/TV-Out Pins ..................................................................... 31
Power Signals................................................................................................................ 32
Clock Signals................................................................................................................. 32
Miscellaneous Interface Signals .................................................................................... 33
GMCH2-M Power-Up/Reset Strap Options................................................................... 33
Multiplexed Display Cache and AGP Signal Mapping ................................................... 34
PCI Configuration Registers....................................................................................................... 35
3.1.
3.2.
3.3.
3.4.
3.5.
Datasheet
®
The Intel 815EM Chipset System ................................................................................ 13
®
Intel 815EM Chipset GMCH2-M Overview .................................................................. 14
Host Interface ................................................................................................................ 15
System Memory Interface ............................................................................................. 16
Multiplexed AGP and Display Cache Interface.............................................................. 16
AGP Interface................................................................................................................ 16
1.6.1.
Display Cache Interface .............................................................................. 17
Hub Interface................................................................................................................. 17
GMCH2-M Integrated Graphics (GFX) Support ............................................................ 17
1.8.1.
Intelâ Dynamic Video Memory Technology (D.V.M.T.) ............................... 18
1.8.2.
Display......................................................................................................... 18
1.8.3.
Digital Video Out Port (DVO)....................................................................... 18
System Clocking............................................................................................................ 18
GMCH2-M Power Delivery ............................................................................................ 18
References .................................................................................................................... 18
Register Nomenclature and Access Attributes.............................................................. 35
GMCH2-M Register Introduction ................................................................................... 36
I/O Mapped Registers ................................................................................................... 36
3.3.1.
CONFIG_ADDRESSConfiguration Address Register ............................. 37
3.3.2.
CONFIG_DATAConfiguration Data Register........................................... 38
PCI Bus Configuration Mechanism ............................................................................... 39
PCI Configuration Space Access .................................................................................. 39
3
®
Intel 82815EM GMCH
R
3.6.
3.7.
4
3.5.1.
Logical PCI Bus #0 Configuration Mechanism ............................................40
3.5.2.
Primary PCI (PCI0) and Downstream Configuration Mechanism ................40
3.5.3.
Internal Graphics Device (GFX) Configuration Mechanism.........................40
Host-Hub Interface Bridge/DRAM Controller Device Registers (Device #0) .................40
3.6.1.
VID—Vendor Identification Register (Device 0)...........................................43
3.6.2.
DID—Device Identification Register (Device 0) ...........................................43
3.6.3.
PCICMD—PCI Command Register (Device 0) ...........................................44
3.6.4.
PCISTS—PCI Status Register (Device 0) ...................................................45
3.6.5.
RID—Revision Identification Register (Device 0) ........................................46
3.6.6.
SUBC—Sub-Class Code Register (Device 0) .............................................46
3.6.7.
BCC—Base Class Code Register (Device 0) ..............................................46
3.6.8.
MLT—Master Latency Timer Register (Device 0) .......................................47
3.6.9.
HDR—Header Type Register (Device 0) .....................................................47
3.6.10.
APBASE—Aperture Base Configuration Register (Device 0 - AGP MODE
ONLY)
.....................................................................................................................48
3.6.11.
SVID—Subsystem Vendor Identification Register (Device 0) .....................49
3.6.12.
SID—Subsystem Identification Register (Device 0) ....................................49
3.6.13.
CAPPTR—Capabilities Pointer (Device 0) ..................................................50
3.6.14.
GMCHCFG—GMCH2-M Configuration Register (Device 0) .......................50
3.6.15.
APCONT—Aperture Control (Device 0) ......................................................52
3.6.16.
DRP—DRAM Row Population Register (Device 0) .....................................53
3.6.17.
DRAMT—DRAM Timing Register (Device 0) ..............................................54
3.6.18.
DRP2—DRAM Row Population Register 2 (Device 0) ................................55
3.6.19.
FDHC  Fixed DRAM Hole Control Register (Device 0) ............................56
3.6.20.
PAMProgrammable Attributes Map Registers (Device 0) ........................56
3.6.21.
C3STATUS —C3 Control and Status Register (Device #0) ........................59
3.6.22.
SMRAM - System Management RAM Control Register (Device 0).............60
3.6.23.
MISCC—Miscellaneous Control Register (Device 0) ..................................62
3.6.24.
CAPID—Capability Identification (Device 0 - AGP MODE ONLY)...............63
3.6.25.
BUFF_SC—System Memory Buffer Strength Control Register (Device 0) .65
3.6.26.
BUFF_SC2-System Memory Buffer Strength Control Register 2 (Device 0)68
3.6.27.
ACAPID—AGP Capability Identifier Register (Device 0) .............................69
3.6.28.
AGPSTAT—AGP Status Register (Device 0)..............................................70
3.6.29.
AGPCMD—AGP Command Register (Device 0) ........................................71
3.6.30.
AGPCTRL—AGP Control Register (Device 0) ............................................72
3.6.31.
APSIZE—Aperture Size (Device 0) .............................................................73
3.6.32.
ATTBASE-Aperture Translation Table Base Register (Device 0) ...............74
3.6.33.
AMTT—AGP Multi-Transaction Timer (Device 0) .......................................75
3.6.34.
LPTT—AGP Low Priority Transaction Timer Register (Device 0) ...............76
3.6.35.
GMCHCFG—GMCH2-M Configuration Register (Device 0) .......................77
3.6.36.
ERRCMD—Error Command Register (Device 0)........................................78
AGP/PCI Bridge Registers – (Device #1 - Visible in AGP Mode Only) ..........................79
3.7.1.
VID1—Vendor Identification Register (Device 1).........................................80
3.7.2.
DID1—Device Identification Register (Device 1) .........................................80
3.7.3.
PCICMD1—PCI-PCI Command Register (Device 1) ..................................81
3.7.4.
PCISTS1—PCI-PCI Status Register (Device 1) ..........................................82
3.7.5.
RID1—Revision Identification Register (Device 1) ......................................83
3.7.6.
SUBC1—Sub-Class Code Register (Device 1) ...........................................83
3.7.7.
BCC1—Base Class Code Register (Device 1) ............................................84
3.7.8.
MLT1—Master Latency Timer Register (Device 1) .....................................84
3.7.9.
HDR1—Header Type Register (Device 1) ...................................................84
3.7.10.
PBUSN—Primary Bus Number Register (Device 1)....................................85
3.7.11.
SBUSN—Secondary Bus Number Register (Device 1)...............................85
3.7.12.
SUBUSN—Subordinate Bus Number Register (Device 1) ..........................85
Datasheet
®
Intel 82815EM GMCH
R
3.8.
4.
Functional Description.............................................................................................................. 111
4.1.
4.2.
4.3.
Datasheet
3.7.13.
SMLT—Secondary Master Latency Timer Register (Device 1) .................. 86
3.7.14.
IOBASE—I/O Base Address Register (Device 1) ....................................... 87
3.7.15.
IOLIMIT—I/O Limit Address Register (Device 1) ........................................ 88
3.7.16.
SSTS—Secondary PCI-PCI Status Register (Device 1) ............................ 89
3.7.17.
MBASE—Memory Base Address Register (Device 1) ................................ 90
3.7.18.
MLIMIT—Memory Limit Address Register (Device 1)................................. 91
3.7.19.
PMBASE—Prefetchable Memory Base Address Register (Device 1) ........ 92
3.7.20.
PMLIMIT—Prefetchable Memory Limit Address Register (Device 1) ......... 93
3.7.21.
BCTRL—PCI-PCI Bridge Control Register (Device 1) ................................ 94
3.7.22.
ERRCMD1—Error Command Register (Device 1) ..................................... 96
Graphics Device Registers (Device 2 - VISIBLE IN GFX MODE ONLY) ...................... 96
3.8.1.
VID2—Vendor Identification Register (Device 2) ........................................ 97
3.8.2.
DID2—Device Identification Register (Device 2)......................................... 98
3.8.3.
PCICMD2—PCI Command Register (Device 2) ......................................... 98
3.8.4.
PCISTS2—PCI Status Register (Device 2)............................................... 100
3.8.5.
RID2—Revision Identification Register (Device 2).................................... 101
3.8.6.
PI—Programming Interface Register (Device 2)....................................... 101
3.8.7.
SUBC2—Sub-Class Code Register (Device 2)......................................... 101
3.8.8.
BCC2—Base Class Code Register (Device 2) ......................................... 102
3.8.9.
CLS—Cache Line Size Register (Device 2).............................................. 102
3.8.10.
MLT2—Master Latency Timer Register (Device 2)................................... 102
3.8.11.
HDR2—Header Type Register (Device 2) ................................................ 103
3.8.12.
BIST—BIST Register (Device 2) ............................................................... 103
3.8.13.
GMADR-Graphics Memory Range Address Register (Device 2)............. 104
3.8.14.
MMADR—Memory Mapped Range Address Register (Device 2) ............ 105
3.8.15.
SVID—Subsystem Vendor Identification Register (Device 2) .................. 105
3.8.16.
SID—Subsystem Identification Register (Device 2) .................................. 106
3.8.17.
ROMADR - Video BIOS ROM Base Address Registers (Device 2).......... 106
3.8.18.
CAPPOINT—Capabilities Pointer Register (Device 2).............................. 106
3.8.19.
INTRLINE—Interrupt Line Register (Device 2) ......................................... 107
3.8.20.
INTRPIN—Interrupt Pin Register (Device 2) ............................................. 107
3.8.21.
MINGNT—Minimum Grant Register (Device 2) ........................................ 107
3.8.22.
MAXLAT—Maximum Latency Register (Device 2) ................................... 107
3.8.23.
PM_CAPID—Power Management Capabilities ID Register (Device 2) .... 108
3.8.24.
PM_CAP—Power Management Capabilities Register (Device 2) ............ 108
3.8.25.
PM_CS - Power Management Control/Status Register (Device 2) ........... 109
System Memory and I/O Address Map ....................................................................... 111
4.1.1.
Memory Address Space ............................................................................ 111
DOS Compatibility Memory Space .............................................................................. 114
4.2.1.1.
DOS Area (00000h-9FFFh) .......................................................... 115
4.2.1.2.
Video Buffer Area (A0000h-BFFFFh) ........................................... 115
4.2.1.3.
Monochrome Adapter (MDA) Range (B0000h - B7FFFh) ............ 115
4.2.1.4.
Expansion Area (C0000h-DFFFFh).............................................. 116
4.2.1.5.
Extended System BIOS Area (E0000h-EFFFFh) ......................... 116
4.2.1.6.
System BIOS Area (F0000h-FFFFFh).......................................... 116
Extended Memory Area............................................................................................... 116
4.3.1.
Main DRAM Address Range (0010_0000h to TOM)................................. 116
4.3.1.1.
15MB-16MB Hole Area................................................................. 116
4.3.1.2.
Extended SMRAM Address Range .............................................. 116
4.3.1.3.
HSEG (High Segment) ................................................................. 117
4.3.1.3.1.
TSEG (Top of Memory Segment)............................... 117
4.3.2.
PCI Memory Address Range (Top of Main Memory to 4 GB)................... 117
5
®
Intel 82815EM GMCH
R
4.3.2.1.
4.4.
4.5.
4.6.
4.7.
4.8.
4.9.
4.10.
4.11.
4.12.
6
APIC Configuration Space (FEC0_0000h -FECF_FFFFh,
FEE0_0000h- FEEF_FFFFh)........................................................118
4.3.2.2.
High BIOS Area (FFE0_0000h -FFFF_FFFFh) ............................118
System Management Mode (SMM) Memory Range....................................................118
4.4.1.
SMM Space Definition ...............................................................................119
4.4.2.
SMM Space Restrictions ...........................................................................119
4.4.3.
SMM Space Combinations ........................................................................120
4.4.4.
Initialization and Usage of SMRAM and Graphics Local Memory .............120
Memory Shadowing .....................................................................................................120
I/O Address Space.......................................................................................................120
4.6.1.
AGP/PCI - I/O Address Mapping ...............................................................121
GMCH2-M Address Decode Rules and Cross-Bridge Address Mapping....................121
4.7.1.
Address Decode Rules ..............................................................................121
4.7.2.
The Hub Interface Accesses to GMCH2-M that Cross Device Boundaries122
4.7.3.
AGP Interface Decode Rules.....................................................................122
4.7.3.1.
Cycles Initiated Using PCI Protocol...............................................122
4.7.3.2.
Cycles Initiated Using AGP Protocol .............................................123
4.7.3.3.
AGP Accesses to GMCH2-M that Cross Device Boundaries .......123
4.7.4.
Legacy VGA Ranges .................................................................................124
Host Interface...............................................................................................................125
4.8.1.
Host Bus Device Support...........................................................................125
4.8.2.
Special Cycles ...........................................................................................127
System Memory DRAM Interface ................................................................................128
4.9.1.
DRAM Organization and Configuration......................................................128
4.9.1.1.
Configuration Mechanism SO-DIMMs ..........................................129
4.9.1.2.
DRAM Register Programming ......................................................129
4.9.2.
DRAM Address Translation and Decoding ................................................130
4.9.3.
SDRAMT Register Programming ..............................................................131
4.9.4.
SDRAM Paging Policy ...............................................................................131
Intelâ Dynamic Video Memory Technology (D.V.M.T.) ................................................131
Display Cache Interface...............................................................................................132
4.11.1.
Supported DRAM Types for Display Cache Memory.................................132
4.11.2.
Memory Configurations..............................................................................133
4.11.3.
Address Translation...................................................................................133
4.11.4.
Display Cache Interface Timing.................................................................134
Internal Graphics Device..............................................................................................134
4.12.1.
3D/2D Instruction Processing ....................................................................134
4.12.2.
3D Engine ..................................................................................................135
4.12.3.
Buffers .......................................................................................................135
4.12.4.
Setup..........................................................................................................136
4.12.5.
Texturing....................................................................................................136
4.12.6.
2D Operation..............................................................................................138
4.12.7.
Fixed Blitter (BLT) and Stretch Blitter (STRBLT) Engines .........................139
4.12.7.1. Fixed BLT Engine..........................................................................139
4.12.7.2. Arithmetic Stretch BLT Engine ......................................................139
4.12.8.
Hardware Motion Compensation ...............................................................140
4.12.9.
Hardware Cursor and Popup Support .......................................................140
4.12.10. Overlay Engine ..........................................................................................140
4.12.11. Display .......................................................................................................141
4.12.12. Digital Video Out (DVO) Port .....................................................................142
4.12.12.1. VCH interface................................................................................142
4.12.12.2. DVO Port Data Format..................................................................143
2
4.12.12.3. DVO Port I C Functionality............................................................145
Datasheet
®
Intel 82815EM GMCH
R
4.13.
4.14.
4.15.
4.16.
4.17.
4.18.
5.
Pinout and Package Information.............................................................................................. 153
5.1.
5.2.
6.
GMCH2-M Pinout ........................................................................................................ 153
GMCH2-M Package Dimensions ................................................................................ 160
Testability ................................................................................................................................. 162
6.1.
6.2.
Datasheet
4.12.13. DDC (Display Data Channel) .................................................................... 145
System Reset for the GMCH2-M................................................................................. 145
System Clock Description ........................................................................................... 146
4.14.1.
External Clock Sources ............................................................................. 146
4.14.2.
Internal Clock Sources .............................................................................. 146
Power Management .................................................................................................... 146
4.15.1.
Specifications Supported........................................................................... 147
General Description of ACPI Power States ................................................................. 147
Power State Transition Rules at Platform Level.......................................................... 148
ACPI Support............................................................................................................... 149
4.18.1.
Full on (C0 State) ...................................................................................... 149
4.18.2.
Stop Grant or Quick Start (C2 State) ........................................................ 149
4.18.3.
Stop Clock (C3 State)................................................................................ 149
4.18.4.
C3 Support AGP Port Signal ..................................................................... 150
4.18.5.
Power-on-suspend (POS) (S1 State) ........................................................ 150
4.18.6.
Suspend-to-RAM (STR) (S3 State) ........................................................... 151
4.18.7.
Suspend to DISK (STD) S4 State ............................................................. 151
4.18.8.
Graphics Controller Requirements............................................................ 151
4.18.8.1. The D0 State ................................................................................ 151
4.18.8.2. The D3 State ................................................................................ 151
XOR Chain .................................................................................................................. 163
All Z ............................................................................................................................. 163
7
®
Intel 82815EM GMCH
R
Figures
®
Figure 1. Intel 815EM Chipset System Block Diagram ...........................................................14
®
Figure 2. Intel 815EM Chipset GMCH2-M Block Diagram......................................................15
Figure 3. PAM Registers .........................................................................................................58
Figure 4. System Memory Address Map ................................................................................112
Figure 5. Detailed Memory System Address Map ..................................................................113
Figure 6. MCH2-M Display Cache Interface to 4MB ..............................................................133
Figure 7. 3D/2D Pipeline Preprocessor ..................................................................................135
Figure 8. Data Flow for the 3D Pipeline..................................................................................136
Figure 9. Digital Video Out Port Mobile Application Block Diagram With VCH ......................143
Figure 10. Digital Video Out Port Block Diagram Without VCH .............................................143
Figure 11. GMCH2-M Pinout (Top View-Left Side) ................................................................154
Figure 12. GMCH2-M Pinout (Top View-Right Side) ..............................................................155
Figure 13. GMCH2-M GMCH BGA Package Dimensions (Top and Side Views) ..................160
®
Figure 14. Intel 815EM Chipset GMCH2-M BGA Package Dimensions (Bottom View) .......161
Tables
Table 1. AGP Data Rate and Signaling Levels.........................................................................16
Table 2. Voltage Levels for Each Interface...............................................................................20
Table 3. Display Cache and AGP signal Mapping....................................................................34
Table 4. GMCH2-M PCI Configuration Space (Device #0) ......................................................41
Table 5. Supported System Memory DIMM Configurations .....................................................53
Table 6. Attribute Bit Assignments ...........................................................................................57
Table 7. PAM Registers and Associated Memory Segments...................................................58
Table 8. GMCH2-M Configuration Space (Device #1) .............................................................79
Table 9. Device 2 Configuration Space Address Map (Internal Graphics) ...............................96
Table 10. Memory Segments and Their Attributes .................................................................114
Table 11. SMM Space Abbreviations .....................................................................................119
Table 12. Summay of Transactions Supported By GMCH2-M...............................................125
Table 13. Host Responses Supported by the GMCH2-M.......................................................126
Table 14. Special Cycles ........................................................................................................127
Table 15. Data Bytes on DIMM Used for Programming DRAM Registers .............................129
Table 16. GMCH2-M DRAM Address Mux Function ..............................................................130
Table 17. Programmable SDRAM Timing Parameters ..........................................................131
Table 18. Memory Size for each configuration: ......................................................................133
Table 19, GMCH2-M Local Memory Address Mapping..........................................................134
Table 20. Partial List of Display Modes Supported.................................................................141
Table 21. Partial List of Flat Panel Modes Supported ............................................................144
Table 22. Partial List of TV-Out Modes Supported.................................................................144
Table 23. Supported Frequencies and Corresponding Phase Alignments ............................146
Table 24. General Description of ACPI Power States ............................................................147
Table 25. State Transition Rules at Platform Level ................................................................148
®
®
Table 26. Ballout differences between Intel 815 Chipset GMCH and Intel 815EM Chipset
GMCH2-M .......................................................................................................................153
Table 27. Alphabetical Pin Assignment (by Signal Name) .....................................................156
8
Datasheet
®
Intel 82815EM GMCH
R
Revision History
Rev.
-001
Datasheet
Description
• Initial Release
Date
January 2000
9
®
Intel 82815EM GMCH
R
Intel® 82815EM GMCH2-M Features
§ Processor/Host Bus Support
 Optimized for the mobile Intel Pentium III processor
and mobile Intel® Celeron™ processors.
 Supports 32-Bit System Bus Addressing
 4 deep in-order queue; 4 or 1 deep request queue
 Supports Uni-processor systems only
 In-order and Dynamic Deferred Transaction Support
 100MHz System Bus Frequency
 GTL+ I/O Buffer
§ Integrated SDRAM Controller










32 to 512MB using 16/64/128/256 Mbit technology
Supports up to 3 double sided SO-DIMMs @ 100Mhz
64-bit data interface
100MHz system memory bus frequency
Support for Asymmetrical SDRAM addressing only
Support for x8 and x16 SDRAM device width
Unbuffered, Non-ECC SDRAM only supported
Refresh Mechanism: CBR supported
Enhanced Open page Arbitration SDRAM paging scheme
Suspend to RAM support
§ Accelerated Graphics Port (AGP) Interface
Multiplexed with Internal Graphics
 Supports a single AGP device
 Supports AGP 2.0 including 4x AGP data transfers
 AGP 2.0 support via dual mode buffers to allow 3.3v or
1.5v signaling
 AGP PIPE# or SBA initiated accesses to SDRAM are not
snooped
 AGP FRAME# initiated accesses to SDRAM are snooped
 High priority access support
 Hierarchical PCI configuration mechanism
 Delayed transaction support for AGP-to-SDRAM reads
that can not be serviced immediately
§ Arbitration Scheme and Concurrency
 Intelligent Centralized Arbitration Model for Optimum
Concurrency Support
 Concurrent operations of processor and System busses
supported via dedicated arbitration and data buffering
§ Data Buffering
 Distributed Data Buffering Model for optimum
concurrency
 SDRAM Write Buffer with read-around-write capability
 Dedicated processor-SDRAM, hub interface-SDRAM and
Graphics-SDRAM Read Buffers
§ Power Management Functions
 SMRAM space remapping to A0000h (128 KB)
 Optional Extended SMRAM space above 256 MB,
additional 512K/1MB TSEG from Top of Memory,
cacheable
 Stop Clock Grant and Halt special cycle translation from
the host to the hub interface
 ACPI Compliant power management
 Dynamic (independent) SCKE support
 AGPBUSY# support
§ Integrated Graphics Controller Multiplexed with
AGP Controller





3D Hyper Pipelined Architecture
Parallel Data Processing (PDP)
Precise Pixel Interpolation (PPI)
Full 2D H/W Acceleration
Motion Video Acceleration







Flat & Gouraud Shading
Mip Maps with Trilinear and Anisotropic Filtering
Full Color Specular
Fogging Atmospheric Effects
Z Buffering
3D Pipe 2D Clipping
Backface Culling
§ 3D Graphics Visual Enhancements
§ 3D Graphics Texturing Enhancements
 Per Pixel Perspective Correction Texture Mapping
 Texture Compositing
 Texture Color Keying/Chroma Keying
§ Digital Video Output Port (DVO)
 Support for an external TV encoder
 Video Controller Hub (VCH) Interface
 196 Pin mini-BGA package
§ Display
 Integrated 24-bit 230MHz RAMDAC
 Gamma Corrected Video
 DDC2B Compliant
§ 2D Graphics




Up to 1600x1200 in 8-bit Color at 75 Hz Refresh
Hardware Accelerated Functions
3 Operand Raster BitBLTs
64x64x3 Color Transparent Cursor
 Hardware Cursor and Popup window support
§ Arithmetic Stretch Blitter Video




H/W Assisted Motion Compensation for MPEG2 Decode
H/W assisted DVD playback at 30fps
H/W Overlay Engine with Bilinear Filtering
Independent gamma correction, saturation, brightness &
contrast for overlay
Graphics Memory Controller
§ Integrated
â
 Intel D.V.M. Technology
§ Display Cache Interface




32-bit data interface
100/133MHz SDRAM interface
Support for 2 1Mx16, or 1 2Mx32 SDRAM
4MB max addressable
§ Supporting I/O Bridge
 360 Pin eBGA I/O Controller Hub (ICH2-M)
§ Packaging/Power
 544 BGA
 1.8V core with 3.3V CMOS I/O
 APIC Buffer Management
10
Datasheet
®
Intel 82815EM GMCH
R
Intel® 815EM Chipset GMCH2-M Simplified Block Diagram
HA[31:3]#
HD[63:0]#
ADS#
BNR#
BPRI#
DBSY#
DEFER#
DRDY#
HIT#
HITM#
HLOCK#
HREQ[4:0]#
HTRDY#
RS[2:0]#
CPURST#
SMAA[12:0]
SMAB[7:4]#
SMAC[7:4]#
SBS[1:0]
SMD[63:0]
SDQM[7:0]
SCSA[5:0]#
SCSB[5:0]#
SRAS#
SCAS#
SWE#
SCKE[5:0]
HCLK
SCLK
LTCLK[1:0]
LOCLK
LRCLK
DCLKREF
HLCLK
GTLREF0
GTLREF1
RESET#
HUBREF
HL[10:0]
HLSTRB
HLSTRB#
HLZCOMP
HCOMP
Datasheet
Analog
Display
Interface
System Bus
Interface
Digital Video
Output Port
System
Memory
Interface
*(Display
Cache
Interface)
Clock
Signals
Misc.
Interface
Signals
Hub
Interface
AGP
Interface
VSYNC
HSYNC
IREF
RED
GREEN
BLUE
DDCK
DDDA
LTVCL
LTVCK
LTVDA
LTVCLKIN/STAL
LTVCLKOUT[1:0
LTVBLANK#
LTVDATA[11:0]
LTVSYNC
LTVHSYNC
INTRPT#
LCS#
LDQM[3:0]
LRAS#
LCAS#
LMA[11:0]
LWE#
LMD[31:0]
PIPE#
SBA[7:0]
RBF#
WBF#
ST[2:0]
ADSTB[1:0]
ADSTB[1:0]#
SBSTB
SBSTB#
GFRAME#
GIRDY#
GTRDY#
GSTOP#
GDEVSEL#
GREQ#
GGNT#
GAD[31:0]
GC/BE[3:0]
GPAR
GRCOMP
AGPREF
AGPBUSY#
11
®
Intel 82815EM GMCH
R
This page left intentionally blank.
12
Datasheet
®
Intel 82815EM GMCH
R
1.
Overview
The Intel® 815EM chipset is a high-flexibility chipset designed to extend from the basic graphics /
multimedia mobile PC platform up to the mainstream performance mobile PC platform. The Intel®
815EM chipset consists of a Graphics and Memory Controller Hub (GMCH2-M) Bridge and an I/O
Controller Hub (ICH2-M) Bridge for the I/O subsystem.
The GMCH2-M integrates a Display Cache SDRAM controller that supports a 32-bit 100MHz SDRAM
array for enhanced integrated 3D graphics performance. Multiplexed with the display cache interface is
an AGP controller interface to enable graphics configuration and upgrade flexibility with the Intel®
815EM chipset. The AGP interface and the internal graphics device are mutually exclusive. When the
AGP port is populated with an AGP graphics device, the integrated graphics is disabled and thus the
display cache interface is not needed.
In this document the terms “GMCH2-M” and “Intel® 815EM chipset” refer to the Intel® 815EM chipset
Graphics and Memory Controller Hub interchangeably, unless otherwise specified. Also the term GFX
and internal graphics device refers to the Intel® 815EM chipset internal graphics device, unless otherwise
specified.
The Intel® 815EM chipset may contain design defects or errors known as errata that may cause the
product to deviate from published specifications. Current characterized errata are available on request.
1.1.
The Intel® 815EM Chipset System
The Intel® 815EM chipset uses a hub architecture with the GMCH2-M as the host bridge hub and the I/O
Controller Hub (ICH2-M) as the I/O hub. The ICH2-M is a highly integrated multifunctional I/O
Controller Hub that provides the interface to the PCI Bus and integrates many of the functions needed in
today’s PC platforms. The Intel® 815EM chipset GMCH2-M and ICH2-M communicate over a dedicated
hub interface.
ICH2-M functions and capabilities include:
• PCI Rev 2.2 compliant with support for 33 MHz PCI operations
• ICH2-M supports up to 6 Req/Gnt pairs (PCI bus)
• Power Management Logic Support
• Enhanced DMA Controller, Interrupt Controller & Timer Functions
• Integrated IDE controller; ICH2-M supports Ultra ATA/66/100
• USB host interface with support for four USB ports
• System Management Bus (SMBus) compatible with most I2C devices
• AC’97 2.1 Compliant Link for Audio and Telephony CODECs
• Low Pin Count (LPC) interface
• Firmware Hub (FWH) interface support
• Alert On LAN*
Datasheet
13
®
Intel 82815EM GMCH
R
The following figure shows a block diagram of the typical mobile platform based on the Intel® 815EM
chipset. The GMCH2-M supports processor bus frequencies of 100MHz. The ICH2-M provides extensive
I/O support. The ICH2-M provides support for 6 PCI bus Req/Gnt pairs, increased IDE capability from
Ultra ATA/33 to Ultra ATA/100, and Alert On LAN*.
®
Figure 1. Intel 815EM Chipset System Block Diagram
Processor
Flat Panel
TV Encoder
100 MHz PSB
VCH
DVO
CRT
AGP4X
External
GFX
Intel® 815EM
Chipset
GMCH2-M
544 BGA
HUB I/F
Display Cache
4 x USB
2 x IDE
ICH2-m
ICH2-M
LAN
SMBus
Audio Codec
Modem Codec
PCI Bus
360
360 BGA
BGA
AC97
LPC
FWH
Flash BIOS
1.2.
100 MHz
SDRAM
LPC
Super I/O
8214EM Sys Blk
Intel® 815EM Chipset GMCH2-M Overview
Figure 1 is a block diagram of the Intel® 815EM chipset. The Intel® 815EM chipset GMCH2-M
functions and capabilities include:
Support for a single processor configuration
• 64-bit GTL+ based Processor Side Bus Interface at 100 MHz
• 32-bit Host Address Support
• 64-bit System Memory Interface with optimized support for SDRAM at 100 MHz
• Integrated 2D & 3D Graphics Engines
• Integrated H/W Assisted Motion Compensation Engine
• Integrated 230 MHz DAC
• Digital Video Out (DVO) Interface Port to communicate to VCH for mobile LVDS flat panel
interface support and TV encoder support
14
Datasheet
®
Intel 82815EM GMCH
R
• Communicates to ICH2-M via Hub interface
• Local memory Display Cache at 100/133 MHz
• 2X/4X AGP Controller Port
• ACPI 1.0 power management
®
Figure 2. Intel 815EM Chipset GMCH2-M Block Diagram
Processor Side Bus (100 Mhz)
Processor
Interface
System
Memory
Interface
SDRAM
100 MHz
64 bit
Primary Display
Display
Cache or
AGP 2X/4X
card
AGP
Interface
Local
Memory
Interface
Data
Stream
Control &
Dispatch
Overlay
Popup + Cursor
RAMDAC
Monitor
DVO port
Digital
Video Out
Port
3D Pipeline
2D (Blit Engine)
Internal
Graphics
C3/C2/S1/S3
Support
Hub Interface
Intel® 815EM Chipset
815EM CS Blk Diag
1.3.
Host Interface
The host interface of the Intel® 815EM chipset GMCH2-M is optimized to support the mobile Intel®
Pentium® III processor, and the mobile Intel® Celeronä processor in uBGA and uPGA packages. The
GMCH2-M implements the host address, control, and data bus interfaces within a single device. The
GMCH2-M supports a 4-deep in-order queue (i.e., supports pipelining of up to four outstanding
transaction requests on the host bus). Host bus addresses are decoded by the GMCH2-M for access to
system memory, PCI memory, I/O (via hub interface), PCI configuration space and Graphics memory.
The GMCH2-M takes advantage of the pipelined addressing capability of the processor to improve the
overall system performance.
Datasheet
15
®
Intel 82815EM GMCH
R
1.4.
System Memory Interface
The Intel® 815EM chipset GMCH2-M integrates a system memory controller that supports a 64-bit, 100MHz SDRAM array. The only DRAM type supported is industry standard Synchronous DRAM
(SDRAM). The SDRAM controller interface is fully configurable through a set of control registers.
Complete descriptions of these registers will be available in a future revision of this document.
The GMCH2-M supports industry standard 64-bit wide DIMMs for desktop platforms and SO-DIMMs
for mobile platforms with SDRAM devices. The twelve multiplexed address lines, SMAA[12:0], along
with the two bank select lines, SBS[1:0], allow the GMCH2-M to support 2M, 4M, 8M, and 16M x64
DIMMs. Only asymmetric addressing is supported. The GMCH2-M has 12 SCS# lines enabling the
support of up to six 64-bit rows of SDRAM. The GMCH2-M targets SDRAM with CL2 and CL3 and
supports both single and double-sided DIMMs for desktop and SO-DIMMs for mobile platforms.
Additionally, the GMCH2-M also provides a seven deep refresh queue. The GMCH2-M can be
configured to keep multiple pages open within the memory array. Pages can be kept open in any one
bank of memory.
SCKE[5:0] is used in configurations requiring powerdown mode for the SDRAM. Each SCKE can be
dynamically powerdown if not in use. This scheme will save significant amount of power since only one
SO-DIMM at any given time will be functional and all the rest will be powered down.
1.5.
Multiplexed AGP and Display Cache Interface
The Intel® 815EM chipset GMCH2-M multiplexes a display cache interface for internal graphics 3D
performance improvement with an AGP controller interface. The Display Cache is used only in the
internal graphics. When an AGP card is populated in the system, the Intel® 815EM chipset GMCH2-M
internal graphics will be disabled and the AGP controller enabled.
1.6.
AGP Interface
A single AGP port is supported by the GMCH2-M AGP interface. The AGP buffers operate in one of
two selectable modes:
• 3.3V drive, not 5 volt safe - This mode is compliant to the AGP 1.0 and 2.0 specs.
• 1.5V drive, not 3.3 volt safe - This mode is compliant with the AGP 2.0 spec.
The following table shows the AGP Data Rate and the Signaling Levels supported by the GMCH2-M:
Table 1. AGP Data Rate and Signaling Levels
Signaling Level
16
Data Rate
1.5v
3.3v
1x AGP
Yes
Yes
2x AGP
Yes
Yes
4x AGP
Yes
No
Datasheet
®
Intel 82815EM GMCH
R
The AGP interface supports 4x AGP signaling. AGP semantic (PIPE# or SBA[7:0]) cycles to SDRAM
are not snooped on the host bus. AGP FRAME# cycles to SDRAM are snooped on the host bus. The
GMCH2-M supports PIPE# or SBA[7:0] AGP address mechanisms, but not both simultaneously. Either
the PIPE# or the SBA[7:0] mechanism must be selected during system initialization. High priority
accesses are supported. Only memory writes from the hub interface to AGP are allowed. No transactions
from AGP to the hub interface are allowed.
1.6.1.
Display Cache Interface
The Intel® 815EM chipset integrates a Display Cache SDRAM controller for enhanced 3D performance.
The maximum memory support is 4 MB. The Display Cache Interface is multiplexed with the AGP
controller interface to provide options for both the value mobile segment and main stream performance
platforms configuration with the Intel® 815EM chipset. When the AGP port is populated with an AGP
graphics device the integrated graphics is disabled and thus the display cache interface is not needed.
The Intel® 815EM chipset GMCH2-M supports a Display Cache SDRAM controller with a 32-bit
133MHz SDRAM array. The DRAM type supported is industry standard Synchronous DRAM (SDRAM)
like that of the system memory. The local memory SDRAM controller interface is fully configurable
through a set of control registers. For more details on the registers, consult the Extensions to the
Pentium® Pro Processor BIOS Writer’s Guide Revision 3.6 and higher.
1.7.
Hub Interface
The hub interface is a private interconnect between the GMCH2-M and the ICH2-M.
1.8.
GMCH2-M Integrated Graphics (GFX) Support
The GMCH2-M includes a highly integrated 2D/3D graphics accelerator (GFX) and is PC99a compliant.
Its architecture consists of dedicated multi-media engines executing in parallel to deliver high
performance graphics and video capabilities which includes integrated 3D graphics engine, 2D graphics
engine, video engine, display pipeline, and motion compensation HW accelerator.
The 3D and 2D engines are managed by a 3D/2D pipeline preprocessor allowing a sustained flow of
graphics data to be rendered and displayed. The deeply pipelined 3D accelerator engine provides 3D
graphics quality and performance via per-pixel 3D rendering and parallel data paths which allow each
pipeline stage to simultaneously operate on different primitives or portions of the same primitive. The
GMCH2-M graphics accelerator engine supports perspective-correct texture mapping, trilinear and
anisotropic filtering, Mip-Mapping, Gouraud shading, alpha-blending, fogging and Z-buffering. A rich
set of 3D instructions permit these features to be independently enabled or disabled.
The GMCH2-M integrated graphics accelerator’s 2D capabilities include BLT and arithmetic STRBLT
engines, a hardware cursor, a popup window and an extensive set of 2D registers and instructions. The
high performance 64-bit BitBLT engine provides hardware acceleration for many common Windows
operations.
In addition to its 2D/3D capabilities, the GMCH2-M integrated graphics accelerator also supports full
MPEG-2 motion compensation for software-assisted DVD video playback, a VESA DDC2B compliant
display interface and a digital video out port to interface to the VCH for panel support.
Datasheet
17
®
Intel 82815EM GMCH
R
1.8.1.
Intelâ Dynamic Video Memory Technology (D.V.M.T.)
The internal graphics device on the GMCH2-M supports Intelâ Dynamic Video Memory Technology.
Intel® D.V.M.T. dynamically responds to application requirements by allocating the proper amount of
display and texturing memory taken from system memory. For the GMCH2-M, a Display Cache (DC)
can be used for Z-buffers (Textures and display buffer are located in system memory). If the display
cache is not used, the Z-buffer is located in system memory.
1.8.2.
Display
The GMCH2-M provides interfaces to a standard progressive scan monitor. This interface is only active
when running in internal graphics mode. The Intel® 815EM chipset GMCH2-M directly drives a standard
progressive scan monitor up to a resolution of 1600x1200 pixels.
1.8.3.
Digital Video Out Port (DVO)
The GMCH2-M provides a Digital Video Out port to connect to the Video Controller Hub (VCH) to
connect to CMOS or LVDS flat panel interfaces. The interface has 1.8V signaling to allow it to operate
at higher frequencies. The VCH also has an interface to connect to a TV encoder. The TV encoder must
support the DVO interface.
1.9.
System Clocking
The GMCH2-M has a new type of clocking architecture. It has integrated SDRAM buffers that runs at
100 MHz, regardless of processor side bus frequency. The GMCH2-M uses a 48-MHz clock as the
reference clock (DCLKREF) input for the graphics pixel clock PLL.
1.10.
GMCH2-M Power Delivery
The GMCH2-M core voltage is 1.8V. System Memory runs off of a 3.3V supply. Display cache
memory runs off of the AGP 3.3V supply. AGP 1X/2X I/O can run off of either a 3.3V or a 1.5V
supply. AGP 4X I/O requires a 1.5V supply. The AGP interface voltage is determined by the VDDQ
generation on the motherboard.
1.11.
References
• GTL+ I/O Specification: Contained in the Pentium® II Processor Databook
• PCI Local bus Specification 2.2: Contact www.pcisig.com
18
Datasheet
®
Intel 82815EM GMCH
R
2.
Signal Description
This section provides a detailed description of the Intel® 815EM chipset GMCH2-M signals. The signals
are arranged in functional groups according to their associated interface. The states of all of the signals
during reset are provided in the System Reset section.
The “#” symbol at the end of a signal name indicates that the active, or asserted state occurs when the
signal is at a low voltage level. When “#” is not present after the signal name the signal is asserted when
at the high voltage level.
The following notations are used to describe the signal type:
I
Input pin
O
Output pin
I/OD
Input / Open Drain Output pin. This pin requires a pullup 3.3V.
I/O
Bi-directional Input/Output pin
s/t/s
Sustained Tristate. This pin is driven to its inactive state prior to tri-stating.
as/t/s
Active Sustained Tristate. This applies to some of the hub interface signals. This
pin is weakly driven to its last driven value.
The signal description also includes the type of buffer used for the particular signal:
Note:
Datasheet
GTL+
Open Drain GTL+ interface signal. Refer to the GTL+ I/O Specification for
complete details. These signals will be actively driven high for a short period of
time to assist timing when the Intel® 815EM chipset is configured for 100-MHz
Host interface. GTL+ signals are inverted bus signals where a low voltage
represents a logical “1”.
AGP
AGP interface signals. These signals can be programmed to be compatible with
AGP 2.0 3.3V or 1.5V Signaling Environment DC and AC Specifications. In
3.3V mode the buffers are not 5V tolerant. In 1.5V mode the buffers are not
3.3V tolerant.
CMOS
The CMOS buffers are Low Voltage TTL compatible signals. These are 3.3V
only.
LVTTL
Low Voltage TTL compatible signals. There are 3.3V only.
1.8V
1.8V signals for the digital video interface
Analog
Analog CRT Signals
That the processor address and data bus signals (Host Interface) are logically inverted signals (i.e., the
actual values are inverted of what appears on the processor bus). This must be taken into account and the
addresses and data bus signals must be inverted inside the Intel® 815EM chipset GMCH2-M. All
processor control signals follow normal convention. A 0 indicates an active level (low voltage) if the
19
®
Intel 82815EM GMCH
R
signal is followed by # symbol and a 1 indicates an active level (high voltage) if the signal has no #
suffix.
The following table shows the Vtt/Vdd and Vref levels for the various interfaces:
Table 2. Voltage Levels for Each Interface
2.1.
Interface
Vtt/Vdd
(nominal)
Vref
GTL+
1.5v
2/3 * Vtt
AGP
1.5v
3.3v
1.5v: 0.5 * Vagpdd
3.3v: 0.4 * Vagpdd
Hub Interface
1.8v
0.5 * Vdd
Host Interface Signals
Signal Name
Type
Description
CPURST#
O
GTL+
CPU Reset: The GMCH2-M asserts CPURST# while RESET# (PCIRST# from
ICH2-M) is asserted and for approximately 1ms after RESET# is deasserted. The
GMCH2-M also pulses CPURST# for approximately 1ms when requested via a hub
interface special cycle. The CPURST# allows the processor to begin execution in a
known state.
HA [31:3]#
I/O
GTL+
Host Address Bus: HA[31:3]# connect to the processor address bus. During
processor cycles, HA[31:3]# are inputs. The GMCH2-M drives HA[31:3]# during
snoop cycles on behalf of Primary PCI. Note that the address bus is inverted on the
processor bus.
A low value on HA[15]# sampled at the rising edge of CPURST# informs the
processor to support Quick-Start stop clock mode. If HA[15]# is sampled high at
CPURST# rising edge, it informs the processor to support Stop-Grant mode.
20
HD [63:0]#
I/O
GTL+
Host Data: These signals are connected to the processor data bus. Note that the
data signals are inverted on the processor bus.
ADS#
I/O
GTL+
Address Strobe: The processor bus owner asserts ADS# to indicate the first of two
cycles of a request phase.
BNR#
I/O
GTL+
Block Next Request: Used to block the current request bus owner from issuing a
new request. This signal is used to dynamically control the processor bus pipeline
depth.
BPRI#
O
GTL+
Priority Agent Bus Request: The GMCH2-M is the only Priority Agent on the
processor bus. It asserts this signal to obtain the ownership of the address bus.
This signal has priority over symmetric bus requests and will cause the current
symmetric owner to stop issuing new transactions unless the HLOCK# signal was
asserted.
DBSY#
I/O
GTL+
Data Bus Busy: Used by the data bus owner to hold the data bus for transfers
requiring more than one cycle.
DEFER#
O
GTL+
Defer: The GMCH2-M will generate a deferred response as defined by the rules of
the GMCH2-M dynamic defer policy. The GMCH2-M will also use the DEFER#
signal to indicate a processor retry response.
DRDY#
I/O
GTL+
Data Ready: Asserted for each cycle that data is transferred.
Datasheet
®
Intel 82815EM GMCH
R
Signal Name
Description
HIT#
I/O
GTL+
Hit: Indicates that a caching agent holds an unmodified version of the requested
line. Also driven in conjunction with HITM# by the target to extend the snoop
window.
HITM#
I/O
GTL+
Hit Modified: Indicates that a caching agent holds a modified version of the
requested line and that this agent assumes responsibility for providing the line.
HITM# is also driven in conjunction with HIT# to extend the snoop window.
HLOCK#
I
GTL+
Host Lock: All processor bus cycles sampled with the assertion of HLOCK# and
ADS#, until the negation of HLOCK# must be atomic (i.e. no Hub interface or
GMCH2-M graphics snoopable access to SDRAM is allowed when HLOCK# is
asserted by the processor).
HREQ [4:0]#
I/O
GTL+
Host Request Command: Asserted during both clocks of request phase. In the
first clock, the signals define the transaction type to a level of detail that is sufficient
to begin a snoop request. In the second clock, the signals carry additional
information to define the complete transaction type. The transactions supported by
the GMCH2-M are defined in the Host Interface section of this document.
HTRDY#
I/O
GTL+
Host Target Ready: Indicates that the target of the processor transaction is able to
enter the data transfer phase.
RS [2:0]#
I/O
GTL+
Response Signals: Indicates type of response as shown below:
GTLREF[1:0]
Datasheet
Type
I
RS[2:0]
Response type
000
Idle state
001
Retry response
010
Deferred response
011
Reserved (not driven by the GMCH2-M)
100
Hard Failure (not driven by the GMCH2-M)
101
No data response
110
Implicit Writeback
111
Normal data response
GTL Reference: Reference voltage input for the Host GTL+ interface. GTLREF is
2/3* VTT. VTT is nominally 1.5v.
21
®
Intel 82815EM GMCH
R
2.2.
System Memory Interface Signals
Signal Name
SMAA [12:0]
SMAB [7:4]#
SMAC [7:4]#
Type
O
CMOS
Description
Memory Address: SMAA [12:0], SMAB [7:4]#, and SMAC [7:4]# are used to
provide the multiplexed row and column address to SDRAM. SMAB[7:4]# and
SMAC[7:4]# are inverted (180 degrees out of phase) versions of SMAA[7:4].
SMAC[5]# is default high when sampled for the 815EM to drive HA[15]# high at
CPURST# deassertion time to indicate to the processor to support C2 StopGrant mode. When SMAC[5]# is sampled low, the 815EM will drive HA[15]# low
at CPURST# deassertion time for the processor to support C2 Quick-Start
mode.
SBS[1:0]
O
CMOS
Memory Bank Select: These signals define the banks that are selected within
each DRAM row. The SMAn and SBS signals combine to address every
possible location within a DRAM device.
SBS[1:0] may be heavily loaded and require 2 SDRAM clock cycles for setup
time to the SDRAM’s. For this reason, all chip select signals (SCSA[5:0]# and
SCSB[5:0]#) must be deasserted on any SDRAM clock cycle that one of these
signals change.
2.3.
SMD [63:0]
I/O
CMOS
Memory Data: These signals are used to interface to the SDRAM data bus.
SDQM [7:0]
O
CMOS
Input/Output Data Mask: These pins act as synchronized output enables
during read cycles and as a byte enables during write cycles.
SCSA [5:0]#
SCSB [5:0]#
O
CMOS
Chip Select: For the memory row configured with SDRAM, these pins perform
the function of selecting the particular SDRAM components during the active
state.
SRAS#
O
CMOS
SDRAM Row Address Strobe: These signals drive the SDRAM array directly
without any external buffers.
SCAS#
O
CMOS
SDRAM Column Address Strobe: These signals drive the SDRAM array
directly without any external buffers.
SWE#
O
CMOS
Write Enable Signal: SWE# is asserted during writes to SDRAM.
SCKE [5:0]
O
CMOS
System Memory Clock Enable: SCKE SDRAM Clock Enable is used to signal
a self-refresh or power-down command to an SDRAM array when entering
system suspend. SCKE is also used to dynamically power down inactive
SDRAM rows.
SRCOMP
I/O
System Memory RCOMP: Used to calibrate the System memory I/O buffers.
AGP Interface Signals
For more details on the operation of these signals, refer to the AGP Interface Specification Revision 2.0.
Some of the AGP interfaces are multiplexed with Display Cache interface signals. AGP interface signals
only function as documented in this section when Intel® 815EM chipset AGP interface is enabled (Intel®
815EM chipset integrated graphics disabled). Refer to Section 2.12 for multiplexing map of AGP to
Display Cache interface signals.
22
Datasheet
®
Intel 82815EM GMCH
R
2.3.1.
AGP Addressing Signals
There are two mechanisms by which the AGP master can enqueue AGP requests: PIPE# and SBA (sideband addressing). Upon initialization, one of the methods is chosen. The master may not switch methods
without a full reset of the system. When PIPE# is used to enqueue addresses, the master is not allowed to
queue addresses using the SBA bus. For example, during configuration time, if the master indicates that
it can use either mechanism, the configuration software will indicate which mechanism the master will
use. Once this choice has been made, the master will continue to use the mechanism selected until the
system is reset (and reprogrammed) to use the other mode. This change of modes is not a dynamic
mechanism but rather a static decision when the device is first being configured after reset.
Signal Name
PIPE#
Type
I
AGP
Description
Pipeline Read
During PIPE# Operation: This signal is asserted by the AGP master to indicate a
full-width address is to be enqueued on by the target using the AD bus. One
address is placed in the AGP request queue on each rising clock edge while PIPE#
is asserted.
During SBA Operation: This signal is not used if SBA (Side Band Addressing) is
selected.
During FRAME# Operation: This signal is not used during AGP FRAME#
operation.
SBA[7:0]
I
AGP
Side-band Addressing
During PIPE# Operation: These signals are not used during PIPE# operation.
During SBA Operation: These signals (the SBA, or side-band addressing, bus) are
used by the AGP master (graphics component) to place addresses into the AGP
request queue. The SBA bus and AD bus operate independently. That is,
transactions can proceed on the SBA bus and the AD bus simultaneously.
During FRAME# Operation: These signals are not used during AGP FRAME#
operation.
Datasheet
23
®
Intel 82815EM GMCH
R
2.3.2.
AGP Flow Control Signals
Signal Name
RBF#
Type
I
AGP
Description
Read Buffer Full
During PIPE# and SBA Operation: Read buffer full indicates if the master is ready
to accept previously requested low priority read data. When RBF# is asserted the
GMCH2-M is not allowed to initiate the return low priority read data. That is, the
GMCH2-M can finish returning the data for the request currently being serviced,
however it cannot begin returning data for the next request. RBF# is only sampled
at the beginning of a cycle.
If the AGP master is always ready to accept return read data then it is not required
to implement this signal.
During FRAME# Operation: This signal is not used during FRAME# operation.
WBF#
I
AGP
Write-Buffer Full
During PIPE# and SBA Operation: Write buffer full indicates if the master is ready
to accept Fast Write data from the GMCH2-M. Intel 815EM chipset does not
support the use of AGP Fast Writes.
If the AGP master is always ready to accept fast write data then it is not required to
implement this signal.
During FRAME# Operation: This signal is not used during FRAME# operation.
GMCK
I/OD
CMOS
2.3.3.
GMBus: When configured by register GMBUS[2:0], GMCK becomes a bidirectional
clock signal between master GMCH2-M and slave VCH.
AGP Status Signals
Signal Name
ST[2:0]
Type
O
AGP
Description
Status Bus
During PIPE# and SBA Operation: Provides information from the arbiter to a AGP
Master on what it may do. ST[2:0] only have meaning to the master when its GNT#
is asserted. When GNT# is deasserted these signals have no meaning and must be
ignored. Refer to the AGP Interface Specificaiton revision 2.0 for further explanation
of the ST[2:0] values and their meanings.
During FRAME# Operation: These signals are not used during FRAME# based
operation; except that a ‘111’ indicates that the master may begin a FRAME#
transaction.
24
Datasheet
®
Intel 82815EM GMCH
R
2.3.4.
AGP Clocking Signals - Strobes
Signal Name
ADSTB0
Type
I/O
s/t/s
AGP
Description
AD Bus Strobe-0
During 2X Operation: During 2X operation, this signal provides timing for the
GAD[15:0] and GCBE[1:0]# signals. The agent that is providing the data will drive this
signal.
During 4X Operation: During 4X operation, this is one-half of a differential strobe pair
that provides timing information for GAD[15:0] and GCBE[1:0]# signals.
ADSTB0#
I/O
s/t/s
AGP
AD Bus Strobe-0 Compliment
During 2X Operation: During 2X operation, this signal is not used.
During 4X Operation: During 4X operation, this is one-half of a differential strobe pair
that provides timing information for GAD[15:0] and GCBE[1:0]# signals. The agent
that is providing the data will drive this signal.
ADSTB1
I/O
s/t/s
AGP
AD Bus Strobe-1
During 2X Operation: During 2X operation, this signal provides timing for AD[16:31]
and C/BE[2:3]# signals. The agent that is providing the data will drive this signal.
During 4X Operation: During 4X operation, this is one-half of a differential strobe pair
that provides timing information for the GAD[16:31] and GCBE[2:3]# signals. The
agent that is providing the data will drive this signal.
ADSTB1#
I/O
s/t/s
AGP
AD Bus Strobe-1 Compliment
During 2X Operation: During 2X operation, this signal is not used
During 4X Operation: During 4X operation, this is one-half of a differential strobe pair
that provides timing information for the GAD[16:31] and GCBE[2:3]# signals. The
agent that is providing the data will drive this signal.
SBSTB
I
AGP
SBA Bus Strobe
During 2X Operation: During 2X operation, this signal provides timing for the SBA
bus signals. The agent that is driving the SBA bus will drive this signal.
During 4X Operation: During 4X operation, this is one-half of a differential strobe pair
that provides timing information for the SBA bus signals. The agent that is driving the
SBA bus will drive this signal.
SBSTB#
I
AGP
SBA Bus Strobe Compliment
During 2X Operation: During 2X operation, this signal is not used.
During 4X Operation: During 4X operation, this is one-half of a differential strobe pair
that provides timing information for the SBA bus signals. The agent that is driving the
SBA bus will drive this signal.
Datasheet
GRCOMP
I/O
AGP RCOMP: Used to calibrate AGP I/O buffers. This signal pin must be connected
to a PCB trace representative of the AGP bus data signal traces but sufficiently long
to present a long shelf before signal reflection occurs. The AGP buffers are calibrated
based on the measured shelf voltage.
AGPREF
I
AGP Reference: Reference voltage input for the AGP interface. AGPREF should be
0.4*VDDAGP when VDD is 3.3V, or 0.5* VDDAGP when VDD is 1.5V.
25
®
Intel 82815EM GMCH
R
2.3.5.
AGP FRAME# Signals
Signal Name
GFRAME#
Type
I/O
s/t/s
AGP
Description
FRAME:
During PIPE# and SBA Operation: Not used by AGP SBA and PIPE# operation.
During FRAME# Operation: GFRAME# is an output when the GMCH2-M acts as
an initiator on the AGP Interface. GFRAME# is asserted by the GMCH2-M to
indicate the beginning and duration of an access. GFRAME# is an input when the
GMCH2-M acts as a FRAME# based AGP target. As a FRAME# based AGP target,
the GMCH2-M latches the C/BE[3:0]# and the AD[31:0] signals on the first clock
edge on which it samples FRAME# active.
GIRDY#
I/O
s/t/s
AGP
Initiator Ready:
During PIPE# and SBA Operation: Not used while enqueueing requests via AGP
SBA and PIPE#, but used during the data phase of PIPE# and SBA transactions.
GIRDY# indicates the AGP compliant master is ready to provide all write data for
the current transaction. Once IRDY# is asserted for a write operation, the master is
not allowed to insert wait states. The assertion of IRDY# for reads indicates that the
master is ready to transfer to a subsequent block (32 bytes) of read data. The
master is never allowed to insert a wait state during the initial data transfer (32
bytes) of a read transaction. However, it may insert wait states after each 32 byte
block is transferred. (There is no relationship between GFRAME# and GIRDY# for
AGP transactions.)
GTRDY#
I/O
s/t/s
AGP
Target Ready:
During PIPE# and SBA Operation: Not used while enqueueing requests via AGP
SBA and PIPE#, but used during the data phase of PIPE# and SBA transactions.
GTRDY# indicates the AGP compliant target is ready to provide read data for the
entire transaction (when the transfer size is less than or equal to 32 bytes). In write
case, it is ready to transfer the initial or subsequent block (32 bytes) of data when
the transfer size is greater than 32 bytes. The target is allowed to insert wait states
at the end of each block data transfer(32 bytes). Each 32-byte block is transferred
on both read and write transactions.
GSTOP#
I/O
s/t/s
AGP
Stop:
During PIPE# and SBA Operation:This signal is not used for PIPE# or SBA
operation.
During FRAME# Operation: STOP# is an input when the GMCH2-M acts as a
FRAME# based AGP initiator and an output when the GMCH2-M acts as a
FRAME# based AGP target. STOP# is used for disconnect, retry, and abort
sequences on the AGP interface.
GDEVSEL#
I/O
s/t/s
AGP
Device Select:
During PIPE# and SBA Operation:This signal is not used during PIPE# or SBA
operation.
During FRAME# Operation: GDEVSEL#, when asserted, indicates that a FRAME#
based AGP target device has decoded its address as the target of the current
access. The GMCH2-M asserts GDEVSEL# based on the SDRAM address range
being accessed by a PCI initiator. As an input it indicates whether any device on the
bus has been selected.
GREQ#
I
AGP
Request:
During SBA Operation: This signal is not used during SBA operation.
During PIPE# and FRAME# Operation: GREQ#, when asserted, indicates that a
FRAME# or PIPE# based AGP master is requesting use of the AGP interface.
26
Datasheet
®
Intel 82815EM GMCH
R
Signal Name
GGNT#
GMDA
Type
O
AGP
I/OD
CMOS
GAD [31:0]
I/O
AGP
Description
Grant:
During SBA, PIPE# and FRAME# Operation: GGNT# along with the information on
the ST[2:0] signals (status bus) indicates how the AGP interface will be used next.
Refer to the AGP Interface Specificaiton revision 2.0 for further explanation of the
ST[2:0] values and their meanings.
GMBUS: When configured by register GMBUS[2:0], GMDA becomes a bidirectional
I/O data signal between master GMCH2-M and slave VCH.
Address/Data Bus:
During PIPE# and FRAME# Operation: GAD[31:0] are used to transfer both
address and data information on the AGP inteface.
During SBA Operation: GAD[31:0] are used to transfer data on the AGP interface.
GCBE [3:0]#
I/O
AGP
Command/Byte Enable:
During FRAME# Operation: During the address phase of a transaction, GCBE[3:0]#
define the bus command. During the data phase GCBE[3:0]# are used as byte
enables. The byte enables determine which byte lanes carry meaningful data. The
commands issued on the GCBE# signals during FRAME# based AGP are the same
GCBE# command described in the PCI 2.1 and 2.2 specifications.
During PIPE# Operation: When an address is enqueued using PIPE#, the C/BE#
signals carry command information. Refer to the AGP 2.0 Interface Specification
Revision 2.0 for the definition of these commands. The command encoding used
during PIPE# based AGP is DIFFERENT than the command encoding used during
FRAME# based AGP cycles (or standard PCI cycles on a PCI bus).
During SBA Operation: These signals are not used during SBA operation.
GPAR
I/O
AGP
Parity:
During FRAME# Operation: GPAR is driven by the GMCH2-M when it acts as a
FRAME# based AGP initiator during address and data phases for a write cycle, and
during the address phase for a read cycle. GPAR is driven by the GMCH2-M when
it acts as a FRAME# based AGP target during each data phase of a FRAME#
based AGP memory read cycle. Even parity is generated across GAD[31:0] and
GCBE[3:0]#.
During SBA and PIPE# Operation: This signal is not used during SBA and PIPE#
operation.
NOTES:
1. LOCK#, SERR#, and PERR# signals are not supported on the AGP Interface (even for PCI operations).
2. PCI signals described in this table behave according to PCI 2.1 specifications when used to perform PCI
transactions on the AGP interface.
Datasheet
27
®
Intel 82815EM GMCH
R
2.3.6.
AGP C3 support Signals
The following pin is associated with ACPI C3 support.
Signal Name
AGPBUSY#
Type
Description
OD
AGP Bus Busy:
CMOS
This signal is generated either by the AGP Graphics Chip in AGP mode or by the
GMCH2-M in graphics mode, but not by both. This is an input to the ICH2-M.
In AGP mode,
Asserted (active low): When asserted, AGPBUSY# indicates that the AGP
device is currently busy and requests that the system not transition to the
C3 state. However, assertion of AGPBUSY# does not guarantee that the
system will not enter the C3 state or perform an Intel SpeedStep™
technology transition. If system is in C3 state, then the assertion of
AGPBUSY# is used to request that the system exit from the C3 state.
The AGP GC must assert AGPBUSY# whenever the AGP GC has a
pending request to use the AGP interface. The AGP GC must assert
AGPBUSY# regardless of which protocol it intends to use on the AGP
interface: SBA, PIPE#, or PCI.
AGPBUSY# may only be asserted by the AGP GC when AGPBUSY# is in
the D0 state and should not be asserted in the D1, D2, or D3 states.
Deasserted (high): When deasserted, AGPBUSY# indicates that the AGP
device is not busy and has no need to use the AGP interface.
In Graphics mode,
Asserted (active low): When asserted, AGPBUSY# indicates the internal
graphics unit is requesting snoop or having interrupt request to be serviced.
Therefore, it requests that the system not transition to the C3 state.
However, assertion of AGPBUSY# does not guarantee that the system will
not enter the C3 state or perform an Intel SpeedStep™ technology
transition. If system is in C3 state, then the assertion of AGPBUSY# is
used to request that the system exit from the C3 state.
Deasserted (high): When deasserted, AGPBUSY# indicates the internal graphics
unit has no pending snoop request nor graphics interrupt request.
28
Datasheet
®
Intel 82815EM GMCH
R
2.4.
Display Cache Interface Signals
Some of the Display Cache interface signals are multiplexed with AGP interface signals. Display Cache
interface signals only function as documented in this section when Intel® 815EM chipset integrated
graphics is enabled (Intel® 815EM chipset AGP interface disabled). Refer to Section 2.12 for
multiplexing map of AGP to Display Cache interface signals.
Signal Name
Type
Description
LCS#
O
CMOS
Chip Select: For the memory row configured with SDRAM, this pin performs the
function of selecting the particular SDRAM components during the active state.
LDQM[3:0]
O
AGP
Input/Output Data Mask: These pins control the memory array and act as
synchronized output enables during read cycles and as a byte enables during write
cycles.
LRAS#
O
CMOS
SDRAM Row Address Strobe: The LRAS# signal is used to generate SDRAM
Command encoded on LRAS#/LCAS#/LWE# signals. When LRAS# is sampled
active at the rising edge of the SDRAM clock, the row address is latched into the
SDRAMs.
LCAS#
O
CMOS
SDRAM Column Address Strobe: The LSCAS# signal is used to generate
SDRAM Command encoded on LSRAS#/LSCAS#/LWE# signals. When LSCAS# is
sampled active at the rising edge of the SDRAM clock, the column address is
latched into the SDRAMs.
LMA[11:0]
O
AGP
Memory Address: LMA[11:0] is used to provide the multiplexed row and column
address to SDRAM.
LWE#
O
CMOS
Write Enable Signal: LWE# is asserted during writes to SDRAM.
LMD[31:0]
I/O
AGP
Memory Data: These signals are used to interface to the SDRAM data bus of
SDRAM array.
L_FSEL
I
CMOS
Display Cache Frequency Select: This signal indicates whether the display cache
is to run at 100MHz or 133MHz. The value of this pin is sampled at de-assertion of
CPURST# to determine display cache frequency.
HIGH = 133MHz (Default)
LOW = 100MHz
Note: L_FSEL has a weak internal pull-up enabled during reset.
Note: 100MHz display cache is a non-validated feature and should be implemented
only if OEM performs validation specifically on this feature.
Datasheet
29
®
Intel 82815EM GMCH
R
2.5.
Hub Interface Signals
Signal Name
2.6.
Description
HL[10:0]
I/O
Hub Interface Signals: Signals used for the hub interface.
HLSTRB
I/O
Packet Strobe: One of two differential strobe signals used to transmit or receive
packet data.
HLSTRB#
I/O
Packet Strobe Compliment: One of two differential strobe signals used to transmit
or receive packet data.
HLREF
I
Ref
HUB reference: Sets the differential voltage reference for the hub interface.
Display Interface Signals
Signal Name
30
Type
Type
Description
VSYNC
O
3.3V
CRT Vertical Synchronization: This signal is used as the vertical sync (polarity is
programmable) or “ Vsync Interval”.
HSYNC
O
3.3V
CRT Horizontal Synchronization: This signal is used as the horizontal sync
(polarity is programmable) or “ Hsync Interval”.
IWASTE
I
Ref
Waste Reference: This signal must be tied to ground.
IREF
I
Ref
Set pointer resistor for the internal color palette DAC. A 174 ohm 1% resistor is
recommended
RED
O
Analog
CRT Analog video output from the internal color palette DAC: The DAC is
designed for a 37.5 ohms equivalent load on each pin (e.g. 75 ohms resistor on the
board, in parallel with the 75 ohms CRT load)
GREEN
O
Analog
CRT Analog video output from the internal color palette DAC: The DAC is
designed for a 37.5 ohms equivalent load on each pin (e.g. 75 ohms resistor on the
board, in parallel with the 75 ohms CRT load)
BLUE
O
Analog
CRT Analog video output from the internal color palette DAC: The DAC is
designed for a 37.5 ohms equivalent load on each pin (e.g., 75 ohms resistor on the
board, in parallel with the 75 ohms CRT load)
DDCK
I/O
CMOS
CRT Monitor DDC Interface Clock: (Also referred to as VESA™ “Display Data
Channel”, also referred to as the “Monitor Plug-n-Play” interface.) For DDC1, DDCK
and DDDA provide a unidirectional channel for Extended Display ID. For DDC2,
DDCK and DDDA can be used to establish a bi-directional channel based on I2C
protocol. The host can request Extended Display ID or Video Display Interface
information over the DDC2 channel.
DDDA
I/O
CMOS
CRT Monitor DDC Interface Data: See DDCK Description
Datasheet
®
Intel 82815EM GMCH
R
2.7.
Digital Video Output Signals/TV-Out Pins
Signal Name
LTVCLKIN/STALL
(DVOCLKIN)
Type
I
1.8V
Description
Low Voltage TV Clock In (TV-Out Mode): In 1.8V TV-Out usage, the
TVCLKIN pin functions as a pixel clock input to the GMCH2-M from the TV
encoder. The TVCLKIN frequency ranges from 20MHz to 40MHz
depending on the mode (e.g., NTSC or PAL) and the overscan
compensation values in the TV Encoder. CLKIN has a worse case duty
cycle of 60%/40% coming in to the GMCH2-M.
Flat Panel Interrupt (LCD Mode): STALL flow control: This signal comes
from the external VCH chip. This signal is asserted when VCH is not ready
to take in display data from Intel® 815EM chipset.
LTVCLKOUT[1:0]
(DVOCLKOUT[1:0])
O
1.8V
LCD/TV Port Clock Out: These pins provide a differential pair reference
clock that can run up to 85MHz.
LTVBLANK#
(DVOBLANK)
O
1.8V
Flicker Blank or Border Period Indication: BLANK# is a programmable
output pin driven by the graphics control. When programmed as a blank
period indication, this pin indicates active pixels excluding the border. When
programmed as a border period indication, this pin indicates active pixel
including the border pixels.
LTVDATA[11:0]
(DVODATA[11:0])
O
1.8V
LCD/TV Data: These signals are used to interface to the LCD/TV-out data
bus.
LTVVSYNC
(DVOVSYNC)
O
1.8V
Vertical Sync: VSYNC signal for the LTV interface. The active polarity of
the signal is programmable.
LTVHSYNC
(DVOHSYNC)
O
1.8V
Horizontal Sync: HSYNC signal for the LTV interface. The active polarity
of the signal is programmable.
LTVCK
(DVOI2CCLK)
I/OD
CMOS
LCD/TV Clock: Clock pin for 2-wire interface.
LTVDA
(DVOI2CDATA)
I/OD
CMOS
LCD/TV Data: Data pin for 2-wire interface.
INTRPT# (DVOINT#)
I
INTRPT#: This dedicated pin indicates an interrupt when low. This is used
to support display device hot plug.
CMOS
Datasheet
31
®
Intel 82815EM GMCH
R
2.8.
Power Signals
Signal Name
2.9.
Description
V_1.8
Power
Core Power (1.8V)
VDDQ
Power
AGP I/O and Display Cache Buffer Supply Power
VSUS_3.3
Power
System Memory Buffer Power (Separate 3.3V power plane for power down modes)
VCCDA
Power
Display Power Signal (Connect to an isolated 1.8V plane with VCCDACA1 and
VCCDACA2)
VCCDACA1
Power
Display Power Signal (Connect to an isolated 1.8V plane with VCCDA and
VCCDACA2)
VCCBA
Power
AGP/Hub I/F Power (1.8V)
VCCDACA2
Power
Display Power Signal (Connect to an isolated 1.8V plane with VCCDA and
VCCDACA1)
VCCDPLL
Power
System Memory PLL Power (1.8V)
VSSDA
Power
Display Ground Signal
VSSDACA
Power
Display Ground Signal
VSS
Power
Core Ground
VSSDPLL
Power
System Memory PLL Ground
VSSBA
Power
AGP/Hub I/F Ground
Clock Signals
Signal Name
32
Type
Type
Description
HCLK
I
CMOS
Host Clock Input: Clock used on the host interface. This pin receives a buffered
host clock from the external clock synthesizer. Externally generated 100MHz clock.
This clock goes to Intel® 815EM chipset PSB logic. This clock is also the reference
for System Memory, Local Memory, and internal graphics core clocks. The clock
synthesizer drives this to 2.5V.
SCLK
I
CMOS
System Memory Clock: Clock used on the output buffers of system memory.
Externally generated 100MHz clock.
LTCLK[1:0]
O
CMOS
Display Cache Transmit Clocks: LTCLK[1:0] are internally generated display
cache clocks used to clock the input buffers of the SDRAM devices.
LOCLK
O
CMOS
Output Clock: LOCLK is an internally generated clock used to drive LRCLK.
LRCLK
I
CMOS
Receive Clock: LRCLK is a display cache clock used to clock the input buffers of
the GMCH2-M.
DCLKREF
I
CMOS
Display Interface Clock: DCLKREF is a 48MHz clock generated by an external
clock synthesizer to the GMCH2-M.
HLCLK
I
CMOS
Hub Interface Clock: 66MHz hub interface clock generated by an external clock
synthesizer.
Datasheet
®
Intel 82815EM GMCH
R
2.10.
Miscellaneous Interface Signals
Signal Name
RESET#
2.11.
Type
I
Description
Global Reset: Driven by the ICH/ICH0 when PCIRST# is active.
Reserved
Reserved for future use, Needs to be connected to V_1.8 for backward
compatiblity to GMCH2-M.
NC
No Connect
GMCH2-M Power-Up/Reset Strap Options
Pin Name
Strap Description
Configuration
Interface
Type
Buffer Type
SBA[7]
Local Memory Frequency
Select
High = 133MHz (default)
Low = 100MHz
AGP/LM
Input
SCAS#
Host Frequency
High = Reserved
Low = 100MHz
System
Memory
Bi-directional
SMAA [11]
IOQ Depth
High = 4 (default)
Low = 1
System
Memory
Bi-directional
SMAA [10]
ALL Z
High = Normal
System
Memory
Bi-directional
System
Memory
Bi-directional
System
Memory
Bi-directional
Active Low
SMAA [9]
SMAC[5]#
FSB P-MOS Kicker
Enable
Enable Quick Start
Support
High = Reserved
®
Low = Pentium III processor
and Celeron processor
support
High = Stop-Grant Support
(default)
Low = Quick-Start Support
NOTES:
1. External reset signal used to sample the straps is Reset#.
2. All system memory reset straps have internal 50K ohm pull-ups during reset.
Datasheet
33
®
Intel 82815EM GMCH
R
2.12.
Multiplexed Display Cache and AGP Signal Mapping
Table 3. Display Cache and AGP signal Mapping
Local Memory
Signal Name
34
AGP Signal Name
Local Memory
Signal Name
AGP Signal Name
LCAS#
G_AD26
LMD16
G_AD21
LCKE
G_AD24
LMD17
G_AD23
LCS#
G_STOP#
LMD18
G_AD25
LDQM0
G_AD0
LMD19
G_AD27
LDQM1
G_AD10
LMD2
G_AD5
LDQM2
SBA2
LMD20
G_AD29
LDQM3
ST1
LMD21
G_AD31
L_FSEL
SBA7
LMD22
SBA6
LMA0
G_GAD22
LMD23
SBA4
LMA1
G_AD15
LMD24
PIPE#
LMA10
G_FRAME#
LMD25
SBA1
LMA11/LBA
G_AD18
LMD26
SBA3
LMA2
G_AD11
LMD27
G_REQ#
LMA3
G_BE0#
LMD28
ST0
LMA4
G_AD9
LMD29
ST2
LMA5
G_AD13
LMD3
G_AD3
LMA6
G_PAR
LMD30
RBF#
LMA7
G_TRDY#
LMD31
SBA0
LMA8
G_AD16
LMD4
G_AD1
LMA9
G_AD20
LMD5
G_AD6
LMD0
G_AD8
LMD6
G_AD4
LMD1
G_AD7
LMD7
G_AD2
LMD10
G_BE1#
LMD8
G_AD12
LMD11
G_DEVSEL#
LMD9
G_AD14
LMD12
G_IRDY#
LRAS#
G_BE3#
LMD13
G_BE2#
LTCLK0
G_AD30
LMD14
G_AD17
LTCLK1
G_AD28
LMD15
G_AD19
LWE#
SBA5
Datasheet
®
Intel 82815EM GMCH
R
3.
PCI Configuration Registers
This section describes the PCI Configuration Register set.
The GMCH2-M contains PCI configuration registers for Device 0 (Host-hub interface Bridge/DRAM
Controller), Device 1 (AGP Bridge), and Device 2 (GMCH2-M internal graphics device).
The GMCH2-M also contains an extensive set of registers and instructions for controlling its graphics
operations. Intel® graphics drivers provide the software interface at this architectural level. The
register/instruction interface is transparent at the Application Programmers Interface (API) level and
thus, beyond the scope of this document.
3.1.
Datasheet
Register Nomenclature and Access Attributes
RO
Read Only. If a register is read only, writes to this register have no effect.
R/w
Read/Write. A register with this attribute can be read and written
R/WC
Read/Write Clear. A register bit with this attribute can be read and written. However, a write of a
1 clears (sets to 0) the corresponding bit and a write of a 0 has no effect.
R/WO
Read/Write Once. A register bit with this attribute can be written to only once after power up.
After the first write, the bit becomes read only.
Reserved
Bits
Some of the GMCH2-M registers described in this section contain reserved bits. These bits are
labeled "Reserved” or “Intel® Reserved”. Software must deal correctly with fields that are
reserved. On reads, software must use appropriate masks to extract the defined bits and not rely
on reserved bits being any particular value. On writes, software must ensure that the values of
reserved bit positions are preserved. That is, the values of reserved bit positions must first be
read, merged with the new values for other bit positions and then written back. Note the software
does not need to perform read, merge, write operation for the configuration address register.
Reserved
Registers
In addition to reserved bits within a register, the GMCH2-M contains address locations in the
configuration space of the Host-hub interface Bridge/DRAM Controller and the internal graphics
device entities that are marked either "Reserved” or Intel® Reserved”. When a “Reserved”
register location is read, a random value can be returned. (“Reserved” registers can be 8-, 16-, or
32-bit in size). Registers that are marked as “Reserved” must not be modified by system
software. Writes to “Reserved” registers may cause system failure.
Default
Value Upon
Reset
Upon a Full Reset, the GMCH2-M sets all of its internal configuration registers to predetermined
default states. Some register values at reset are determined by external strapping options. The
default state represents the minimum functionality feature set required to successfully bring up
the system. Hence, it does not represent the optimal system configuration. It is the responsibility
of the system initialization software (usually BIOS) to properly determine the DRAM
configurations, operating parameters and optional system features that are applicable, and to
program the GMCH2-M registers accordingly.
35
®
Intel 82815EM GMCH
R
3.2.
GMCH2-M Register Introduction
The GMCH2-M contains two sets of software accessible registers, accessed via the Host I/O address
space:
• I/O mapped Control registers in the host I/O space: These register controls the access to PCI
configuration space (see section entitled I/O Mapped Registers)
• Internal GMCH2-M configuration registers: These are partitioned into three logical device register
sets (“logical” since they reside within a single physical device).
 Device #0: Host-hub interface Bridge/DRAM Controller functionality controls PCI bus 0 such
as PCI registers, DRAM configuration, other chip-set operating parameters and optional
features.
 Device #1: The second register block is dedicated to the AGP interface.
 Device #2: The third block is dedicated to the internal graphics device (GFX) in the GMCH2M.
The GMCH2-M supports PCI configuration space accesses using the mechanism denoted as
Configuration Mechanism #1 in the PCI specification. The GMCH2-M internal registers (both I/O
Mapped and Configuration registers) are accessible by the host. The registers can be accessed as Byte,
Word (16-bit), or Dword (32-bit) quantities, with the exception of CONFIG_ADDRESS, which can only
be accessed as a Dword. All multi-byte numeric fields use "little-endian" ordering (i.e., lower addresses
contain the least significant parts of the field).
3.3.
I/O Mapped Registers
The GMCH2-M contains two registers that reside in the processor I/O address space
• Configuration Address Register (CONFIG_ADDRESS):
CONFIG_ADDRESS is a 32 bit register accessed only when referenced as a Dword. A Byte or
Word reference will "pass through" the Configuration Address Register onto the PCI0 bus as an
I/O cycle. The CONFIG_ADDRESS register contains the PCI Bus Number, PCI Device
Number, PCI Function Number, and PCI Register Number for which a subsequent configuration
access is intended. The register numbers, bit 7:2 in CONFIG_ADDRESS, is the upper 6 bits of
the dword-aligned byte offset address of the selected dword.
• Configuration Data Register (CONFIG_DATA):
CONFIG_DATA is a 32 bit Read/Write window into the specified dword of the pCI
configuration space selected by CONFIG_ADDRESS. The portion of configuration space that is
referenced by CONFIG_DATA is determined by the contents of CONFIG_ADDRESS. Reads
and writes to the PCI Configuration space can be byte, word, or dword, via I/O read or write
instructions to the CONFIG_DATA port.
The Configuration Address Register enables/disables the configuration space and determines what
portion of configuration space is visible through the Configuration Data window.
36
Datasheet
®
Intel 82815EM GMCH
R
3.3.1.
CONFIG_ADDRESS
Configuration Address Register
I/O Address:
Default Value:
Access:
Size:
0CF8h Accessed as a DWord
00000000h
Read/Write
32 bits
The CONFIG_ADDRESS register contains the Bus Number, Device Number, Function Number, and
Register Number for which a subsequent configuration access is intended.
31
30
24
CFGE
16
Reserved (0)
15
11
Device Number
Bit
23
10
Bus Number
8
Function Number
7
2
Register Number
1
0
Reserved
Descriptions
31
Configuration Enable (CFGE). When this bit is set to 1 accesses to PCI configuration space are
enabled. If this bit is reset to 0 accesses to PCI configuration space are disabled.
30:24
Reserved. These bits are read only and have a value of 0.
23:16
Bus Number. When the Bus Number is programmed to 00h the target of the Configuration Cycle is
one of the three devices in the GMCH2-M or the PCI Bus (the hub interface is logically a PCI bus) that
is directly connected to the GMCH2-M, depending on the Device Number field.
A Type 0 configuration cycle is generated on the hub interface if the Bus Number is programmed to
00h and the GMCH2-M is not the target.
A Type 1 configuration cycle is generated on the hub interface if the Bus Number is non-zero, and is
less than the value programmed into the SECONDARY BUS NUMBER or is greater than the value
programmed into the SUBORDINATE BUS NUMBER Register.
A Type 0 PCI configuration cycle is generated on the AGP bridge if the Bus Number is non-zero and
matches the value programmed into the SECONDARY BUS NUMBER Register.
A Type 1 PCI configuration cycle is generated on the AGP bridge if the Bus Number is non-zero,
greater than the value in the SECONDARY BUS NUMBER register and less than or equal to the value
programmed into the SUBORDINATE BUS NUMBER Register.
Datasheet
37
®
Intel 82815EM GMCH
R
Bit
15:11
Descriptions
Device Number. This field selects one agent on the PCI bus selected by the Bus Number.
During a Type 1 Configuration cycle this field is mapped to AD[15:11].
During a Type 0 Configuration Cycle this field is decoded and one bit among AD[31:11] is driven to a
1.
The GMCH2-M is always Device Number 0 for the Host bridge (GMCH2-M) entity,
The AGP bridge entity is always Device Number 1 and,
The Internal Graphics Device entity is always Device number 2.
If the Bus Number is non-zero and matches the value programmed into the SECONDARY BUS
NUMBER Register a Type 0 PCI configuration cycle will be generated on the AGP bridge. The Device
Number field is decoded and the GMCH2-M asserts one and only one GADxx signal as an IDSEL.
GAD16 is asserted to access Device #0, GAD17 for Device #1, GAD18 for Device #2 and so forth up
to Device #15 which will assert AD31. All device numbers higher than 15 cause a type 0 configuration
access with no IDSEL asserted, which will result in a Master Abort reported in the GMCH2-M’s “virtual”
PCI-PCI bridge registers.
For Bus Numbers resulting in hub interface configuration cycles the GMCH2-M propagates the Device
Number field as A[15:11]. For Bus Numbers resulting in AGP bridge Type 1 Configuration cycles the
Device Number is propagated as GAD[15:11].
3.3.2.
10:8
Function Number. This field is mapped to AD[10:8] during PCIx configuration cycles. This allows the
configuration registers of a particular function in a multi-function device to be accessed. The GMCH2M only responds to configuration cycles with a function number of 000b; all other function number
values attempting access to the GMCH2-M (Device Number = 0, 1 or 2, Bus Number = 0) will generate
a master abort.
7:2
Register Number. This field selects one register within a particular Bus, Device, and Function as
specified by the other fields in the Configuration Address Register. This field is mapped to AD[7:2]
during PCI configuration cycles.
1:0
Reserved.
CONFIG_DATA
Configuration Data Register
I/O Address:
Default Value:
Access:
Size:
0CFCh
00000000h
Read/Write
32 bits
CONFIG_DATA is a 32 bit Read/Write window into configuration space. The portion of configuration
space that is referenced by CONFIG_DATA is determined by the contents of CONFIG_ADDRESS.
Bit
31:0
38
Descriptions
Configuration Data Window (CDW). If bit 31 of CONFIG_ADDRESS is 1 any I/O reference that falls in
the CONFIG_DATA I/O space will be mapped to configuration space using the contents of
CONFIG_ADDRESS.
Datasheet
®
Intel 82815EM GMCH
R
3.4.
PCI Bus Configuration Mechanism
The PCI Bus defines a slot based "configuration space" that allows each device to contain up to 8
functions with each function containing up to 256 8-bit configuration registers. The PCI specification
defines two bus cycles to access the PCI configuration space: Configuration Read and Configuration
Write. Memory and I/O spaces are supported directly by the processor. Configuration space is supported
by a mapping mechanism implemented within the GMCH2-M.
The PCI specification defines two mechanisms to access configuration space, Mechanism #1 and
Mechanism #2. The GMCH2-M supports only Mechanism #1. The configuration access mechanism
makes use of the CONFIG_ADDRESS Register and CONFIG_DATA Register. To reference a
configuration register a Dword I/O write cycle is used to place a value into CONFIG_ADDRESS that
specifies the PCI bus, the device on that bus, the function within the device, and a specific configuration
register of the device function being accessed. CONFIG_ADDRESS[31] must be 1 to enable a
configuration cycle. CONFIG_DATA then becomes a window into the four bytes of configuration space
specified by the contents of CONFIG_ADDRESS. Any read or write to CONFIG_DATA will result in
the GMCH2-M translating the CONFIG_ADDRESS into the appropriate configuration cycle.
The GMCH2-M is responsible for translating and routing the processor I/O accesses to the
CONFIG_ADDRESS and CONFIG_DATA registers to internal GMCH2-M configuration registers, the
internal graphic device, or the hub interface.
3.5.
PCI Configuration Space Access
The GMCH2-M and the ICH2-M are physically connected via the hub interface. From a configuration
standpoint, the hub interface connecting the GMCH2-M and the ICH2-M is logically PCI bus #0. All
devices internal to the GMCH2-M and ICH2-M appear to be on PCI bus #0. The system primary PCI
expansion bus is physically attached to the ICH2-M and, from a configuration standpoint appears as a
hierarchical PCI bus behind a PCI-to-PCI bridge. The primary PCI expansion bus connected to the
ICH2-M has a programmable PCI Bus number.
The GMCH2-M contains three PCI devices within a single physical component. The configuration
registers for both Devices 0, 1 and 2 are mapped as devices residing on PCI bus #0.
Note:
Datasheet
• Device 0:
Host-hub interface Bridge/DRAM Controller.
• Device 1:
AGP Bridge supporting 1X/2X/4X transactions.
• Device 2:
GMCH2-M internal graphics device (GFX).
Even though the primary PCI expansion bus is referred to as PCI0 in this document it is not PCI bus #0
from a configuration standpoint. Note that a physical PCI bus #0 does not exist. The hub interface and
the internal devices in the GMCH2-M and ICH2-M logically constitute PCI Bus #0 to configuration
software.
39
®
Intel 82815EM GMCH
R
3.5.1.
Logical PCI Bus #0 Configuration Mechanism
The GMCH2-M decodes the Bus Number (bits 23:16) and the Device Number fields of the CONFIG_
ADDRESS register. If the Bus Number field of CONFIG_ADDRESS is 0 the configuration cycle is
targeting a PCI Bus #0 device.
Device #0: The Host-hub Bridge/DRAM Controller entity is hardwired as Device #0 on PCI Bus #0.
Device #1: The AGP interface entity is hardwired as Device #1 on PCI Bus #0.
Device #2: The internal graphics device (GFX) entity is hardwired as Device #2 on PCI Bus #0.
Configuration cycles to one of the GMCH2-M internal devices are confined to the GMCH2-M and not
sent over the Hub interface. Accesses to devices #3 to #31 on PCI Bus #0 will be forwarded over the
Hub interface.
3.5.2.
Primary PCI (PCI0) and Downstream Configuration Mechanism
If the Bus Number in the CONFIG_ADDRESS is non-zero the GMCH2-M will generate a configuration
cycle over the hub interface. The ICH2-M compares the non-zero Bus Number with the SECONDARY
BUS NUMBER and SUBORDINATE BUS NUMBER registers of its P2P bridges to determine if the
configuration cycle is meant for Primary PCI expansion bus (PCI0), or a downstream PCI bus.
3.5.3.
Internal Graphics Device (GFX) Configuration Mechanism
From the chipset configuration perspective the internal graphics device is seen as a PCI device (device
#2) on PCI Bus #0. Configuration cycles that target device #2 on PCI Bus #0 will be claimed by the
internal graphics device and will not be forwarded via hub interface to the ICH2-M.
3.6.
Host-Hub Interface Bridge/DRAM Controller Device
Registers (Device #0)
Device #0 registers come in two categories
• Those visible in both GFX mode and AGP mode.
• Those visible only in AGP mode. These are always appended with the message AGP Mode Only.
When GMCH2-M is in internal graphics (GFX) mode most of the configuration bits needed to configure
Device #0 are needed in AGP mode as well. The few exceptions are don’t care situations in AGP. In
AGP mode, a number of bits to control the AGP port are made visible only in AGP mode. They will
read back 0s in GFX mode. The following table shows the GMCH2-M configuration space for device
#0.
40
Datasheet
®
Intel 82815EM GMCH
R
Table 4. GMCH2-M PCI Configuration Space (Device #0)
Address
Offset
Register
Symbol
00–01h
VID
02–03h
DID
04–05h
06–07h
Register Name
Default Value
Access
Vendor Identification
8086h
RO
Device Identification (Device 0)
1130h
RO
PCICMD
PCI Command Register
0006h
Read/Write
PCISTS
PCI Status Register
0090h
Read/Write,
RO
08h
RID (A0)
Revision Identification for A0-step
10h
RO
08h
RID (A1)
Revision Identification for A1-step
11h
RO
09h



0Ah
SUBC
Sub-Class Code
00h
RO
0Bh
BCC
Base Class Code
06h
RO
0Ch



0Dh
MLT
Master Latency Timer
00h
RO
0Eh
HDR
Header Type
00h
RO
0Fh



10-13h
APBASE
00000008h
(AGP)
Read/Write,
Reserved
Reserved
Reserved
Aperture Base Configuration
RO
00000000h (GFX)
14-2Bh

2C–2Dh
SVID
2E–2Fh
SID
30–33h

34h
CAPPTR
35–4Fh

50h
GMCHCFG
51h
APCONT


Subsystem Vendor Identification
0000h
Read/Write
Subsystem Identification
0000h
Read/Write


88h
RO


01ss 0s00b
Read/Write
00h
Read/Write,
Reserved
Reserved
Capabilities Pointer
Reserved
GMCH2-M Configuration
Aperture Control
RO
52h
DRP
DRAM Row Population
00h
Read/Write
53h
DRAMT
DRAM Timing Register
00h
Read/Write
54h
DRP2
DRAM Row Population Register 2
00h
Read/Write
55–57h



58h
FDHC
Fixed DRAM Hole Control
00h
Read/Write
59-5Fh
PAM
Programmable Attributes Map
00h
Read/Write
60h
C3STAT
C3 control and status
00h
Read/Write,
Reserved
RO
61–6Fh
Datasheet

Reserved


41
®
Intel 82815EM GMCH
R
Address
Offset
Register
Symbol
70h
SMRAM
Register Name
System Management RAM Control
Default Value
Access
00h
Read/Write,
RO
71h

72-73h
MISCC
Reserved
Miscellaneous Control Register


0000h
Read/Write,
RO
74–87h



88-8Bh
CAPID
F104 A009h
RO
8C–91h



92–93h
BUFF_SC
Buffer Strength Control
FFFFh
Read/Write
94–95h
BUFF_SC2
Buffer Strength Control 2
FFFFh
Read/Write
96–9Fh



A0-A3h
ACAPID
AGP Capability Identifier
0020 0002h
RO (AGP
only)
A4-A7h
AGPSTAT
AGP Status
1F00 0207h
RO (AGP
only)
A8-ABh
AGPCMD
AGP Command
0000 0000h
Read/Write
Reserved
Capability Identification
Reserved
Reserved
(AGP only)
AC-AFh

B0-B3h
AGPCTRL
Reserved
AGP Control


0000 0000h
Read/Write
(AGP only)
B4h
APSIZE
Aperture Size
00h
Read/Write
(AGP only)
B5-B7h

B8-BBh
ATTBASE
Reserved
Aperture Translation Table Base


0000 0000h
Read/Write
(AGP only)
BCh
AMTT
AGP Multi-Transaction Timer
00h
Read/Write
(AGP only)
BDh
LPTT
Low Priority Transaction Timer
00h
Read/Write
(AGP only)
BEh
MCHCFG
MCH Configuration
0000 x000b
Read/Write,
RO
BF-CAh

CBh
ERRCMD
Reserced
Error Command


00h
Read/Write
(AGP only)
CC-FFh
42

Reserved


Datasheet
®
Intel 82815EM GMCH
R
3.6.1.
VID—Vendor Identification Register (Device 0)
Address Offset:
Default Value:
Attribute:
Size:
00 - 01h
8086h
Read Only
16 bits
The VID Register contains the vendor identification number. This 16-bit register combined with the
Device Identification Register uniquely identifies any PCI device. Writes to this register have no effect.
Bit
15:0
3.6.2.
Description
Vendor Identification Number. This is a 16-bit value assigned to Intel®. Intel® VID = 8086h.
DID—Device Identification Register (Device 0)
Address Offset:
Default Value:
Attribute:
Size:
02 - 03h
1130h
Read Only
16 bits
This 16-bit register combined with the Vendor Identification register uniquely identifies any PCI device.
Writes to this register have no effect.
Bit
15:0
Datasheet
Description
Device Identification Number. This is a 16 bit value assigned to the GMCH2-M Host-Hub Interface
Bridge / DRAM Controller Device 0.
1130h = Device ID for Device 0.
43
®
Intel 82815EM GMCH
R
3.6.3.
PCICMD—PCI Command Register (Device 0)
Address Offset:
Default:
Access:
Size:
04-05h
0006h
Read/Write
16 bits
The PCICMD Register enables and disables the SERR# signal.
15
10
Reserved (0)
9
8
(HW=0)
SERR En
7
6
5
4
3
2
1
0
(HW=0)
(HW=0)
(HW=0)
(HW=0)
(HW=0)
(HW=1)
(HW=1)
(HW=0)
Bit
Descriptions
15:10
Reserved.
9
Fast Back-to-Back. Not implemented by GMCH2-M; hardwired to 0.
8
SERR Enable (SERRE). This bit is a global enable bit for Device 0 SERR messaging. The GMCH2-M does not have an
SERR# signal. The GMCH2-M communicates the SERR# condition by sending an SERR message to the ICH. If this bit
is set to a 1, the GMCH2-M is enabled to generate SERR messages over the Hub interface for specific Device 0 error
conditions If SERRE is reset to 0, then the SERR message is not generated by the GMCH2-M for Device 0.
NOTE: This bit only controls SERR messaging for the Device 0. Device 1 has its own SERRE bit to control error
reporting for error conditions occurring on Device 1. The two control bits are used in a logical OR manner to
enable the SERR hub interface message mechanism.
7
Address/Data Stepping. Not implemented by GMCH2-M; hardwired to 0
6
Parity Error Enable (PERRE). Not implemented by GMCH2-M;hardwired to 0. Writes to this bit position have no affect.
5
VGA Palette Snoop. Not implemented by GMCH2-M; hardwired to 0. Writes to this bit position have no affect.
4
Memory Write and Invalidate Enable. Not implemented, is hardwired to 0. Writes to this bit position have no affects.
3
Special Cycle Enable. (Not implemented by GMCH2-M; hardwired to 0). GMCH2-M ignores all special cycles
generated on the PCI.
2
Bus Master Enable (BME). (Not implemented by GMCH2-M; hardwired to 1). GMCH2-M is always allowed to be a
Bus Master. . Writes to this bit position have no affect.
1
Memory Access Enable (MAE). (Not implemented by GMCH2-M; hardwired to 1). GMCH2-M always allows access
to main memory. Writes to this bit position have no affect.
0
I/O Access Enable (IOAE). (Not implemented; hardwired to 0). Writes to this bit position have no affect.
44
Datasheet
®
Intel 82815EM GMCH
R
3.6.4.
PCISTS—PCI Status Register (Device 0)
Address Offset:
Default Value:
Access:
Size:
06-07h
0090h
Read Only, Read/Write Clear
16 bits
PCISTS is a 16-bit status register that reports the occurrence of a PCI master abort and PCI target abort
on the PCI0 bus. PCISTS also indicates the DEVSEL# timing that has been set by the GMCH2-M
hardware for target responses on the PCI0 bus. Bits [15:12] and bit 8 are Read/Write clear and bits
[10:9] are read only.
15
14
13
12
11
0
SSE
RMAS
RTAS
(HW=0)
7
6
(HW=1)
Bit
Datasheet
5
Reserved
4
10
9
00
8
(HW=0)
3
(HW=1)
0
Reserved
Descriptions
15
Detected Parity Error (DPE). This bit is hardwired to a 0. Writes to this bit position have no affect.
14
Signaled System Error (SSE). This bit is set to 1 when GMCH2-M Device 0 generates an SERR
message over the hub interface for any enabled Device 0 error condition. Device 0 error conditions are
enabled in the PCICMD register. Device 0 error flags are read/reset from the PCISTS register. Software
sets SSE to 0 by writing a 1 to this bit.
13
Received Master Abort Status (RMAS). This bit is set when the GMCH2-M generates a hub
interface request that receives a Master Abort completion packet. Software clears this bit by writing a 1
to it.
12
Received Target Abort Status (RTAS). This bit is set when the GMCH2-M generates a hub interface
request that receives a Target Abort completion packet. Software clears this bit by writing a 1 to it.
11
Signaled Target Abort Status (STAS). (Not implemented in GMCH2-M; is hardwired to a 0). Writes to
this bit position have no affect.
10:9
DEVSEL# Timing (DEVT). These bits are hardwired to 00. Writes to these bit positions have no affect.
Device 0 does not physically connect to PCI0. These bits are set to 00 (fast decode) so that optimum
DEVSEL timing for PCI0 is not limited by GMCH2-M.
8
Data Parity Detected (DPD). This bit is hardwired to a 0. Writes to this bit position have no affect.
7
Fast Back-to-Back (FB2B). This bit is hardwired to 1. Writes to these bit positions have no affect.
Device 0 does not physically connect to PCI. This bit is set to 1 (indicating fast back-to-back capability)
so that the optimum setting for PCI is not limited by GMCH2-M.
6:5
Reserved.
4
Capability List (CLIST). This bit is hardwired to ‘1’ to indicate that GMCH2-M always has a capability
list. The list of capabilities is accessed via register CAPPTR at configuration address offset 34h.
Register CAPPTR contains an offset pointing to the address of the first of a linked list of capability
registers. Writes to this bit position have no affect.
3:0
Reserved.
45
®
Intel 82815EM GMCH
R
3.6.5.
RID—Revision Identification Register (Device 0)
Address Offset:
Default Value:
Access:
Size:
08h
10h for A0-step silicon
11h for A1-step silicon
Read Only
8 bits
This register contains the revision number of the Device 0. These bits are read only and writes to this
register have no effect.
Bit
7:0
3.6.6.
Description
Revision Identification Number. -RO- This is an 8-bit value that indicates the revision identification
number for Device 0. This value is 10h.
SUBC—Sub-Class Code Register (Device 0)
Address Offset:
Default Value:
Access:
Size:
0Ah
00h
Read Only
8 bits
This register contains the Sub-Class Code for the GMCH2-M Function #0. This code is 00h indicating a
Host Bridge device. The register is read only.
Bit
7:0
3.6.7.
Description
Sub-Class Code (SUBC). –RO- This is an 8-bit value that indicates the category of Bridge into which
GMCH2-M falls. The code is 00h indicating a Host Bridge.
BCC—Base Class Code Register (Device 0)
Address Offset:
Default Value:
Access:
Size:
0Bh
06h
Read Only
8 bits
This register contains the Base Class Code of the GMCH2-M Function #0. This code is 06h indicating a
Bridge device. This register is read only.
Bit
7:0
46
Description
Base Class Code (BASEC). This is an 8-bit value that indicates the Base Class Code for GMCH2-M.
This code has the value 06h, indicating a Bridge device.
Datasheet
®
Intel 82815EM GMCH
R
3.6.8.
MLT—Master Latency Timer Register (Device 0)
Address Offset:
Default Value:
Access:
Size:
0Dh
00h
Read Only
8 bits
Device 0 in GMCH2-M is not a PCI master. Therefore this register is not implemented.
Bit
7:0
3.6.9.
Descriptions
Master Latency Timer Value. This read only field always returns 0 when read and writes have no
affect.
HDR—Header Type Register (Device 0)
Address Offset:
Default:
Access:
Size:
0Eh
00h
Read Only
8 bits
This register identifies the header layout of the configuration space. No physical register exists at this
location.
Bit
7:0
Datasheet
Descriptions
Header Type. This read only field always returns 0 when read and writes have no affect.
47
®
Intel 82815EM GMCH
R
3.6.10.
APBASE—Aperture Base Configuration Register (Device 0 AGP MODE ONLY)
Address Offset:
Default Value (AGP Mode):
Default Value (GFX Mode):
Access:
Size:
10-13h
00000008h
00000000h
Read/Write, Read Only
32 bits
The APBASE is a standard PCI Base Address register that is used to set the base of the AGP Aperture.
The standard PCI Configuration mechanism defines the base address configuration register such that only
a fixed amount of space can be requested (dependent on which bits are hardwired to “0” or behave as
hardwired to “0”). To allow for flexibility (of the aperture) an additional register called APSIZE is used
as a “back-end” register to control which bits of the APBASE will behave as hardwired to “0”. This
register will be programmed by the GMCH2-M specific BIOS code that will run before any of the
generic configuration software is run.
Note that bit 1 of the APCONT register is used to prevent accesses to the aperture range before this
register is initialized by the configuration software and the appropriate translation table structure has
been established in the main memory.
31
26
Upper Prog. Base Address Bits
15
24
Bit
16
Lower
“HW”/Prog
Base
Address
4
(HW=0)
31:26
25
(HW=0)
3
2
Prefetch
able
1
Type
0
Mem
Space
Indicator
(HW=0)
Description
Upper Programmable Base Address bits (Read/Write). These bits are used to locate the range size
selected via lower bits 25:4.
Default = 0000
25
Lower “Hardwired”/Programmable Base Address bit. This bit behaves as a “hardwired” or as a
programmable depending on the contents of the APSIZE register as defined below:
Bit 25 is controlled by the bit 3 of the APSIZE register in the following manner:
If APSIZE[3]=0 then APBASE[25]=0 indicating 64 MB aperture size:
à 0 (default)
Aperture Size = 64 MB
If APSIZE[3]=1 then APBASE[25] = (Read/Write) allowing 32 MB aperture size: Aperture Size = 32 MB
à Read/Write
Default for APSIZE[3]=0b forces default APBASE[25] = 0b (bit responds as “hardwired” to 0). This
provides a default to the maximum aperture size of 64 MB. The GMCH2-M specific BIOS is responsible
for selecting smaller size (if required) before PCI configuration software runs and establishes the system
address map.
48
Datasheet
®
Intel 82815EM GMCH
R
Bit
Description
24:4
Hardwired to “0”. This forces minimum aperture size selected by this register to be 32 MB.
3
Prefetchable (RO). This bit is hardwired to “1” to identify the Graphics Aperture range as a
prefetchable, i.e.:
There are no side effects on reads, the device returns all bytes on reads regardless of the byte enables,
and the GMCH2-M may merge processor writes into this range without causing errors.
3.6.11.
2:1
Type (RO). These bits determine addressing type and they are hardwired to “00” to indicate that address
range defined by the upper bits of this register can be located anywhere in the 32-bit address space.
0
Memory Space Indicator (RO). Hardwired to “0” to identify aperture range as a memory range.
SVID—Subsystem Vendor Identification Register (Device 0)
Address Offset:
Default:
Access:
Size:
2C-2Dh
0000h
Read/Write Once
16 bits
Bit
15:0
3.6.12.
Description
Subsystem Vendor ID (Read/WriteO). This value is used to identify the vendor of the subsystem.
The default value is 0000h. This field should be programmed by BIOS during boot-up. Once written,
this register becomes Read Only. This Register can only be cleared by a Reset.
SID—Subsystem Identification Register (Device 0)
Address Offset:
Default:
Access:
Size:
Bit
15:0
Datasheet
2E-2Fh
0000h
Read/Write Once
16 bits
Description
Subsystem ID (Read/WriteO). This value is used to identify a particular subsystem. The default value
is 0000h. This field should be programmed by BIOS during boot-up. Once written, this register
becomes Read Only. This Register can only be cleared by a Reset.
49
®
Intel 82815EM GMCH
R
3.6.13.
CAPPTR—Capabilities Pointer (Device 0)
Address Offset:
Default Value:
Access:
Size:
34h
88h
Read Only
8 bits
The CAPPTR provides the offset that is the pointer to the location where the capability identification
register is located.
Bit
Description
7:0
Pointer to the start of the CAPPTR linked list.
The 88h value points to the CAPID register that provides capability information regarding the GMCH2M.
3.6.14.
GMCHCFG—GMCH2-M Configuration Register (Device 0)
Address Offset:
Default:
Access:
Size:
50h
01ss0s00
Read/Write, Read Only
8 bits
7
6
Mem Arb
Gnt Win
Enable
CPU
Latency
Timer
5
Reserved
4
3
2
Local
Memory
Frequency
Select
DRAM
Page
Closing
Policy
System
Memory
Frequency
Select
Bit
7
1
0
Reserved
Description
Memory Arbiter Grant Window Enable (MAGWE). This bit controls the Host vs Low Priority Graphics
timeslice regulation in the arbiter for the System DRAM.
At pre-arbitration (aka, stage 1)
0 = Disabled. Enforce fixed priority.
1 = Limit grant to host-to-graphics stream to 6 consecutive packets.
At main-arbitration (aka, stage 2)
0 = Disabled. Enforce fixed priority.
1 = 24 clocks limiting host, 24 clocks guaranteed to low priority graphics stream.
In fixed mode arbitration, MAGWE=0, the host stream always has higher priority over the low priority
graphics stream for accesses to system memory. In timeslice mode, the host stream and the low
priority graphics stream are both regulated by a time window to provide fairness to the graphics
stream. Fixed priority mode where the host stream is always favored is the recommended mode of
operation, as this setting gives highest system performance without adversely affecting graphics
performance under real life applications workload.
50
Datasheet
®
Intel 82815EM GMCH
R
Bit
6
Description
CPU Latency Timer (CLT).
0 = Deferrable processor cycle will be Deferred immediately after receiving another ADS#
1 = Deferrable processor cycle will be Deferred after it has been held in a “Snoop Stall” for 31 clocks
and another ADS# has arrived (default).
5
Reserved
4
Local Memory Frequency Select (LMFS). This bit selects the operating frequency for the Local
Memory Controller. Default is set by sampling LM_FREQ_SEL strap (AGP SBA[7] pin) at reset. It has
a weak internal pull-up enabled during reset.
0 = 100MHz. This is a reflection of LM_FREQ_SEL strap being pulled down.
1 = 133MHz, (default). This is a reflection of LM_FREQ_SEL strap being pulled up (default).
Note: This bit has meaning only when operating in Internal graphics modes w/ display cache.
Note: The value of this bit should only be changed when the Internal Graphics device is
disabled
(i.e., GMS = 00).
3
DRAM Page Closing Policy (DPCP). When this bit is a 0 GMCH2-M will tend to leave the DRAM
pages open. In this mode the only times that GMCH2-M will close memory pages are:
0 = Precharge Bank during service of a “Page Miss” access.
Precharge All when changing from one Row to another if any Pages are open.
Precharge All at leadin to a Refresh operation
When this bit is a 1 GMCH2-M will tend to leave the DRAM pages closed. In this mode GMCH2-M will:
1 = Precharge All during the service of any “Page Miss” access.
Precharge All when changing from one Row to another if any Pages are open.
Precharge All at leadin to a Refresh operation.
2
System Memory Frequency Select (SMFS). This bit selects the operating frequency for the main
system memory. Default is set by sampling SBS0# pin at reset. When GMCH2-M is in Graphics mode,
this bit is hardwired to ‘0’ (100MHz) and read only.
0 = 100MHz.
1 = 133MHz.
1:0
Datasheet
Reserved.
51
®
Intel 82815EM GMCH
R
3.6.15.
APCONT—Aperture Control (Device 0)
Address Offset:
Default Value:
Access:
Size:
51h
00h
Read/Write, Write Once, Read Only
8 bits
The Aperture Control Register controls selection and access to aperture space.
7
3
Reserved
Bit
2
1
0
AGP
Select
Lock
Aperture
Access
Global EN
AGP
Select
Description
7:3
Reserved
2
GFX AGP Select Lock (WO). This GFX AGP Select (bit 0) can be made read only by this bit. This
is a write once bit, after it is written to it cannot be changed without a system reset.
0 = GFX AGP Select remains writeable.
1 = GFX AGP Select is read only.
1
Aperture Access Global Enable (Read/Write). This bit is used to prevent access to the aperture
from any port (processor, PCI0 or AGP/PCI1) before the aperture range is established by the
configuration software and appropriate translation table in the main DRAM has been initialized.
Default is “0”. It must be set after system is fully configured for aperture accesses.
0
GFX AGP Select. Read/Write, RO if GFX AGP Select Lock (bit 2 =1) This field selects the
graphics device to be either AGP or Internal Graphics (GFX).
0 = AGP Mode. AGP interface device is enabled. All registers in device 0 and device 1 are visible.
No device 2 registers are visible, reads from those addresses will return 1’s.
1 = GFX Mode. Internal Graphics device is enabled. All non-AGP related device 0 registers and all
device 2 registers are visible. No device 1 registers are visible, reads from those addresses will
return 1’s. Reads from AGP related device 0 registers will return 0’s. The internal graphics device will
not respond to any configuration cycles unless SMRAM[7:6] @ 70h are NOT 00 AND APCONT[0] @
is 1.
GFX AGP Select must be programmed before any other access is made to the configuration space.
The two possible modes are mutually exclusive. This bit determines whether other configuration
registers are enabled or disabled.This bit must be set as part of the initialization sequence.
52
Datasheet
®
Intel 82815EM GMCH
R
3.6.16.
DRP—DRAM Row Population Register (Device 0)
Address Offset:
Default Value:
Access:
Size:
52h
00h
Read/Write (Read_Only if D_LCK = 1)
8 bits
The DRAM Row Population Register defines the population of each side of each SO-DIMM. Note that
this entire register becomes RO when the D_LCK bit (SMRAM register Device 0, address offset 70h) is
set.
7
4
3
0
DIMM 1 Population
DIMM 0 Population
Bit
Description
7:4
DIMM 1 Population. This field indicates the population of DIMM 1. (See table below)
3:0
DIMM 0 Population. This field indicates the population of DIMM 0. (See table below)
Table 5. Supported System Memory DIMM Configurations
GMCH2M
Register
Code
DIMM
Capacity
# of
Devices
/ DIMM
# of
Sides
DRAM
Technology
Front Side
Population
Count
0
0
1
32 MB
16
2
32 MB
3
Back Side Population
Config
Count
Row
Bank
Column
Config
N/A
Empty
N/A
N/A
N/A
DS
16 Mb
8-
2 Mb
x8
11
1
9
4
SS
64 Mb
4-
4 Mb
x 16
12
2
8
48 MB
12
DS
64/16 Mb
4-
4 Mb
x 16
8-
2 Mb
x8
12
2/1
8
4
64 MB
8
DS
64 Mb
4-
4 Mb
x 16
4-
4 Mb
x 16
12
2
8
5
64 MB
8
SS
64 Mb
8-
8 Mb
x8
12
2
9
5
64 MB
4
SS
128 Mb
4-
8 Mb
x 16
12
2
9
6
96 MB
12
DS
64 Mb
8-
8 Mb
x8
4-
4 Mb
x 16
12
2
9/8
6
96 MB
8
DS
128/64 Mb
4-
8 Mb
x 16
4-
4 Mb
x 16
12
2
9/8
7
128 MB
16
DS
64 Mb
8-
8 Mb
x8
8-
8 Mb
x8
12
2
9
7
128 MB
8
DS
128 Mb
4-
8 Mb
x 16
4-
8 Mb
x 16
12
2
9
9
128 MB
8
SS
128 Mb
8-
16 Mb
x8
12
2
10
A
128 MB
4
SS
256 Mb
4-
16 Mb
x 16
13
2
9
B
192 MB
12
DS
128 Mb
8-
16 Mb
x8
4-
8 Mb
x 16
12
2
10/9
B
192 MB
16
DS
128/64 Mb
8-
16 Mb
x8
8-
8 Mb
x8
12
2
10/9
C
256 MB
16
DS
128 Mb
8-
16 Mb
x8
8-
16 Mb
x8
12
2
10
D
256 MB
8
DS
256 Mb
4-
16 Mb
x 16
4-
16 Mb
x 16
13
2
9
E
256 MB
8
SS
256 Mb
8-
32 Mb
x8
13
2
10
F
512 MB
16
DS
256 Mb
8-
32 Mb
x8
13
2
10
Datasheet
Empty
8-
8-
2 Mb
32 Mb
x8
x8
53
®
Intel 82815EM GMCH
R
3.6.17.
DRAMT—DRAM Timing Register (Device 0)
Address Offset:
Default Value:
Access:
Size:
53h
00h
Read/Write
8 bits
This register controls the operating mode and the timing of the DRAM Controller.
7
5
SDRAM Mode Select
4
3
2
1
0
DRAM
Cycle
Time
Reserved
CAS#
Latency
SDRAM
RAS# to
CAS# Dly
SDRAM
RAS#
Precharge
Bit
7:5
Description
SDRAM Mode Select (SMS). These bits select the operational mode of the GMCH2-M DRAM
interface. The special modes are intended for initialization at power up.
000 = DRAM in Self-Refresh Mode,
Refresh Disabled (Default)
001 = Normal Operation,
100Mhz System memory - Refresh interval 15.6 uSec
010 = Normal Operation,
100Mhz System memory - Refresh interval 7.8
011 = Normal Operation,
100Mhz System memory - Refresh interval 1.28 uSec
100 = NOP Command Enable. In this mode all processor cycles to SDRAM result in a NOP
Command.
101 = All Banks Precharge Enable. In this mode all processor cycles to SDRAM result in an All
Banks Precharge Command on the SDRAM interface.
110 = Mode Register Set Enable. In this mode all processor cycles to SDRAM result in a mode
register set command on the SDRAM interface. The Command is driven on the MA[12:0] lines. MA[2:0]
must always be driven to 010 for burst of 4 mode. MA3 must be driven to 1 for interleave wrap type.
MA4 needs to be driven to the value programmed in the CAS# Latency bit. MA[6:5] should always be
driven to 01. MA[12:7] must be driven to 00000. BIOS must calculate and drive the correct host
address for each row of memory such that the correct command is driven on the MA[12:0] lines. Note
that MAB[7:4]# are inverted from MAA[7:4]; BIOS must account for this.
111 = CBR Enable. In this mode all processor cycles to SDRAM result in a CBR cycle on the SDRAM
interface.
4
DRAM Cycle Time (DCT). This bit controls the number of SCLKs for an access cycle.
0 = Tras = 5 SCLKs & Trc = 7 SCLKs (Default)
1 = Tras = 7 SCLKs & Trc = 9 SCLKs.
3
Intel® Reserved.
2
CAS# Latency (CL). This bit controls the number of CLKs between when a read command is sampled
by the SDRAMs and when GMCH2-M samples read data from the SDRAMs. 0 = CAS# latency is 3
SCLKs.
1 = CAS# latency is 2 SCLKs.
1
SDRAM RAS# to CAS# Delay (SRCD). This bit controls the number of SCLKs from a Row Activate
command to a read or write command.
0 = 3 clocks will be inserted between a row activate command and either a read or write command.
1 = 2 clocks will be inserted between a row activate and either a read or write command.
54
Datasheet
®
Intel 82815EM GMCH
R
Bit
0
Description
SDRAM RAS# Precharge (SRP). This bit controls the number of SCLKs for RAS# precharge.
0 = 3 clocks of RAS# precharge are provided.
1 = 2 clocks of RAS# precharge are provided
3.6.18.
DRP2—DRAM Row Population Register 2 (Device 0)
Address Offset:
Default Value:
Access:
Size:
54h
00h
Read/Write (Read_Only if D_LCK = 1)
8 bits
This second DRAM Row Population Register (DRP2) defines the population of each side of DIMM 2.
7
4
3
Reserved
Bit
0
DIMM 2 Population
Description
7:4
Reserved.
3:0
DIMM 2 Population. This field indicates the population of DIMM 2. Please refer to Supported System
Memory DIMM Configurations table located with the DRP register definition. Note that not some of the
larger capacity DIMMs may not be supported in DIMM 2 based on the capacities of DIMM 0 and DIMM
1.
The max supported main memory capacity is 512 MB.
Datasheet
55
®
Intel 82815EM GMCH
R
3.6.19.
FDHC  Fixed DRAM Hole Control Register (Device 0)
Address Offset:
Default Value:
Access:
Size:
58h
00h
Read/Write
8 bits
This 8-bit register controls a single fixed DRAM hole: 15-16 MB.
7
6
0
Hole EN
Reserved
Bit
7
Description
Hole Enable (HEN). This field enables a memory hole in DRAM space.
Host cycles matching an enabled hole are passed on to ICH through the hub interface. Hub interface
and PCI cycles matching an enabled hole will be ignored by the GMCH2-M. Note that a selected hole is
not re-mapped.
0 = No Hole Enabled
1 = 15 MB-16 MB (1MB) Hole Enabled
6:0
3.6.20.
Reserved.
PAM
Programmable Attributes Map Registers (Device 0)
Address Offset:
Default Value:
Attribute:
Size:
59 - 5Fh
00h
Read/Write
4 bits/register
The GMCH2-M allows programmable memory attributes on 13 Legacy memory segments of various
sizes in the 768 KB to 1 MB address range. Seven Programmable Attribute Map (PAM) Registers are
used to support these features. Cacheability of these areas is controlled via the MTRR registers in the
mobile Intel Pentium® III processor. Two bits are used to specify memory attributes for each memory
segment. These bits apply to both host, AGP/PCI and hub interface initiator accesses to the PAM areas.
These attributes are:
• RE - Read Enable. When RE = 1, the processor read accesses to the corresponding memory segment
are claimed by the GMCH2-M and directed to main memory. Conversely, when RE = 0, the host
read accesses are directed to the hub interface/PCI0.
• WE - Write Enable. When WE = 1, the host write accesses to the corresponding memory segment
are claimed by the GMCH2-M and directed to main memory. Conversely, when WE = 0, the host
write accesses are directed to the hub interface/PCI0.
The RE and WE attributes permit a memory segment to be Read Only, Write Only, Read/Write, or
disabled. For example, if a memory segment has RE = 1 and WE = 0, the segment is Read Only.
Each PAM Register controls two regions, typically 16 KB in size. Each of these regions has a 4-bit field.
The four bits that control each region have the same encoding and are defined in the following table.
56
Datasheet
®
Intel 82815EM GMCH
R
Table 6. Attribute Bit Assignments
7
6
5
4
3
2
1
0
Description
R
R
0
0
R
R
0
0
Disabled. DRAM is disabled and all accesses are directed to the hub interface. The
GMCH2-M does not respond as a AGP/PCI or hub interface target for any read or write
access to this area.
R
R
0
1
R
R
0
1
Read Only. Reads are forwarded to DRAM and writes are forwarded to the hub
interface for termination. This write protects the corresponding memory segment. The
GMCH2-M will respond as a AGP/PCI or hub interface target for read accesses but not
for any write accesses.
R
R
1
0
R
R
1
0
Write Only. Writes are forwarded to DRAM and reads are forwarded to the hub
interface for termination. The GMCH2-M will respond as a AGP/PCI or hub interface
target for write accesses but not for any read accesses.
R
R
1
1
R
R
1
1
Read/Write. This is the normal operating mode of main memory. Both read and write
cycles from the host are claimed by the GMCH2-M and forwarded to DRAM. The
GMCH2-M will respond as a AGP/PCI or hub interface target for both read and write
accesses.
As an example, consider a BIOS that is implemented on the expansion bus. During the initialization
process, the BIOS can be shadowed in main memory to increase the system performance. When BIOS is
shadowed in main memory, it should be copied to the same address location. To shadow the BIOS, the
attributes for that address range should be set to write only. The BIOS is shadowed by first doing a read
of that address. This read is forwarded to the expansion bus. The host then does a write of the same
address, which is directed to main memory. After the BIOS is shadowed, the attributes for that memory
area are set to read only so that all writes are forwarded to the expansion bus. The table above and the
figure below show the PAM registers and the associated attribute bits:
Datasheet
57
®
Intel 82815EM GMCH
R
Figure 3. PAM Registers
Offset
5Fh
5Eh
5Dh
5Ch
5Bh
5Ah
59h
PAM6
PAM5
PAM4
PAM3
PAM2
PAM1
PAM0
7
6
R
R
5
4
WE RE
3
2
R
R
1
0
WE RE
Reserved
Reserved
Write Enable (R/W)
1=Enable
0=Disable
Read Enable (R/W)
1=Enable
0=Disable
Read Enable (R/W
1=Enable
0=Disable
Write Enable (R/W)
1=Enable
0=Disable
Reserved
Reserved
PAM Reg
Table 7. PAM Registers and Associated Memory Segments
PAM Reg
Attribute Bits
Memory Segment
Comments
Offset
PAM0[3:0]
Reserved
PAM0[7:4]
R
R
WE
RE
0F0000h - 0FFFFFh
BIOS Area
59h
PAM1[3:0]
R
R
WE
RE
0C0000h - 0C3FFFh
ISA Add-on BIOS
5Ah
PAM1[7:4]
R
R
WE
RE
0C4000h - 0C7FFFh
ISA Add-on BIOS
5Ah
PAM2[3:0]
R
R
WE
RE
0C8000h - 0CBFFFh
ISA Add-on BIOS
5Bh
PAM2[7:4]
R
R
WE
RE
0CC000h- 0CFFFFh
ISA Add-on BIOS
5Bh
PAM3[3:0]
R
R
WE
RE
0D0000h- 0D3FFFh
ISA Add-on BIOS
5Ch
PAM3[7:4]
R
R
WE
RE
0D4000h- 0D7FFFh
ISA Add-on BIOS
5Ch
PAM4[3:0]
R
R
WE
RE
0D8000h- 0DBFFFh
ISA Add-on BIOS
5Dh
PAM4[7:4]
R
R
WE
RE
0DC000h- 0DFFFFh
ISA Add-on BIOS
5Dh
PAM5[3:0]
R
R
WE
RE
0E0000h- 0E3FFFh
BIOS Extension
5Eh
PAM5[7:4]
R
R
WE
RE
0E4000h- 0E7FFFh
BIOS Extension
5Eh
PAM6[3:0]
R
R
WE
RE
0E8000h- 0EBFFFh
BIOS Extension
5Fh
PAM6[7:4]
R
R
WE
RE
0EC000h- 0EFFFFh
BIOS Extension
5Fh
58
59h
Datasheet
®
Intel 82815EM GMCH
R
Expansion Area (C0000h-DFFFFh)
This 128 KByte ISA Expansion region is divided into eight 16 KByte segments. Each segment can be
assigned one of four Read/Write states: read-only, write-only, Read/Write, or disabled. Typically, these
blocks are mapped through GMCH2-M and are subtractively decoded to ISA space. Memory that is
disabled is not remapped.
Note:
That GMCH2-M has support for more PAM regions than Intel® 810 chipset.
Extended System BIOS Area (E0000h-EFFFFh)
This 64 KByte area is divided into four 16 KByte segments. Each segment can be assigned independent
read and write attributes so it can be mapped either to main DRAM or to the hub interface. Typically,
this area is used for RAM or ROM. Memory segments that are disabled are not remapped elsewhere.
System BIOS Area (F0000h-FFFFFh)
This area is a single 64 KByte segment. This segment can be assigned read and write attributes. It is by
default (after reset) Read/Write disabled and cycles are forwarded to the hub interface. By manipulating
the Read/Write attributes, the GMCH2-M can “shadow” BIOS into the main DRAM. When disabled,
this segment is not remapped.
3.6.21.
C3STATUS —C3 Control and Status Register (Device #0)
Address Offset:
Default Value:
Access:
Size:
60h
(see table)
Read/Write (some bits Read Only)
32 bits
31
17
Reserved
15
11
Reserved
SDRAM
Power
Down
10
9
8
AGPBUSY
Active
Self Ref
Disable
Reserved
Bit
16
7
0
C3 Entry Delay Counter
Description
31:17
Reserved
16
SDRAM Power Down Feature
0 = SDRAM allowed to enter power down via CKE#
1 = SDRAM not allowed to enter power down (Default)
15:11
Reserved
10
Force AGPBUSY# active.
0 = AGPBUSY# operates normally (Default)
1 = AGPBUSY# asserted (low) continuously.
Datasheet
59
®
Intel 82815EM GMCH
R
Bit
9
Description
Self-Refresh Disable.
0 = Self-Refresh enabled during C3 state. (Default)
1 = Self-Refresh disabled.
8
Reserved: need to be programmed as ‘0’
7:0
C3 entry delay counter. Adjusts the delay for entering the C3 state.
40h = Optimal setting
3.6.22.
SMRAM - System Management RAM Control Register (Device 0)
Address Offset:
Default Value:
Access:
Size:
70h
00h
Read/Write, Read Only
8 bits
The SMRAM register controls how accesses to Compatible and Extended SMRAM spaces are treated,
and how much (if any) memory is used from the System to support both SMRAM and Graphics Local
Memory needs.
7
6
Graphics Mode Select
Bit
7:6
5
4
Upper SMM Select
3
2
Lower SMM Select
1
0
SMM
Space
Locked
E_
SMRAM_
ERR
Description
Graphics Mode Select (GMS). This field is used to enable/disable the Internal Graphics device
(GFX) and select the amount of Main Memory that is “Stolen” to support the Internal Graphics device
in VGA (non-linear) mode only. These 2 bits only have meaning if we are not in AGP mode.
00 = Internal Graphics Device Disabled, No memory “Stolen”
01 = Internal Graphics Device Enabled, No memory “Stolen”
10 = Internal Graphics Device Enabled, 512K of memory “Stolen” for frame buffer.
11 = Internal Graphics Device Enabled, 1M of memory “Stolen” for frame buffer.
Note: When the Internal Graphics Device is Disabled (00) the Graphics Device and all of its
memory and I/O functions are disabled and the clocks to this logic are turned off, memory accesses
to the VGA range (A0000-BFFFF) will be forwarded on to the hub interface, and the Graphics Local
Memory space is NOT “stolen” from main memory. Any change to the SMRAM register will not affect
AGP mode or cause the controller to go into AGP mode.
When this field is non-0 the Internal Graphics Device and all of its memory and I/O functions are
enabled, all non-SMM memory accesses to the VGA range will be handled internally and the
selected amount of Graphics Local Memory space (0, 512K or 1M) is “stolen” from the main
memory. Graphics Memory is “stolen” AFTER TSEG Memory is “stolen”.
Once D_LCK is set, these bits becomes read only. GMCH2-M does not support VGA on local
memory. Software must not use the 01 mode for VGA
60
Datasheet
®
Intel 82815EM GMCH
R
Bit
5:4
Description
Upper SMM Select (USMM). This field is used to enable/disable the various SMM memory ranges
above 1 MB. TSEG is a block of memory (“Stolen” from Main Memory at [TOM-Size]: [TOM]) that is
only accessable by the processor and only while operating in SMM mode. HSEG is a Remap of the
AB segment at FEEA0000: FEEBFFFF. Both of these areas, when enabled, are usable as SMM
RAM.
00 = TSEG and HSEG are both Disabled
01 = TSEG is Disabled, HSEG is Conditionally Enabled
10 = TSEG is Enabled as 512 KB and HSEG is Conditionally Enabled
11 = TSEG is Enabled as 1 MB and HSEG is Conditionally Enabled
Note: Non-SMM Operations (SMM processor accesses and all other access) that use these address
ranges are forwarded to the hub interface.
Once D_LCK is set, these bits becomes read only.
HSEG is ONLY enabled if LSMM = 00.
3:2
Lower SMM Select (LSMM). This field controls the definition of the A&B segment SMM space
00 = AB segment Disabled (no one can write to it).
01 = AB segment Enabled as General System RAM (anyone can write to it).
10 = AB segment Enabled as SMM Code RAM Shadow. Only SMM Code Reads can access DRAM
in the AB segment (processor code reads only). SMM Data operations and all Non-SMM Operations
go to either the internal graphics device or are broadcast on the hub interface.
11 = AB segment Enabled as SMM RAM. All SMM operations to the AB segment are serviced by
DRAM, all Non-SMM Operations go to either the internal Graphics Device or are broadcast on the
hub interface (processor SMM Read/Write can access SMM space).
When D_LCK is set bit 3 becomes Read Only, and bit 2 is Writable ONLY if bit 3 is a “1”.
When bit 3 is set only the processor can access it.
Datasheet
1
SMM Space Locked (D_LCK): When D_LCK is set to 1 then D_LCK, GMS, USMM, and the most
significant bit of LSMM become read only. D_LCK can be set to 1 via a normal configuration space
write but can only be cleared by a reset. The combination of D_LCK and LSMM provide convenience
with security. The BIOS can use LSMM=01 to initialize SMM space and then use D_LCK to “lock
down” SMM space in the future so that no application software (or BIOS itself) can violate the
integrity of SMM space, even if the program has knowledge of the LSMM function. This bit also
Locks the DRP and DRP2 registers.
0
E_SMRAM_ERR (E_SMERR): This bit is set when processor accesses the defined memory ranges
in Extended SMRAM (HSEG or TSEG) while not in SMM mode. It is software’s responsibility to clear
this bit. The software must write a 1 to this bit to clear it This bit is Not set for the case of an Explicit
Write Back operation.
61
®
Intel 82815EM GMCH
R
3.6.23.
MISCC—Miscellaneous Control Register (Device 0)
Address Offset:
Default Value:
Access:
Size:
72-73h
0000h
Read/Write, Read Only
16 bits
This register holds all of the miscellaneous control bits for GMCH2-M.
15
8
Reserved
7
6
Read PWR Throttle Cntl
5
4
Write PWR Throttle Cntl
Bit
3
2
1
0
Throttle
Lock
Reserved
BNR
Looka
head
GFX Tr
Win Size
Sel
Description
15:8
Reserved.
7:6
Read Power Throttle Control. Read/Write,RO if Throttle Lock (bit 3=1). These bits are locked
(RO) when bit 3 (Throttle Lock) is set to 1.These bits select the Power Throttle Bandwidth Limits for
Read operations to System Memory.
00 = No Limit
(800 MB/Sec) (Default Value)
01 = Limit at 87 ½ % (700 MB/Sec)
10 = Limit at 75 %
(600 MB/Sec)
11 = Limit at 62 ½ % (500 MB/Sec)
5:4
3
Write Power Throttle Control. Read/Write,RO if Throttle Lock (bit 3=1). These bits are locked
(RO) when bit 3 (Throttle Lock) is set to 1. These bits select the Power Throttle Bandwidth Limits for
Write operations to System Memory.
00 = No Limit
(800 MB/Sec) (Default Value)
01 = Limit at 62 ½ %
( 500 MB/Sec)
10 = Limit at 50 %
( 400 MB/Sec)
11 = Limit at 37 ½ %
( 300 MB/Sec)
Throttle Lock. Read/Write, RO if Throttle Lock (bit 3 =1). Once this bit is set, it can only be
cleared by a reset.
0 = Bits [7:3] remain writeable
1 = Block writes to bits [7:3]
2
Reserved (RO)
1
BNR Lookahead (Read/Write). This enables the HT unit to look further up the data path in order to
optimize the BNR (Block New Requests) signal in order to increase our effective IOQ (In Order Queue)
depth.
0 = Normal Behavior (default)
1 = BNR Lookahead Enable
62
Datasheet
®
Intel 82815EM GMCH
R
Bit
Description
0
Graphics Translation Window Size Select (GTWSS). (Read/Write) In GFX mode this would be the
size of the GTT (Graphics Translation Table). Not a valid bit in AGP mode.
0 = 64 MB (default)
1 = 32 MB.
3.6.24.
CAPID—Capability Identification (Device 0 - AGP MODE ONLY)
Address Offset:
Default Value:
Access:
Size:
88-8Bh
1 F205 A009h – 1 1205 0009h
Read Only
64 bits (40 bits implemented)
31
30
29
28
27
Reserved
Display
Cache
Capability
AGP
Capability
Internal
Graphics
Capability
24
CAPID Version
23
16
CAPID Length
15
8
Next Capability Pointer
7
0
CAP_ID
Bit
Description
39-33
Reserved
32
Mobile Capable
0 = Not mobile capable
1= Mobile capable (bit hard wired internally)
31
Reserved
Default = 0.
30
Display Cache Capability (R/O)
0 = Only supports UMA mode (no Local Memory).
1 = Component is Local Memory (Display Cache) and UMA capable.
29
AGP Capability (R/O)
0 = AGP mode not supported.
1 = AGP mode supported.
Datasheet
63
®
Intel 82815EM GMCH
R
Bit
28
Description
Internal Graphics Capability (R/O)
0 = Internal graphic controller not supported.
1 = Internal graphic controller supported.
27-24
CAPID Version (R/O)
This field has the value 0010b to identify the first revision of the CAPID register definition.
23-16
CAPID Length (R/O)
This field has the value 05h to indicate the structure length.
15-8
Next Capability Pointer (R/O) This field has two possible values base on (APCONT[0] at offset 51h):
A0h when APCONT[0] = 0 (AGP Mode) meaning the next capability pointer is ACAPID.
00h when APCONT[0] = 1 (GFX Mode) meaning that this was the last capability pointer in the list.
7-0
CAP_ID (R/O)
This field has the value 1001b to identify the CAP_ID assigned by the PCI SIG for vendor dependent
capability pointers.
64
Datasheet
®
Intel 82815EM GMCH
R
3.6.25.
BUFF_SC—System Memory Buffer Strength Control Register
(Device 0)
Address Offset:
Default Value:
Access:
Size:
92-93h
FFFFh
Read/Write
16 bits
This register programs the system memory DRAM interface signal buffer strengths, with the exception of
the CKEs. The programming of these bits should be based on DRAM density (x8 or x16), DRAM
technology (16Mb, 64Mb, 128Mb or 256 Mb), rows populated, etc. Note that x4 & x32 DRAMs are not
supported. Registered DIMMs and DIMMS with ECC are also not supported and BIOS upon detection
of ECC via SPD, should report to the user that ECC DIMM timings are not supported by the GMCH2-M.
Note SO-DIMM only support x16 DRAM density.
In the descriptions below, the term “Row” is equivalent to one side of one DIMM. In other words, a
“single-sided” DIMM contains one populated row (always an odd numbered), and one empty row (even
numbered). A “double-sided” DIMM contains two populated rows.
All buffer strengths are based on the number of “loads” connected to each pin of a given signal group. A
“load” represents one pin of one SDRAM Device. The GMCH2-M pin is implied and not counted in the
load equations. The number of loads on a given signal for a given configuration can be determined
entirely from the width of the SDRAM devices that populate each row in the configuration. This
information is readily available for each row via the Serial Presence Detect mechanism.
15
14
13
12
11
10
SCS[5]#
Buffer
Strength
SCS[4]#
Buffer
Strength
SCS[3]#
Buffer
Strength
SCS[2]#
Buffer
Strength
SCS[1]#
Buffer
Strength
SCS[0]#
Buffer
Strength
7
6
5
SMAB[7:4]# Buffer
Strength
4
SMAA[7:4] Buffer
Strength
Bit
15
3
9
2
SMD and SDQM Buffer
Strengths
8
SMAC[7:4]# Buffer
Strength
1
0
Control Buffer Strengths
Description
SCS[5]# Buffer Strength (Row 5).
0 = 1.7x 4 loads
1 = 1.0x 0 or 2 loads
Each Row is actually selected by a pair of Chip Select signals (SCSA[n]# and SCSB[n]#).
The number of SCS# loads for a given Row can be determined from SPD data using the following
equation:
Loads = 32 / (width of SDRAM devices in row)
14
SCS[4]# Buffer Strength (Row 4).
0 = 1.7x 4 loads
1 = 1.0x 0 or 2 loads
Datasheet
65
®
Intel 82815EM GMCH
R
Bit
13
Description
SCS[3]# Buffer Strength (Row 3).
0 = 1.7x 4 loads
1 = 1.0x 0 or 2 loads
12
SCS[2]# Buffer Strength (Row 2).
0 = 1.7x 4 loads
1 = 1.0x 0 or 2 loads
11
SCS[1]# Buffer Strength (Row 1).
0 = 1.7x 4 loads
1 = 1.0x 0 or 2 loads
10
SCS[0]# Buffer Strength (Row 0).
0 = 1.7x 4 loads
1 = 1.0x 0 or 2 loads
9:8
SMAC[7:4]# Buffer Strength (Rows 4/5).
00 = 2.7x > 8 loads
01 = 1.7x 8 loads
10 = 1.0x 0 or 4 loads
11 = 1.0x 0 or 4 loads
Separate copies of these SMA*[7:4] “Command-Per-Clock” signals are provided for each DIMM. So
the loads for each copy are determined by the number of SDRAM devices on the corresponding DIMM
(4, 8, 12, or 16 loads).
The number of loads for each SMA*[7:4] signal group can be determined from SPD data using the
following equation:
Loads = (64 / (SDRAM Device Width for 1st row)) + (64 / (SDRAM Device Width for 2nd row))
7:6
SMAB[7:4]# Buffer Strength (Rows 2/3).
00 = 2.7x 16 loads
01 = 1.7x 8 loads
10 = 1.0x 0 or 4 loads
11 = 1.0x 0 or 4 loads
5:4
SMAA[7:4] Buffer Strength (Rows 0/1).
00 = 2.7x 16 loads
01 = 1.7x 8 loads
10 = 1.0x 0 or 4 loads
11 = 1.0x 0 or 4 loads
66
Datasheet
®
Intel 82815EM GMCH
R
Bit
3:2
Description
SMD[63:0] and SDQM[7:0] Buffer Strengths (All Rows).
00 = 1.7x > 2 loads
01 = 0.7x Reserved
10 = 1.0x 0-2 loads
11 = 1.0x 0-2 loads
The load on the SMD and SDQM signals is a function only of the number of populated rows in the
system (range 1 to 6 loads):
Loads = Number of populated Rows.
1:0
SWE#, SCAS#, SRAS#, SMAA[11:8, 3:0], SBS[1:0] Control Buffer Strengths (All Rows)
00 = 1.7x > 16 loads
01 = 0.7x < 8 loads
10 = 1.0x 8-16 loads
11 = 1.0x 8-16 loads
The load on the Address and Control signals (other than SMA*[7:4] above) is simply the number of
devices populated in ALL rows (range from 4 to 48 loads!).
Loads = (64 / Row 0 Device Width) + (64 / Row 1 Device Width) + (64 / Row 2 Device Width) +
(64 / Row 3 Device Width) + (64 / Row 4 Device Width) + (64 / Row 5 Device Width)
Datasheet
67
®
Intel 82815EM GMCH
R
3.6.26.
BUFF_SC2-System Memory Buffer Strength Control Register 2
(Device 0)
Address Offset:
Default Value:
Access:
Size:
94-95h
FFFFh
Read/Write
16 bits
This register programs the system memory DRAM interface CKE signal buffer strengths. See BUFF_SC
register for remainder of buffer strength controls.
15
8
Reserved (R/W)
7
6
Reserved (R/W)
5
4
3
2
1
0
SCKE5
Buffer
Strength
SCKE4
Buffer
Strength
SCKE3
Buffer
Strength
SCKE2
Buffer
Strength
SCKE1
Buffer
Strength
SCKE0
Buffer
Strength
Bit
Description
15:6
Reserved. Generic Read/Write bits with flops for future use.
5
SCKE[5] Buffer Strength (Row 5).
0 = 2.7x 8 loads
1 = 1.7x 0 or 4 loads
The load on a given SCKE signal is equal to the number of SDRAM devices for that particular Row
(either 4 or 8 loads).
Loads = (64 / SDRAM Device Width for this row)
4
SCKE[4] Buffer Strength (Row 4).
0 = 2.7x 8 loads
1 = 1.7x 0 or 4 loads
3
SCKE[3] Buffer Strength (Row 3).
0 = 2.7x 8 loads
1 = 1.7x 0 or 4 loads
2
SCKE[2] Buffer Strength (Row 2).
0 = 2.7x 8 loads
1 = 1.7x 0 or 4 loads
1
SCKE[1] Buffer Strength (Row 1).
0 = 2.7x 8 loads
1 = 1.7x 0 or 4 loads
68
Datasheet
®
Intel 82815EM GMCH
R
Bit
0
Description
SCKE[0] Buffer Strength (Row 0).
0 = 2.7x 8 loads
1 = 1.7x 0 or 4 loads
3.6.27.
ACAPID—AGP Capability Identifier Register (Device 0)
Address Offset:
Default Value:
Access:
Size:
A0-A3h
00200002h
Read Only
32 bits
This register provides standard identifier for AGP capability. (AGP MODE ONLY)
31
24
Reserved
23
20
19
Major AGP Revision Number
16
Minor AGP Revision Number
15
8
Next Capability Pointer
7
0
AGP Capability ID
Bit
Datasheet
Description
31:24
Reserved
23:20
Major AGP Revision Number: These bits provide a major revision number of AGP specification to
which this version of GMCH2-M conforms. These bits are set to the value 0010b to indicate AGP
Rev. 2.x.
19:16
Minor AGP Revision Number: These bits provide a minor revision number of AGP specification to
which this version of GMCH2-M conforms. This number is hardwired to value of “0000” (i.e. implying
Rev x.0)
Together with major revision number this field identifies
GMCH2-M as an AGP REV 2.0 compliant device.
15:8
Next Capability Pointer: AGP capability is the first and the last capability described via the
capability pointer mechanism and therefore these bits are hardwired to “0” to indicate the end of the
capability linked list.
7:0
AGP Capability ID: This field identifies the linked list item as containing AGP registers.
This field has the value 0000_0010b as assigned by the PCI SIG.
69
®
Intel 82815EM GMCH
R
3.6.28.
AGPSTAT—AGP Status Register (Device 0)
Address Offset:
Default Value:
Access:
Size:
A4-A7h
1F000207h
Read Only
32 bits
This register reports AGP device capability/status. (AGP MODE ONLY)
31
24
Request Queue (RQ) (HW=1Fh)
23
16
Reserved
15
10
Reserved
7
6
Reserved
5
4
3
>4 GB
Support
(HW=0)
Fast
Writes
(HW=0)
Reserved
Bit
9
8
SBA
(HW=1)
Reserved
2
0
Data Transfer Rate
(HW=111)
Description
31:24
RQ. This field contains the maximum number of AGP command requests the GMCH2-M is configured
to manage. The lower 6 bits of this field reflect the value programmed in AGPCTRL[12:10]. Only
discrete values of 32, 16, 8, 4 , 2 and 1 can be selected via AGPCTRL.
Upper bits are hardwired to “0”. Default =1Fh to allow a maximum of 32 outstanding AGP command
requests.
23:10
Reserved
9
SBA. This bit indicates that the GMCH2-M supports side band addressing. It is hardwired to 1.
8:6
Reserved
5
4G. This bit indicate that the GMCH2-M does not support addresses greater than 4 gigabytes. It is
hardwired to 0.
4
FW.This bit indicate that the GMCH2-M does not support fast writes from the processor-to-AGP
master. It is hardwired to 0.
3
Reserved
2:0
Rate. After reset the GMCH2-M reports its data transfer rate capability.
Default Value = 111.
Note that the selected data transfer mode applies to both AD bus and SBA bus.
Bit 0 (high) = AGP device supports 1x data transfer mode
Bit 1 (high) = AGP device supports 2x data transfer mode
Bit 2 (high) = AGP device supports 4x data transfer mode. This bit can be masked by the AGPCTRL
register bit 0 (See PCI Configuration Space Table 3-1).
70
Datasheet
®
Intel 82815EM GMCH
R
3.6.29.
AGPCMD—AGP Command Register (Device 0)
Address Offset:
Default Value:
Access:
Size:
A8-ABh
00000000h
Read/Write
32 bits
This register provides control of the AGP operational parameters. (AGP mode only.)
31
10
Reserved
7
6
Reserved
Bit
5
4
3
4 GB
(HW=0)
FW EN
Reserved
9
8
SBA EN
AGP EN
2
0
Data Rate
Description
31:10
Reserved.
9
SBA Enable.
1 = Side Band Addressing mechanism is enabled.
0= Side Band Addressing mechanism is disabled
8
AGP Enable. Any AGP operations received while this bit is set to 1 will be serviced even if this bit is
reset to 0. If this bit transitions from a 1 to a 0 on a clock edge in the middle of an SBA command
being delivered in 1X mode the command will be issued.
0 = GMCH2-M will ignore all AGP operations, including the sync cycle.
1 = GMCH2-M will respond to AGP operations delivered via PIPE#, or to operations delivered via SBA
if the AGP Side Band Enable bit is also set to 1.
7:6
Reserved.
5
4G. The GMCH2-M as an AGP target does not support addressing greater than 4 gigabytes. This bit is
hardwired to 0.
4
FW Enable. This bit must always be programmed to ‘0’. The chip-set will behave unpredictably if this
bit is programmed with ‘1’.
3
Reserved.
2:0
Data Rate: The settings of these bits determines the AGP data transfer rate. One (and only one) bit in
this field must be set to indicate the desired data transfer rate. The same bit must be set on both
master and target.
Bit 0: 1X,
Bit 1: 2X,
Bit 2: 4x.
Configuration software will update this field by setting only one bit that corresponds to the capability of
AGP master (after that capability has been verified by accessing the same functional register within the
AGP master’s configuration space.)
Bit 2 becomes reserved (but will still read 4x, erroneously) when the 4x Override bit in the AGPCTRL
register is set to 1 because this bit will not be updated in 4x Override mode. When the 4x Override bit
is set writes to Data Rate[2] have no functional impact.
Note: This field applies to AD and SBA buses.
Datasheet
71
®
Intel 82815EM GMCH
R
3.6.30.
AGPCTRL—AGP Control Register (Device 0)
Address Offset:
Default Value:
Access:
Size:
B0-B3h
00000000h
Read/Write
32 bits
This register provides for additional control of the AGP interface. (AGP mode only.)
31
8
Reserved
Bit
7
GTLB_EN
6
0
Reserved
Description
31:8
Reserved
7
GTLB Enable (and GTLB Flush Control) (Read/Write):
1 = enables normal operations of the Graphics Translation Lookaside Buffer.
0 = the GTLB is flushed by clearing the valid bits associated with each entry. In this mode of operation
all accesses that require translation bypass the GTLB. All requests that are positively decoded to the
graphics aperture force the GMCH2-M to access the translation table in main memory before
completing the request. Translation table entry fetches will not be cached in the GTLB. (default)
When an invalid translation table entry is read, this entry will still be cached in the GTLB (ejecting the
least recently used entry). The GMCH2-M will flush the GWB whenever software sets or clears this bit
to ensure coherency between the GTLB and main memory.
NOTE: This bit can be changed dynamically (i.e. while an access to GTLB occurs).
6:0
72
Reserved
Datasheet
®
Intel 82815EM GMCH
R
3.6.31.
APSIZE—Aperture Size (Device 0)
Address Offset:
Default Value:
Access:
Size:
B4h
00h
Read/Write
8 bits
This register determines the effective size of the Graphics Aperture used for a particular GMCH2-M
configuration. This register can be updated by the GMCH2-M-specific BIOS configuration sequence
before the PCI standard bus enumeration sequence takes place. If the register is not updated then a
default value will select an aperture of maximum size (i.e. 64 MB). (AGP mode only.)
7
4
Reserved
Bit
3
GFX
Aperture
Size
2
0
Reserved
Description
7:4
Reserved.
3
Graphics Aperture Size. Bit 3 operates on bit 25 of the Aperture Base (APBASE) configuration
register. When this bit is a “0” it forces bit 25 in APBASE to behave as “hardwired” to 0. When this bit is
a “1” it forces bit 25 in APBASE to be Read/Write accessible. Only the following combinations are
allowed:
0 = 64 MB Aperture Size
1 = 32 MB Aperture Size
Default for APSIZE[3]=0b forces default APBASE[25] =0b (responds as “hardwired” to 0). This provides
maximum aperture size of 64 MB. Programming APSIZE[3]=1b enables APBASE[25] as Read/Write
programmable.
2:0
Datasheet
Reserved.
73
®
Intel 82815EM GMCH
R
3.6.32.
ATTBASE-Aperture Translation Table Base Register (Device 0)
Address Offset:
Default Value:
Access:
Size:
B8-BBh
00000000h
Read/Write
32 bits
This register provides the starting address of the Graphics Aperture Translation Table Base located in the
main DRAM. This value is used by the GMCH2-M’s Graphics Aperture address translation logic
(including the GTLB logic) to obtain the appropriate address translation entry required during the
translation of the aperture address into a corresponding physical DRAM address. The ATTBASE register
may be dynamically changed. (AGP mode only.)
Note:
The address provided via ATTBASE is 4KB aligned.
31
29
Reserved
Bit
74
28
12
ATT Base Address
11
0
Reserved
Description
31:29
Reserved.
28:12
ATT Base Address. This field contains a pointer to the base of the translation table used to map
memory space addresses in the aperture range to addresses in main memory.
11:0
Reserved.
Datasheet
®
Intel 82815EM GMCH
R
3.6.33.
AMTT—AGP Multi-Transaction Timer (Device 0)
Address Offset:
Default Value:
Access:
Size:
BCh
00h
Read/Write
8 bits
AMTT is an 8-bit register that controls the amount of time that the GMCH2-M’s arbiter allows AGP/PCI
master to perform multiple back-to-back transactions. The GMCH2-M’s AMTT mechanism is used to
optimize the performance of the AGP master (using PCI semantics) that performs multiple back-to-back
transactions to fragmented memory ranges (and as a consequence it can not use long burst transfers).
The AMTT mechanism applies to the processor-AGP/PCI transactions as well and it guarantees to the
processor a fair share of the AGP/PCI interface bandwidth.
The number of clocks programmed in the AMTT represents the guaranteed time slice (measured in
66MHz clocks) allotted to the current agent (either AGP/PCI master or Host bridge) after which the AGP
arbiter will grant the bus to another agent. The default value of AMTT is 00h and disables this function.
The AMTT value can be programmed with 8 clock granularity. For example, if the AMTT is
programmed to 18h, then the selected value corresponds to the time period of 24 AGP (66-MHz) clocks.
(AGP mode only.)
7
3
Multi-Transaction Timer Count Value
Bit
Datasheet
2
0
Reserved
Description
7:3
Multi-Transaction Timer Count Value. The number programmed in these bits represents the
guaranteed time slice (measured in eight 66MHz clock granularity) allotted to the current agent (either
AGP/PCI master or Host bridge) after which the AGP arbiter will grant the bus to another agent.
2:0
Reserved.
75
®
Intel 82815EM GMCH
R
3.6.34.
LPTT—AGP Low Priority Transaction Timer Register (Device 0)
Address Offset:
Default Value:
Access:
Size:
BDh
00h
Read/Write
8 bits
LPTT is an 8-bit register similar in a function to AMTT. This register is used to control the minimum
tenure on the AGP for low priority data transaction (both reads and writes) issued using PIPE# or SB
mechanisms. The number of clocks programmed in the LPTT represents the guaranteed time slice
(measured in 66MHz clocks) allotted to the current low priority AGP transaction data transfer state. This
does not necessarily apply to a single transaction but it can span over multiple low-priority transactions
of the same type. After this time expires the AGP arbiter may grant the bus to another agent if there is a
pending request. The LPTT does not apply in the case of high-priority request where ownership is
transferred directly to high-priority requesting queue. The default value of LPTT is 00h and disables this
function. The LPTT value can be programmed with 8 clock granularity. For example, if the LPTT is
programmed to 10h, then the selected value corresponds to the time period of 16 AGP (66-MHz) clocks.
(AGP mode only.)
7
3
Low Priority Transaction Timer Count Value
Bit
76
2
0
Reserved
Description
7:3
Low Priority Transaction Timer Count Value. The number of clocks programmed in these bits
represents the guaranteed time slice (measured in eight 66MHz clock granularity) allotted to the current
low priority AGP transaction data transfer state.
2:0
Reserved.
Datasheet
®
Intel 82815EM GMCH
R
3.6.35.
GMCHCFG—GMCH2-M Configuration Register (Device 0)
Address Offset:
Default:
Access:
Size:
BEh
0000 X000b
Read/Write, Read Only
8 bits
7
6
Reserved
5
4
3
MDA
Present
(R/W)
Bit
2
0
Reserved
Description
7:6
Reserved.
5
MDA Present (MDAP) (Read/Write). This bit should not be set when the VGA Enable bit is not
set. This bit works with the VGA Enable bit in the BCTRL register (3Eh, bit 3) of device 1 to control
the routing of processor initiated transactions targeting MDA compatible I/O and memory address
ranges. If the VGA enable bit is set, then accesses to IO address range x3BCh - x3BFh are
forwarded to hub interface. If the VGA enable bit is not set then accesses to IO address range x3BCh
- x3BFh are treated just like any other IO accesses i.e. the cycles are forwarded to AGP if the
address is within IOBASE and IOLIMIT and ISA enable bit is not set. MDA resources are defined as
the following:
Memory:
0B0000h - 0B7FFFh
I/O: 3B4h, 3B5h, 3B8h, 3B9h, 3BAh, 3BFh, (including ISA address aliases, A[15:10] are not used in
decode). Any I/O reference that includes the I/O locations listed above, or their aliases, will be
forwarded to hub interface even if the reference includes I/O locations not listed above. The following
table shows the behavior for all combinations of MDA and VGA:
VGA MDA Behavior
4:0
Datasheet
0
0
All References to MDA and VGA go to hub interface
0
1
Illegal Combination (DO NOT USE)
1
0
All References to VGA go to AGP/PCI. MDA-only references
1
1
VGA References go to AGP/PCI; MDA References go to hub interface
Reserved.
77
®
Intel 82815EM GMCH
R
3.6.36.
ERRCMD—Error Command Register (Device 0)
Address Offset:
Default Value:
Access:
Size:
CBh
00h
Read/Write
8 bits
This register enables various errors to generate a SERR hub interface special cycle. Since the GMCH2-M
does not have an SERR# signal, SERR messages are passed from the GMCH2-M to the ICH2-M over
hub interface. The actual generation of the SERR message is globally enabled for Device 0 via the PCI
Command register. (AGP mode only.)
Note:
An error can generate one and only one hub interface error special cycle. It is software’s responsibility to
make sure that when an SERR error message is enabled for an error condition, SMI and SCI error
messages are disabled for that same error condition.
7
6
Reserved
5
4
3
2
1
0
SRMMRO
SRTA
SLNDM
SAAOGA
SIAA
SAIGATTE
Bit
Description
7:6
Reserved.
5
SERR on Receiving Main Memory Refresh Overrun. Identical functionality in Device 2 memory
mapped space @ 020B8h. This bit allows use of this same functionality in AGP Mode.
0 = reporting of this condition is disabled.
1 = the GMCH2-M generates an SERR hub interface special cycle when a main memory refresh
overrun occurs.
4
SERR on Receiving Target Abort on the hub interface.
0 = reporting of this condition is disabled.
1 = the GMCH2-M generates an SERR hub interface special cycle when an GMCH2-M originated hub
interface cycle is terminated with a Target Abort.
3
SERR on LOCK to non-DRAM Memory.
0 = reporting of this condition is disabled.
1= the GMCH2-M generates an SERR hub interface special cycle when a processor initiated LOCK
transaction targeting non-DRAM memory space occurs.
2
SERR on AGP Access Outside of Graphics Aperture.
0 = reporting of this condition is disabled.
1 = the GMCH2-M generates an SERR hub interface special cycle when an AGP access occurs to an
address outside of the graphics aperture.
1
SERR on Invalid AGP Access.
0 = reporting of this condition is disabled.
1 = GMCH2-M generates an SERR hub interface special cycle when an AGP access occurs to an
address outside of the graphics aperture and either to the 640k - 1M range or above the top of memory.
78
Datasheet
®
Intel 82815EM GMCH
R
Bit
0
Description
SERR on Access to Invalid Graphics Aperture Translation Table Entry.
0 = reporting of this condition via SERR messaging is disabled.When this bit is set to “
1 = the GMCH2-M generates an SERR hub interface special cycle when an invalid translation table
entry was returned in response to a AGP access to the graphics aperture.
3.7.
AGP/PCI Bridge Registers – (Device #1 - Visible in AGP
Mode Only)
These registers are accessible through the configuration mechanism defined in an earlier section of this
document.
Table 8. GMCH2-M Configuration Space (Device #1)
Address
Offset
Datasheet
Register
Symbol
Register Name
Default Value
Access Type
00-01h
VID1
Vendor Identification
8086h
RO
02-03h
DID1
Device Identification
1131h
RO
04-05h
PCICMD1
PCI Command Register
0000h
RO, Read/Write
06-07h
PCISTS1
PCI Status Register
0020h
RO, Read/WriteC
08h
RID1 (A0)
Revision Identification for A0-step
10h
RO
08h
RID1 (A1)
Revision Identification for A1-step
11h
RO
09

Reserved
00h

0Ah
SUBC1
Sub-Class Code
04h
RO
0Bh
BCC1
Base Class Code
06h
RO
0Ch

Reserved
00h

0Dh
MLT1
Master Latency Timer
00h
Read/Write
0Eh
HDR1
Header Type
01h
RO
0F-17h

Reserved
00h

18h
PBUSN
Primary Bus Number
00h
RO
19h
SBUSN
Secondary Bus Number
00h
Read/Write
1Ah
SUBUSN
Subordinate Bus Number
00h
Read/Write
1Bh
SMLT
Secondary Bus Master Latency
Timer
00h
Read/Write
1Ch
IOBASE
I/O Base Address Register
F0h
Read/Write
1Dh
IOLIMIT
I/O Limit Address Register
00h
Read/Write
1E-1Fh
SSTS
Secondary Status Register
02A0h
RO, Read/WriteC
20-21h
MBASE
Memory Base Address Register
FFF0h
Read/Write
22-23h
MLIMIT
Memory Limit Address Register
0000h
Read/Write
79
®
Intel 82815EM GMCH
R
Address
Offset
3.7.1.
Register
Symbol
Register Name
Default Value
Access Type
24-25h
PMBASE
Prefetchable Memory Base
Address
FFF0h
Read/Write
26-27h
PMLIMIT
Prefetchable Memory Limit
Address
0000h
Read/Write
28-3Dh

Reserved
00h

3Eh
BCTRL
Bridge Control Register
00h
Read/Write
3Fh

Reserved
00h

40h
ERRCMD1
Error Command
00h
Read/Write
41-FFh

Reserved
00h

VID1—Vendor Identification Register (Device 1)
Address Offset:
Default Value:
Attribute:
Size:
00 - 01h
8086h
Read Only
16 bits
The VID Register contains the vendor identification number. This 16-bit register combined with the
Device Identification Register uniquely identifies any PCI device. Writes to this register have no effect.
Bit
15:0
3.7.2.
Description
Vendor Identification Number. This is a 16-bit value assigned to Intel®. Intel® VID = 8086h.
DID1—Device Identification Register (Device 1)
Address Offset:
Default Value:
Attribute:
Size:
02 - 03h
1131h
Read Only
16 bits
This 16-bit register combined with the Vendor Identification register uniquely identifies any PCI device.
Writes to this register have no effect.
Bit
15:0
Description
Device Identification Number. This is a 16 bit value assigned to the GMCH2-M AGP interface device.
1131h = Device ID for Device 1.
80
Datasheet
®
Intel 82815EM GMCH
R
3.7.3.
PCICMD1—PCI-PCI Command Register (Device 1)
Address Offset:
Default:
Access:
Size
04-05h
0000h
Read/Write, Read Only
16 bits
15
10
Reserved (0)
9
8
FB2B
(Not Impl)
SERR En
7
6
5
4
3
2
1
0
Addr/Data
Stepping
(Not Impl)
Parity
Error En
(Not Impl)
Reserved
Mem WR
& Inval En
Special
Cycle En
Bus
Master En
Mem Acc
En
I/O Acc En
Bit
Descriptions
15:10
Reserved.
9
Fast Back-to-Back: Not Applicable-hardwired to “0”.
8
SERR Message Enable (SERRE1). This bit is a global enable bit for Device 1 SERR messaging. The
GMCH2-M does not have an SERR# signal. The GMCH2-M communicates the SERR# condition by
sending an SERR message to the ICH2-M.
0 = the SERR message is not generated by the GMCH2-M for Device 1
1 = the GMCH2-M is enabled to generate SERR messages over hub interface for specific Device 1
error conditions that are individually enabled in the ERRCMD1 and BCTRL registers. The error status is
reported in the PCISTS1 register.
NOTE: This bit only controls SERR messaging for the Device 1. Device 0 has its own SERRE bit
to control error reporting for error conditions occurring on Device 0. The two control bits are
used in a logical OR manner to enable the SERR hub interface message mechanism.
7
Address/Data Stepping: Not applicable. Hardwired to “0”.
6
Parity Error Enable (PERRE1): PERR# is not supported on AGP/PCI1. Hardwired to “0”.
5
Reserved.
4
Memory Write and Invalidate Enable: (RO) This bit is implemented as RO and returns a value of “0”
when read.
3
Special Cycle Enable: (RO) This bit is implemented as Read Only and returns a value of “0” when
read.
2
Bus Master Enable (BME1): (Read/Write)
0 =(default) AGP Master initiated FRAME# cycles will be ignored by the GMCH2-M resulting in a
Master Abort. Ignoring incoming cycles on the secondary side of the P2P bridge effectively disables the
bus master on the primary side.
1 = AGP Master initiated FRAME# cycles will be accepted by the GMCH2-M if they hit a valid address
decode range. This bit has no affect on AGP Master originated SBA or PIPE# cycles.
Datasheet
81
®
Intel 82815EM GMCH
R
Bit
Descriptions
1
Memory Access Enable (MAE1): (Read/Write)
0 = all of Device #1’s memory space is disabled.
1 = to enable the Memory and Prefetchable memory address ranges defined in the MBASE, MLIMIT,
PMBASE, and PMLIMIT registers, as well as the VGA window.
0
I/O Access Enable (IOAE1): (Read/Write)
0 = all of Device 1’s I/O space is disabled.
1 = to enable the I/O address range defined in the IOBASE, and IOLIMIT registers.
3.7.4.
PCISTS1—PCI-PCI Status Register (Device 1)
Address Offset:
Default Value:
Access:
Size:
06-07h
0020h
Read Only, Read/Write Clear
16 bits
PCISTS1 is a 16-bit status register that reports the occurrence of error conditions associated with primary
side of the “virtual” PCI-PCI bridge embedded within the GMCH2-M. Since this device does not
physically reside on PCI0 it reports the optimum operating conditions so that it does not restrict the
capability of PCI0.
15
14
13
12
11
Detected
Par Error
(HW=0)
Sig Sys
Error
Rec
Mast
Abort Sta
(HW=0)
Rec
Target
Abort Sta
(HW=0)
Sig Target
Abort Sta
(HW=0)
7
6
5
4
FB2B
(HW=0)
Reserved
66/60 MHz
Cap
(HW=0)
Bit
82
10
9
DEVSEL# Timing
(HW=00)
3
8
Data Par
Detected
(HW=0)
0
Reserved
Descriptions
15
Detected Parity Error (DPE1): Not Applicable - hardwired to “0”.
14
Signaled System Error (SSE1). This bit is set to 1 when Device 1 generates an SERR message over
hub interface for any enabled Device 1 error condition. Device 1 error conditions are enabled in the
PCICMD1, ERRCMD1 and BCTRL registers. Device 1 error flags are read/reset from the SSTS
register. Software clears this bit by writing a 1 to it.
13
Received Master Abort Status (RMAS1): Not Applicable – hardwired to “0”.
12
Received Target Abort Status (RTAS1): Not Applicable – hardwired to “0”.
11
Signaled Target Abort Status (STAS1): Not Applicable – hardwired to “0”.
10:9
DEVSEL# Timing (DEVT1): Not Applicable - hardwired to “00b”.
8
Data Parity Detected (DPD1): Not Applicable - hardwired to “0”.
7
Fast Back-to-Back (FB2B1): Not Applicable - hardwired to “0”.
Datasheet
®
Intel 82815EM GMCH
R
Bit
3.7.5.
Descriptions
6
Reserved.
5
66/60MHz Capability: Not Applicable - Hardwired to “1”.
4:0
Reserved.
RID1—Revision Identification Register (Device 1)
Address Offset:
Default Value:
Access:
Size:
08h
10h for A0-step silicon
11h for A1-step silicon
Read Only
8 bits
This register contains the revision number of the GMCH2-M Device 1. These bits are read only and
writes to this register have no effect. For the A-0 Stepping, this value is 10h.
Bit
7:0
3.7.6.
Description
Revision Identification Number. This is an 8-bit value that indicates the revision identification number
for the GMCH2-M Device 1. Default = 10h
SUBC1—Sub-Class Code Register (Device 1)
Address Offset:
Default Value:
Access:
Size:
0Ah
04h
Read Only
8 bits
This register contains the Sub-Class Code for the GMCH2-M Device 1. This code is 04h indicating a
PCI-PCI Bridge device. The register is read only.
Bit
7:0
Datasheet
Description
Sub-Class Code (SUBC1). This is an 8-bit value that indicates the category of Bridge into which the
GMCH2-M falls. The code is 04h indicating a Host Bridge.
83
®
Intel 82815EM GMCH
R
3.7.7.
BCC1—Base Class Code Register (Device 1)
Address Offset:
Default Value:
Access:
Size:
0Bh
06h
Read Only
8 bits
This register contains the Base Class Code of the GMCH2-M Device 1. This code is 06h indicating a
Bridge device. This register is read only.
Bit
Description
7:0
3.7.8.
Base Class Code (BASEC). This is an 8-bit value that indicates the Base Class Code for the GMCH2M Device 1. This code has the value 06h, indicating a Bridge device.
MLT1—Master Latency Timer Register (Device 1)
Address Offset:
Default Value:
Access:
Size:
0Dh
00h
Read/Write
8 bits
This functionality is not applicable. It is described here since these bits should be implemented as a
Read/Write to prevent standard PCI-PCI bridge configuration software from getting “confused”.
7
6
5
4
3
2
Not Applicable
Bit
7:3
0
Reserved
Description
Not applicable but support Read/Write operations.
(Reads return previously written data.)
2:0
3.7.9.
Reserved.
HDR1—Header Type Register (Device 1)
Address Offset:
Default:
Access:
Size:
0Eh
01h
Read Only
8 bits
This register identifies the header layout of the configuration space. No physical register exists at this
location.
Bit
7:0
84
Descriptions
This read only field always returns 01h when read. Writes have no effect.
Datasheet
®
Intel 82815EM GMCH
R
3.7.10.
PBUSN—Primary Bus Number Register (Device 1)
Address Offset:
Default:
Access:
Size:
18h
00h
Read Only
8 bits
This register identifies that “virtual” PCI-PCI bridge is connected to bus #0.
Bit
7:0
3.7.11.
Descriptions
Bus Number. Hardwired to “0”.
SBUSN—Secondary Bus Number Register (Device 1)
Address Offset:
Default:
Access:
Size:
19h
00h
Read/Write
8 bits
This register identifies the bus number assigned to the second bus side of the “virtual” PCI-PCI bridge
i.e. to PCI1/AGP. This number is programmed by the PCI configuration software to allow mapping of
configuration cycles to PCI1/AGP.
Bit
7:0
3.7.12.
Descriptions
Bus Number. Programmable
SUBUSN—Subordinate Bus Number Register (Device 1)
Address Offset:
Default:
Access:
Size:
1Ah
00h
Read/Write
8 bits
This register identifies the subordinate bus (if any) that resides at the level below PCI1/AGP. This
number is programmed by the PCI configuration software to allow mapping of configuration cycles to
PCI1/AGP.
Bit
7:0
Datasheet
Descriptions
Bus Number. Programmable
85
®
Intel 82815EM GMCH
R
3.7.13.
SMLT—Secondary Master Latency Timer Register (Device 1)
Address Offset:
Default Value:
Access:
Size:
1Bh
00h
Read/Write
8 bits
This register controls the bus tenure of the GMCH2-M on AGP/PCI. SMLT is an 8-bit register that
controls the amount of time the GMCH2-M as an AGP/PCI bus master, can burst data on the AGP/PCI
Bus. The Count Value is an 8 bit quantity, however SMLT[2:0] are reserved and assumed to be 0 when
determining the Count Value. The GMCH2-M’s SMLT is used to guarantee to the AGP master a
minimum amount of the system resources. When the GMCH2-M begins the first PCI bus cycle after
being granted the bus, the counter is loaded and enabled to count from the assertion of FRAME#. If the
count expires while the GMCH2-M’s grant is removed (due to AGP master request), then the GMCH2M will lose the use of the bus, and the AGP master agent may be granted the bus. If GMCH2-M’s bus
grant is not removed, the GMCH2-M will continue to own the AGP/PCI bus regardless of the SMLT
expiration or idle condition. Note that the GMCH2-M must always properly terminate an AGP/PCI
transaction with FRAME# negation prior to the final data transfer.
The number of clocks programmed in the SMLT represents the guaranteed time slice (measured in
66Mhz PCI clocks) allotted to the GMCH2-M, after which it must complete the current data transfer
phase and then surrender the bus as soon as its bus grant is removed. For example, if the SMLT is
programmed to 18h, then the value is 24 AGP clocks. The default value of SMLT is 00h and disables
this function. When the SMLT is disabled, the burst time for the GMCH2-M is unlimited (i.e. the
GMCH2-M can burst forever).
7
3
Secondary MLT Counter Value
Bit
7:3
2
0
Reserved
Description
Secondary MLT Counter Value.
Default=0 i.e. SMLT disabled
2:0
86
Reserved.
Datasheet
®
Intel 82815EM GMCH
R
3.7.14.
IOBASE—I/O Base Address Register (Device 1)
Address Offset:
Default Value:
Access:
Size:
1Ch
F0h
Read/Write
8 bits
This register control the processor to PCI1/AGP I/O access routing based on the following formula:
IO_BASE=< address =<IO_LIMIT
Only upper 4 bits are programmable. For the purpose of address decode, address bits A[11:0] are treated
as 0. Thus the bottom of the defined I/O address range will be aligned to a 4KB boundary.
Note:
BIOS must not set this register to 00h otherwise 0CF8h/0CFCh accesses will be forwarded to AGP
7
4
3
I/O Address Base
Bit
7:4
0
I/O Addressing Capability
Description
I/O Address Base. Corresponds to A[15:12] of the I/O address.
Default=Fh
3:0
Datasheet
I/O Addressing Capability. Hardwired to 0h indicating that only 16 bit I/O addressing is supported. Bits
[31:16] of the I/O base address is assumed to be 0000h.
87
®
Intel 82815EM GMCH
R
3.7.15.
IOLIMIT—I/O Limit Address Register (Device 1)
Address Offset:
Default Value:
Access:
Size:
1Dh
00h
Read/Write
8 bits
This register controls the processor to PCI1/AGP I/O access routing based on the following formula:
IO_BASE=< address =<IO_LIMIT
Only upper 4 bits are programmable. For the purpose of address decode address bits A[11:0] are
assumed to be FFFh. Thus, the top of the defined I/O address range will be at the top of a 4KB aligned
address block.
7
4
3
I/O Address Limit
Bit
7:4
0
Reserved
Description
I/O Address Limit. Corresponds to A[15:12] of the I/O address.
Default=0
3:0
88
Reserved. (Only 16 bit addressing supported.)
Datasheet
®
Intel 82815EM GMCH
R
3.7.16.
SSTS—Secondary PCI-PCI Status Register (Device 1)
Address Offset:
Default Value:
Access:
Size:
1E-1Fh
02A0h
Read Only, Read/Write Clear
16 bits
SSTS is a 16-bit status register that reports the occurrence of error conditions associated with secondary
side (i.e. PCI1/AGP side) of the “virtual” PCI-PCI bridge embedded within GMCH2-M.
15
14
13
12
11
Det. Parity
Error
Rec Sys
Error
(HW=0)
Rec
Master
Abort
Rec
Target
Abort
Sig Target
Abort
(HW=0)
7
6
5
FB2B
(HW=1)
Reserved
66/60 MHz
Cap
(HW=1)
9
DEVSEL Timing
(HW=01b; medium)
8
Data
Parity Det.
(HW=0)
4
Bit
15
10
0
Reserved
Descriptions
Detected Parity Error (DPE1).
This bit is set to a 1 to indicate GMCH2-M’s detection of a parity error in the address or data phase of
PCI1/AGP bus transactions.
Software sets DPE1 to 0 by writing a 1 to this bit.
Note that the function of this bit is not affected by the PERRE1 bit. Also note that PERR# is not
implemented in the GMCH2-M.
14
Received System Error (SSE1). This bit is hardwired to 0 since the GMCH2-M does not have an
SERR# signal pin.
13
Received Master Abort Status (RMAS1).
This bit is set to 1 when the GMCH2-M terminates a Host-to-PCI1/AGP with an unexpected master
abort.
Software resets this bit to 0 by writing a 1 to it.
12
Received Target Abort Status (RTAS1).
When a GMCH2-M initiated transaction on PCI1/AGP is terminated with a target abort, RTAS1 is set to
1.
Software resets RTAS1 to 0 by writing a 1 to it.
Datasheet
11
Signaled Target Abort Status (STAS1). STAS1 is hardwired to a 0, since the GMCH2-M does not
generate target abort on PCI1/AGP.
10:9
DEVSEL# Timing (DEVT1). This 2-bit field indicates the timing of the DEVSEL# signal when the
GMCH2-M responds as a target on PCI1/AGP, and is hard-wired to the value 01b (medium) to indicate
the time when a valid DEVSEL# can be sampled by the initiator of the PCI cycle.
8
Data Parity Detected (DPD1). Hardwired to 0. GMCH2-M does not implement G_PERR# function.
However, data parity errors are still detected and reported using SERR hub interface special cycles (if
enabled by SERRE1 and the BCTRL register, bit 0).
89
®
Intel 82815EM GMCH
R
Bit
3.7.17.
Descriptions
7
Fast Back-to-Back (FB2B1). This bit is hardwired to 1, since GMCH2-M as a target supports fast
back-to-back transactions on PCI1/AGP.
6
Reserved.
5
66/60MHZ Capability: Hardwired to “1”.
4:0
Reserved.
MBASE—Memory Base Address Register (Device 1)
Address Offset:
Default Value:
Access:
Size:
20-21h
FFF0h
Read/Write
16 bits
This register controls the processor to PCI1 non-prefetchable memory access routing based on the
following formula:
MEMORY_BASE=< address =<MEMORY_LIMIT
The upper 12 bits of the register are Read/Write and correspond to the upper 12 address bits A[31:20] of
the 32 bit address. The bottom 4 bits of this register are read-only and return zeroes when read. The
configuration software must initialize this register. For the purpose of address decode address bits
A[19:0] are assumed to be 0. Thus, the bottom of the defined memory address range will be aligned to a
1MB boundary.
15
4
Memory Address Base
Bit
90
3
0
Reserved
Description
15:4
Memory Address Base (MEM_BASE). Corresponds to A[31:20] of the memory address.
3:0
Reserved.
Datasheet
®
Intel 82815EM GMCH
R
3.7.18.
MLIMIT—Memory Limit Address Register (Device 1)
Address Offset:
Default Value:
Access:
Size:
22-23h
0000h
Read/Write
16 bits
This register controls the processor to PCI1 non-prefetchable memory access routing based on the
following formula:
MEMORY_BASE=< address =<MEMORY_LIMIT
The upper 12 bits of the register are Read/Write and correspond to the upper 12 address bits A[31:20] of
the 32 bit address. The bottom 4 bits of this register are read-only and return zeroes when read. The
configuration software must initialize this register. For the purpose of address decode, address bits
A[19:0] are assumed to be FFFFFh. Thus, the top of the defined memory address range will be at the top
of a 1MB aligned memory block.
15
4
Memory Address Limit
Bit
15: 4
3
0
Reserved
Description
Memory Address Limit (MEM_LIMIT). Corresponds to A[31:20] of the memory address.
Default=0
3:0
Reserved.
Memory range covered by MBASE and MLIMIT registers are used to map non-prefetchable PCI1/AGP
address ranges (typically where control/status memory-mapped I/O data structures of the graphics
controller will reside) and PMBASE and PMLIMIT are used to map prefetchable address ranges
(typically graphics local memory). This segregation allows application of USWC space attribute to be
performed in a true plug-and-play manner to the prefetchable address range for improved processor-AGP
memory access performance.
Configuration software is responsible for programming all address range registers (prefetchable, nonprefetchable) with the values that provide exclusive address ranges i.e. prevent overlap with each other
and/or with the ranges covered with the main memory. There is no provision in the GMCH2-M hardware
to enforce prevention of overlap and operations of the system in the case of overlap are not guaranteed.
Datasheet
91
®
Intel 82815EM GMCH
R
3.7.19.
PMBASE—Prefetchable Memory Base Address Register
(Device 1)
Address Offset:
Default Value:
Access:
Size:
24-25h
FFF0h
Read/Write
16 bits
This register controls the processor to PCI1 prefetchable memory accesses routing based on the
following formula:
PREFETCHABLE_MEMORY_BASE=< address =<PREFETCHABLE_MEMORY_LIMIT
The upper 12 bits of the register are Read/Write and correspond to the upper 12 address bits A[31:20] of
the 32 bit address. The bottom 4 bits of this register are read-only and return zeroes when read. The
configuration software must initialize this register. For the purpose of address decode, address bits
A[19:0] are assumed to be 0. Thus, the bottom of the defined memory address range will be aligned to a
1MB boundary.
15
4
Prefetchable Memory Address Base
Bit
15: 4
3
0
Reserved
Description
Prefetchable Memory Address Base (PMEM_BASE). Corresponds to A[31:20] of the memory
address.
Default=FFF
3:0
92
Reserved.
Datasheet
®
Intel 82815EM GMCH
R
3.7.20.
PMLIMIT—Prefetchable Memory Limit Address Register
(Device 1)
Address Offset:
Default Value:
Access:
Size:
26-27h
0000h
Read/Write
16 bits
This register controls the processor to PCI1 prefetchable memory accesses routing based on the
following formula:
PREFETCHABLE_MEMORY_BASE=< address =<PREFETCHABLE_MEMORY_LIMIT
The upper 12 bits of the register are Read/Write and correspond to the upper 12 address bits A[31:20] of
the 32 bit address. The bottom 4 bits of this register are read-only and return zeroes when read. The
configuration software must initialize this register. For the purpose of address decode, address bits
A[19:0] are assumed to be FFFFFh. Thus, the top of the defined memory address range will be at the top
of a 1MB aligned memory block.
15
4
Prefetchable Memory Address Limit
Bit
15: 4
3
0
Reserved
Description
Prefetchable Memory Address Limit (PMEM_LIMIT).Corresponds to A[31:20] of the memory address.
Default=0
3:0
Note:
Datasheet
Reserved.
That prefetchable memory range is supported to allow segregation by the configuration software between
the memory ranges that must be defined as UC and the ones that can be designated as a USWC (i.e.
prefetchable) from the processor perspective.
93
®
Intel 82815EM GMCH
R
3.7.21.
BCTRL—PCI-PCI Bridge Control Register (Device 1)
Address Offset:
Default:
Access:
Size
3Eh
00h
Read/Write
8 bits
This register provides extensions to the PCICMD1 register that are specific to PCI-PCI bridges. The
BCTRL provides additional control for the secondary interface (PCI1/AGP) as well as some bits that
affect the overall behavior of the “virtual” PCI-PCI bridge embedded within GMCH2-M, e.g. VGA
compatible address ranges mapping.
7
6
5
4
3
2
1
0
FB2B EN
(HW=0)
Sec Bus
Reset
(HW=0)
Master
Abort
Mode
(HW=0)
Reserved
VGA EN
ISA
Enable
SERR#
EN
(HW=0)
Parity Err
Response
EN
Bit
Descriptions
7
Fast Back to Back Enable. Since there is only one target allowed on AGP this bit is meaningless.
This bit is hardwired to “0”.
6
Secondary Bus Reset: GMCH2-M does not support generation of reset via this bit on the AGP and
therefore this bit is hardwired to “0”.
Note that the only way to perform a hard reset of the AGP is via the system reset either initiated
by software or hardware via ICH.
94
5
Master Abort Mode: This bit is hardwired to “0”. This means when acting as a master on AGP/PCI1
the GMCH2-M will drop writes on the “floor” and return all “1” during reads when a Master Abort occurs.
4
Reserved.
Datasheet
®
Intel 82815EM GMCH
R
Bit
3
Descriptions
VGA Enable. This bit works with the MDA present bit , GMCHCFG[3], of device 0 to control the
routing of processor initiated transactions targeting VGA compatible I/O and memory address ranges.
When this bit is set , the GMCH2-M will forward the following processor accesses to the AGP:
1) memory accesses in the range 0A0000h to 0BFFFFh
2) I/O addresses where A[9:0] are in the ranges 3B0h to 3BBh and 3C0h to 3DFh
(inclusive of ISA address aliases - A[15:10] are not decoded)
When this bit is set , forwarding of these accesses issued by the processor is independent of the I/O
address and memory address ranges defined by the previously defined base and limit registers.
Forwarding of these accesses is also independent of the settings of the bit 2 (ISA Enable) of this
register if this bit is “1”. If the VGA enable bit is set, then accesses to IO address range x3BCh - x3BFh
are forwarded to hub interface. If the VGA enable bit is not set then accesses to IO address range
x3BCh - x3BFh are treated just like any other IO accesses i.e. the cycles are forwarded to AGP if the
address is within IOBASE and IOLIMIT and ISA enable bit is not set, otherwise they are forwarded to
hub interface.
If this bit is “0” (default) , then VGA compatible memory and I/O range accesses are not forwarded to
AGP but rather they are mapped to primary PCI unless they are mapped to AGP via I/O and memory
range registers defined above (IOBASE, IOLIMIT, MBASE, MLIMIT, PMBASE, PMLIMIT)
The following table shows the behavior for all combinations of MDA and VGA:
VGA
MDA
Behavior
0
0
All References to MDA and VGA Go To hub interface
0
1
Illegal Combination (DO NOT USE)
1
0
All References To VGA Go To AGP MDA-only references (I/O
Address 3BF and aliases) will go to hub interface.
1
Datasheet
1
VGA References go to AGP/PCI; MDA References go to the hub interface
2
ISA Enable: Modifies the response by the GMCH2-M to an I/O access issued by the processor that
target ISA I/O addresses. This applies only to I/O addresses that are enabled by the IOBASE and
IOLIMIT registers. When this bit is set to 1 GMCH2-M will not forward to PCI1/AGP any I/O
transactions addressing the last 768 bytes in each 1KB block even if the addresses are within the
range defined by the IOBASE and IOLIMIT registers. Instead of going to PCI1/AGP these cycles will be
forwarded to the hub interface where they can eventually be subtractively or positively claimed by the
ISA bridge. If this bit is “0” (default) then all addresses defined by the IOBASE and IOLIMIT for
processor I/O transactions will be mapped to PCI1/AGP.
1
SERR# Enable. This bit normally controls forwarding SERR# on the secondary interface to the
primary interface. The GMCH2-M does not support the SERR# signal on the AGP/PCI1 bus. This bit is
hardwired to a “0”.
0
Parity Error Response Enable: Controls GMCH2-M’s response to data phase parity errors on
PCI1/AGP. G_PERR# is not implemented by the GMCH2-M. However, when this bit is set to 1,
address and data parity errors on PCI1 are reported via SERR messaging, if enabled by SERRE1. If
this bit is reset to 0, then address and data parity errors on PCI1/AGP are not reported via SERR
messaging. Other types of error conditions can still be signaled via SERR messaging independent of
this bit’s state.
95
®
Intel 82815EM GMCH
R
3.7.22.
ERRCMD1—Error Command Register (Device 1)
Address Offset:
Default:
Access:
Size
40h
00h
Read/Write
8 bits
7
1
0
Reserved
Bit
3.8.
SERR on
Rec Targ.
Abort on
AGP/PCI
Descriptions
7:1
Reserved.
0
SERR on Receiving Target Abort on AGP/PCI. When this bit is set to “1” the GMCH2-M generates
an SERR hub interface special cycle when an GMCH2-M originated AGP/PCI cycle is terminated with
a Target Abort. If this bit is “0” then reporting of this condition is disabled.
Graphics Device Registers (Device 2 - VISIBLE IN GFX
MODE ONLY)
These registers are accessible through the configuration mechanism defined in an earlier section of this
document.
Table 9. Device 2 Configuration Space Address Map (Internal Graphics)
Address
Offset
96
Register
Symbol
Register Name
Default Value
Access Type
00-01h
VID2
Vendor Identification
8086h
RO
02-03h
DID2
Device Identification
1112h
RO
04-05h
PCICMD2
PCI Command Register
0004h
Read/Write
06-07h
PCISTS2
PCI Status Register
02B0h
RO
08h
RID2 (A0)
Revision Identification for A0-step
10h
RO
08h
RID2 (A1)
Revision Identification for A1-step
11h
RO
09h
PI
Programming Interface
00h
RO
0Ah
SUBC2
Sub-Class Code
00h
RO
0Bh
BCC2
Base Class Code
03h
RO
0Ch
CLS
Cache Line Size Register
00h
RO
0Dh
MLT2
Master Latency Timer
00h
RO
0Eh
HDR2
Header Type
01h
RO
0Fh
BIST
BIST Register
00h
RO
10-13h
GMADR
Graphics Memory Range Address
00000008h
Read/Write
Datasheet
®
Intel 82815EM GMCH
R
Address
Offset
3.8.1.
Register
Symbol
Register Name
Default Value
Access Type
14-17h
MMADR
Memory Mapped Range Address
00000000h
Read/Write
18-2Bh

Reserved
00h

2C-2Dh
SVID
Subsystem Vendor ID
0000h
Read/WriteO
2E-2Fh
SID
Subsystem ID
0000h
Read/WriteO
30-33h
ROMADR
Video Bios ROM Base Address
00000000h
RO
34h
CAPPOINT
Capabilities Pointer
DCh
RO
35-3Bh

Reserved
00h

3Ch
INTRLINE
Interrupt Line Register
00h
Read/Write
3Dh
INTRPIN
Interrupt Pin Register
01h
RO
3Eh
MINGNT
Minimum Grant Register
00h
RO
3Fh
MAXLAT
Maximum Latency Register
00h
RO
40-DBh

Reserved
00h

DC-DDh
PM_CAPID
Power Management Capabilities
0001h
RO
DE-DFh
PM_CAP
Power Management Capabilities
0022h
RO
E0-E1h
PM_CS
Power Management Control
0000h
Read/Write
E2-FFh

Reserved
00h

VID2—Vendor Identification Register (Device 2)
Address Offset:
Default Value:
Attribute:
00h−01h
8086h
Read Only
The VID Register contains the vendor identification number. This 16-bit register combined with the
Device Identification Register uniquely identifies any PCI device. Writes to this register have no effect.
Bit
15:0
Datasheet
Description
Vendor Identification Number. This is a 16-bit value assigned to Intel®.
97
®
Intel 82815EM GMCH
R
3.8.2.
DID2—Device Identification Register (Device 2)
Address Offset:
Default Value:
Attribute:
02h−03h
1132h
Read Only
This 16-bit register combined with the Vendor Identification register uniquely identifies any PCI device.
Writes to this register have no effect.
Bit
Description
15:0
Device Identification Number. This is a 16 bit value assigned to the internal graphics device of the
GMCH2-M.
1132h = Device ID for Device 2.
3.8.3.
PCICMD2—PCI Command Register (Device 2)
Address Offset:
Default:
Access:
04h−05h
0004h
Read Only, Read/Write
This 16-bit register provides basic control over the GMCH2-M's ability to respond to PCI cycles. The
PCICMD Register in GMCH2-M disables GMCH2-M PCI compliant master accesses to main memory.
15
10
Reserved (0)
8
FB2B
(HW=0)
SERR En
(HW=0)
7
6
5
4
3
2
1
0
Addr/Data
Stepping
(HW=0)
Parity
Error En
(HW=0)
VGA Pal
Sn
(HW=0)
Mem WR
& Inval En
(HW=0)
Special
Cycle En
(HW=0)
Bus
Master En
Mem Acc
En
I/O Acc En
Bits
98
9
Description
15:10
Reserved.
9
Fast Back-to-Back (FB2B)
RO. (Not Implemented). Hardwired to 0.
8
SERR# Enable (SERRE) RO. (Not Implemented). Hardwired to 0.
7
Address/Data Stepping
RO. (Not Implemented). Hardwired to 0.
6
Parity Error Enable (PERRE) RO. (Not Implemented). Hardwired to 0. Since GMCH2-M belongs to
the category of devices that does not corrupt programs or data in system memory or hard drives,
GMCH2-M ignores any parity error that it detects and continues with normal operation.
5
Video Palette Snooping (VPS) RO. This bi is hardwired to 0 to disable snooping.
4
Memory Write and Invalidate Enable (MWIE) RO. Hardwired to 0. GMCH2-M does not support
memory write and invalidate commands.
3
Special Cycle Enable (SCE) RO. This bit is hardwired to 0. GMCH2-M ignores Special cycles.
Datasheet
®
Intel 82815EM GMCH
R
2
Bus Master Enable (BME) RO. Hardwired to 1 to enable GMCH2-M to function as a PCI compliant
master.
1
Memory Access Enable (MAE) Read/Write. This bit controls GMCH2-M’s response to memory
space accesses.
0 = Disable (default).
1 = Enable.
0
I/O Access Enable (IOAE) Read/Write. This bit controls GMCH2-M’s response to I/O space
accesses.
0 = Disable (default).
1 = Enable.
Datasheet
99
®
Intel 82815EM GMCH
R
3.8.4.
PCISTS2—PCI Status Register (Device 2)
Address Offset:
Default Value:
Access:
06h−07h
02B0h
Read Only
PCISTS is a 16-bit status register that reports the occurrence of a PCI compliant master abort and PCI
compliant target abort. PCISTS also indicates the DEVSEL# timing that has been set by the GMCH2-M
hardware.
15
14
13
12
11
Detected
Par Error
(HW=0)
Sig Sys
Error
(HW=0)
Recog
Mast Abort
Sta
(HW=0)
Rec
Target
Abort Sta
(HW=0)
Sig Target
Abort Sta
(HW=0)
7
6
5
4
FB2B
(HW=1)
User Def
Format
(HW=0)
66 MHz
PCI Cap
(HW=1)
Cap List
(HW=1)
Bits
100
10
9
DEVSEL# Timing
(HW=01)
8
Data Par
Detected
(HW=0)
3
0
Reserved
Description
15
Detected Parity Error (DPE) RO. Since GMCH2-M does not detect parity, this bit is always set to 0.
14
Signaled System Error (SSE) RO. The GMCH2-M Graphics device never asserts SERR#,
therefore this bit is hardwired to 0.
13
Received Master Abort Status (RMAS) RO. The GMCH2-M Graphics device never gets a Master
Abort, therefore this bit is hardwired to 0.
12
Received Target Abort Status (RTAS) RO. The GMCH2-M Graphics device never gets a Target
Abort, therefore this bit is hardwired to 0.
11
Signaled Target Abort Status (STAS). Hardwired to 0. GMCH2-M does not use target abort
semantics.
10:9
DEVSEL# Timing (DEVT) RO. This 2-bit field indicates the timing of the DEVSEL# signal when
GMCH2-M responds as a target. Hardwired to 01 to indicate that GMCH2-M is a medium decode
device.
8
Data Parity Detected (DPD) Read/WriteC. Since Parity Error Response is hardwired to disabled
(and GMCH2-M does not do any parity detection), this bit is hardwired to 0.
7
Fast Back-to-Back (FB2B). Hardwired to 1. GMCH2-M accepts fast back-to-back when the
transactions are not to the same agent.
6
User Defined Format (UDF). Hardwired to 0.
5
66 MHz PCI Capable (66C). Hardwired to 1 indicating that GMCH2-M is 66MHz PCI capable. Note
that there is nothing in the design that prevents is 66MHz PCI from working, but we support on a
subset of 66MHz PCI that is essentially equivalent to AGP. 66MHz PCI is not a feature we validate.
4
CAP LIST
RO. This bit is set to 1 to indicate that the register at 34h provides an offset into the
function’s PCI Configuration Space containing a pointer to the location of the first item in the list.
3:0
Reserved.
Datasheet
®
Intel 82815EM GMCH
R
3.8.5.
RID2—Revision Identification Register (Device 2)
Address Offset:
Default Value:
Access:
08h
10h for A0-step silicon
11h for A1-step silicon
Read Only
This register contains the revision number of GMCH2-M. These bits are read only and writes to this
register have no effect.
Bits
7:0
3.8.6.
Description
Revision Identification Number. This is an 8-bit value that indicates the revision identification
number for GMCH2-M. The four MSB’s are for process differentiation and the four LSB’s indicate
stepping. Default = 10h
PI—Programming Interface Register (Device 2)
Address Offset:
Default Value:
Access:
09h
00h
Read Only
This register contains the device-programming interface for GMCH2-M.
Bits
7:0
3.8.7.
Description
Programming Interface (PI). 00h=Hardwired as a Display controller.
SUBC2—Sub-Class Code Register (Device 2)
Address Offset:
Default Value:
Access:
Size:
0Ah
00h
Read Only
8 bits
This register contains the Sub-Class Code for the GMCH2-M Device 2.
Bit
7:0
Datasheet
Description
Sub-Class Code (SUBC). This is an 8-bit value that indicates the category of Display controller into
which GMCH2-M falls. The code is 00h indicating a VGA compatible device. Default = 00h
101
®
Intel 82815EM GMCH
R
3.8.8.
BCC2—Base Class Code Register (Device 2)
Address Offset:
Default Value:
Access:
Size:
0Bh
03h
Read Only
8 bits
This register contains the Base Class Code of the GMCH2-M Function #1.
Bit
7:0
3.8.9.
Description
Base Class Code (BASEC). This is an 8-bit value that indicates the Base Class Code for GMCH2-M.
This code has the value 03h, indicating a Display controller.
CLS—Cache Line Size Register (Device 2)
Address Offset:
Default Value:
Access:
0Ch
00h
Read only
GMCH2-M does not support this register as a PCI slave.
Bits
7:0
3.8.10.
Description
Cache Line Size (CLS). Hardwired to 0s. GMCH2-M as a PCI compliant master does not use the
Memory Write and Invalidate command and, in general, does not perform operations based on cache
line size.
MLT2—Master Latency Timer Register (Device 2)
Address Offset:
Default Value:
Access:
Bits
7:0
102
0Dh
00h
Read Only
Description
Master Latency Timer Count Value. Hardwired to 0s. GMCH2-M does not support the
programmability of the master latency timer because it does not perform bursts.
Datasheet
®
Intel 82815EM GMCH
R
3.8.11.
HDR2—Header Type Register (Device 2)
Address Offset:
Default Value:
Access:
0Eh
00h
Read Only
This register contains the Header Type of GMCH2-M.
Bits
Description
7:0
Header Type (HTYPE). This is an 8-bit value that indicates the Header Type for GMCH2-M.
00h = Indicating a basic (i.e., single function) configuration space format.
3.8.12.
BIST—BIST Register (Device 2)
Address Offset:
Default Value:
Access:
0Fh
00h
Read Only
This register is used for control and status of Built In Self Test (BIST).
7
BIST
Supported
(HW=0)
Bits
6
0
Reserved
Description
7
BIST Supported. BIST is not supported. This bit is hardwired to 0.
6:0
Reserved.
Datasheet
103
®
Intel 82815EM GMCH
R
3.8.13.
GMADR-Graphics Memory Range Address Register (Device 2)
Address Offset:
Default Value:
Access:
10−13h
00000008h
Read/Write, Read Only
This register requests allocation for internal Graphics memory. The allocation is for either 32 MB or 64
MB of memory space (selected by bit 0 of the Device 0 MISCC Reg) and the base address is defined by
bits [31:25,24].
31
26
Memory Base Address
15
25
24
16
64 MB
Addr. Mask
4
Address Mask (cont)
(HW=0; 32MB addr range)
Bit
Address Mask
(HW=0; 32MB addr range)
3
2
Prefetch
Mem En
(HW=1)
1
Memory Type
(HW=0; 32MB addr)
0
Mem/IO
Space
(HW=0)
Descriptions
31:26
Memory Base Address
Read/Write. Set by the OS, these bits correspond to address signals
[31:26].
25
64 MB Address Mask
RO , Read/Write.
If Device 0 MISCC Reg bit 0 = 0, then this bit is RO with a value of 0, indicating a memory range of
64 MB.
If Device 0 MISCC Reg bit 0 = 1, then this bit is Read/Write, indicating a memory range of 32 MB.
104
24:4
Address Mask
RO. Hardwired to 0s to indicate 32 MB address range.
3
Prefetchable Memory
RO. Hardwired to 1 to enable prefetching.
2:1
Memory Type
RO. Hardwired to 0 to indicate 32-bit address.
0
Memory/IO Space
RO. Hardwired to 0 to indicate memory space.
Datasheet
®
Intel 82815EM GMCH
R
3.8.14.
MMADR—Memory Mapped Range Address Register (Device 2)
Address Offset:
Default Value:
Access:
14−17h
00000000h
Read/Write, Read Only
This register requests allocation for GMCH2-M registers and instruction ports. The allocation is for 512
KB and the base address is defined by bits [31:19].
31
19
Memory Base Address
(addr bits [31:19])
15
4
Address Mask (cont)
(HW=0; 512 KB addr range)
Bit
3.8.15.
18
16
Address Mask (HW=0;
512 KB addr range)
3
Prefetch
Mem En
(HW=0)
2
1
Memory Type
(HW=0; 32 Mb addr)
0
Mem/IO
Space
(HW=0)
Descriptions
31:19
Memory Base Address
Read/Write. Set by the OS, these bits correspond to address signals
[31:19].
18:4
Address Mask
RO. Hardwired to 0s to indicate 512 KB address range.
3
Prefetchable Memory
RO. Hardwired to 0 to prevent prefetching.
2:1
Memory Type
RO. Hardwired to 0s to indicate 32-bit address.
0
Memory / IO Space
RO. Hardwired to 0 to indicate memory space.
SVID—Subsystem Vendor Identification Register (Device 2)
Address Offset:
Default Value:
Access:
Bit
15:0
2C−2Dh
0000h
Read/Write Once
Descriptions
Subsystem Vendor ID (Read/WriteO). This value is used to identify the vendor of the subsystem.
The default value is 0000h. This field should be programmed by BIOS during boot-up.
Once written, this register becomes Read_Only. This Register can only be cleared by a Reset.
Datasheet
105
®
Intel 82815EM GMCH
R
3.8.16.
SID—Subsystem Identification Register (Device 2)
Address Offset:
Default Value:
Access:
2E−2Fh
0000h
Read/Write Once
Bit
15:0
Descriptions
Subsystem ID (Read/WriteO). This value is used to identify a particular subsystem.
The default value is 0000h. This field should be programmed by BIOS during boot-up.
Once written, this register becomes Read_Only. This Register can only be cleared by a Reset.
3.8.17.
ROMADR - Video BIOS ROM Base Address Registers (Device 2)
Address Offset:
Default Value:
Access:
30−33h
00000000h
Read Only
31
18
ROM Base Address
(addr bits [31:18])
15
11
10
Bit
1
Reserved
0
ROM
BIOS En
(HW=0)
Descriptions
31:18
ROM Base Address
RO. Hardwired to 0s.
17:11
Address Mask
RO. Hardwired to 0s to indicate 256 KB address range.
10:1
Reserved. Hardwired to 0s.
0
ROM BIOS Enable
RO. 0 = ROM not accessible.
CAPPOINT—Capabilities Pointer Register (Device 2)
Address Offset:
Default Value:
Access:
Bit
7:0
106
16
Address Mask (HW=0;
256 KB addr range)
Address Mask (cont)
(HW=0; 256 KB addr range)
3.8.18.
17
34h
DCh
Read Only
Descriptions
Pointer to the start of AGP standard register block. Since there is no AGP bus on GMCH2-M, this
field is set to DCh to point to the Power Management Capabilities ID Register
Datasheet
®
Intel 82815EM GMCH
R
3.8.19.
INTRLINE—Interrupt Line Register (Device 2)
Address Offset:
Default Value:
Access:
3Ch
00h
Read/Write
Bit
7:0
3.8.20.
Descriptions
Interrupt Connection. Used to communicate interrupt line routing information. POST software writes
the routing information into this register as it initializes and configures the system. The value in this
register indicates which input of the system interrupt controller that the device’s interrupt pin is
connected to.
INTRPIN—Interrupt Pin Register (Device 2)
Address Offset:
Default Value:
Access:
3Dh
01h
Read Only
Bit
7:0
Descriptions
Interrupt Pin. As a single function device, GMCH2-M specifies INTA# as its interrupt pin.
01h=INTA#.
3.8.21.
MINGNT—Minimum Grant Register (Device 2)
Address Offset:
Default Value:
Access:
3Eh
00h
Read Only
Bit
7:0
Descriptions
Minimum Grant Value. GMCH2-M does not burst as a PCI compliant master.
Bits[7:0]=00h.
3.8.22.
MAXLAT—Maximum Latency Register (Device 2)
Address Offset:
Default Value:
Access:
Bit
7:0
Datasheet
3Fh
00h
Read Only
Descriptions
Maximum Latency Value. Bits[7:0]=00h. GMCH2-M has no specific requirements for how often it
needs to access the PCI bus.
107
®
Intel 82815EM GMCH
R
3.8.23.
PM_CAPID—Power Management Capabilities ID Register
(Device 2)
Address Offset:
Default Value:
Access:
DCh−DDh
0001h
Read Only
15
8
7
0
NEXT_PTR
CAP_ID
Bits
3.8.24.
Description
15:8
NEXT_PTR. This contains a pointer to next item in capabilities list. This the final capability in the list
and must be set to 00h.
7:0
CAP_ID. SIG defines this ID is 01h for power management.
PM_CAP—Power Management Capabilities Register (Device 2)
Address Offset:
Default Value:
Access:
DEh−DFh
0022h
Read Only
15
11
PME Support (HW=0)
7
6
Reserved
5
4
3
Dev
Specific
Init
(HW=1)
Aux Pwr
Src
(HW=0)
PME
Clock
(HW=0)
Bits
108
10
9
8
D2
(HW=0)
D1
(HW=0)
Reserved
2
0
Version
Description
15:11
PME Support. This field indicates the power states in which GMCH2-M may assert PME#. Hardwired
to 0 to indicate that GMCH2-M does not assert the PME# signal.
10
D2. Hardwired to 0 to indicate D2 power management state is not supported.
9
D1. Hardwired to0 to indicate that D1 power management state is NOT supported.
8:6
Reserved. Read as 0s.
5
Device Specific Initialization (DSI). Hardwired to 1 to indicate that special initialization of GMCH2-M
is required before generic class device driver is to use it.
4
Auxiliary Power Source. Hardwired to 0.
3
PME Clock. Hardwired to 0 to indicate GMCH2-M does not support PME# generation.
2:0
Version. Hardwired to 010b to indicate there are 4 bytes of power management registers
implemented and that this device complies with revision 1.1 of the PCI Power Management Interface
Specification.
Datasheet
®
Intel 82815EM GMCH
R
3.8.25.
PM_CS - Power Management Control/Status Register (Device 2)
Address Offset:
Default Value:
Access:
15
PME Sta
(HW=0)
E0h−E1h
0000h
Read/Write
14
13
12
Data Scale (Reserved)
9
Data_Select (Reserved)
7
2
Reserved
Bits
8
PME En
1
0
PowerState
Description
15
PME_Status RO. This bit is 0 to indicate that GMCH2-M does not support PME# generation from
D3 (cold).
14:13
Data Scale (Reserved) RO. GMCH2-M does not support data register. This bit always returns 0
when read; write operations have no effect.
12:9
Data_Select (Reserved) RO. GMCH2-M does not support data register. This bit always returns 0
when read; write operations have no effect.
8
PME_En
Read/Write. This bit is 0 to indicate that PME# assertion from D3 (cold) is disabled.
7:2
Reserved. Always returns 0s when read; write operations have no effect.
1:0
PowerState
Read/Write. This field indicates the current power state of GMCH2-M and can be used
to set GMCH2-M into a new power state. If software attempts to write an unsupported state to this
field, write operation must complete normally on the bus, but the data is discarded and no state
change occurs.
00 = D0
01 = D1 (Not supported in GMCH2-M)
10 = D2 (Not supported in GMCH2-M)
11 = D3
Datasheet
109
®
Intel 82815EM GMCH
R
This page left intentionally blank.
110
Datasheet
®
Intel 82815EM GMCH
R
4.
Functional Description
This chapter describes the Graphics and Memory Controller Hub (GMCH2-M) interfaces, and boot
sequencing. The “System Address Map” provides a system-level address memory map and describes the
memory space controls provided by the GMCH2-M.
4.1.
System Memory and I/O Address Map
A mobile Intel Pentium® III processor or mobile Intel Celeron™ processor system based on the
GMCH2-M, supports 4 GB of addressable memory space and 64 KB+3 of addressable I/O space. (The
Pentium® III bus I/O addressability is 64KB + 3). There is a programmable memory address (PAM)
space under the 1 MB region which can be controlled with programmable attributes of Write Only, or
Read Only. Attribute programming is described in the Configuration Register Description section. This
section focuses on how the memory space is partitioned and what these separate memory regions are
used for. The I/O address space is explained at the end of this section.
The mobile Intel Pentium® III processor, mobile Intel Celeron™ processor supports addressing of
memory ranges larger than 4 GB. The GMCH2-M Host Bridge claims any access over 4 GB by
terminating transaction (without forwarding it to the hub interface). Writes are terminated by dropping
the data and for reads the GMCH2-M returns all zeros on the host bus.
In the following sections, it is assumed that all of the compatibility memory ranges reside on the hub
interface. The exceptions to this rule are the VGA ranges, which may be mapped to the internal Graphics
Device.
The GMCH2-M Memory Map includes a number of programmable ranges, ALL of these ranges MUST
be unique and NON-OVERLAPPING. There are NO Hardware Interlocks to prevent problems in the
case of overlapping ranges. Accesses to overlapped ranges may produce indeterminate results.
4.1.1.
Memory Address Space
The following figure shows a high-level representation of the system memory address map. Figure 5
provides additional details on mapping specific memory regions as defined and supported by the
GMCH2-M chipset.
Datasheet
111
®
Intel 82815EM GMCH
R
Figure 4. System Memory Address Map
4 GB
PCI
Memory
Address
Range
Local
Memory
Range
Memory
Mapped
Range
Top of the Main Memory
Main
Memory
Address
Range
0
Independently
Programmable NonOverlapping Memory
Windows
mem_map_s
112
Datasheet
®
Intel 82815EM GMCH
R
Figure 5. Detailed Memory System Address Map
System Memory
Space
64 GB
Extended
P6
Memory
4 GB Max TOM
AGP
Window
PCI
Memory
Range
Graphics
Aperture
0FFFFFh
1 MB
Upper BIOS Area
(64 KB)
512 MB
0F0000h
0EFFFFh
960 KB
Lower BIOS Area
(64 KB; 16 KB x 4)
0E0000h
0DFFFFh
896 KB
16 MB
Expansion Card
BIOS
and Buffer Area
(128 KB;
16 KB x 8)
Optional ISA Hole
15 MB
Main Memory
Range
1 MB
DOS
Compatibility
Memory
640 KB
Optionally
mapped to the
internal AGP
0C0000h
0BFFFFh
0A0000h
09FFFFh
Std PCI/ISA
Video Mem
(SMM Mem)
128 KB
768 KB
640 KB
DOS Area
(640 KB)
0 MB
000000h
0 KB
mem_map
Datasheet
113
®
Intel 82815EM GMCH
R
4.2.
DOS Compatibility Memory Space
This area is divided into the following address regions:
• 0 - 640 KB DOS Area
• 640 - 768 KB Video Buffer Area
• 768 - 896 KB in 16KB sections (total of 8 sections) - Expansion Area
• 896 -960 KB in 16KB sections (total of 4 sections) - Extended System BIOS Area
• 960 KB - 1 MB Memory (BIOS Area) - System BIOS Area
The Intel® 815EM chipset supports all sixteen memory segments of interest in the compatibility area.
Thirteen of the memory ranges can be enabled or disabled independently for both read and write cycles.
Table 10. Memory Segments and Their Attributes
Memory
Segments
114
Attributes
Comments
000000H 09FFFFH
Fixed - always mapped to main DRAM
0 to 640K - DOS Region
0A0000H 0BFFFFH
Mapped to the hub interface, AGP/PCI bus or
internal graphics - configurable as SMM space
Video Buffer (physical DRAM
configurable as SMM space)
0C0000H 0C3FFFH
WE RE
Add-on BIOS
0C4000H 0C7FFFH
WE RE
Add-on BIOS
0C8000H 0CBFFFH
WE RE
Add-on BIOS
0CC000H 0CFFFFH
WE RE
Add-on BIOS
0D0000H 0D3FFFH
WE RE
Add-on BIOS
0D4000H 0D7FFFH
WE RE
Add-on BIOS
0D8000H 0DBFFFH
WE RE
Add-on BIOS
0DC000H 0DFFFFH
WE RE
Add-on BIOS
0E0000H 0E3FFFH
WE RE
BIOS Extension
0E4000H 0E7FFFH
WE RE
BIOS Extension
0E8000H 0EBFFFH
WE RE
BIOS Extension
0EC000H 0EFFFFH
WE RE
BIOS Extension
Datasheet
®
Intel 82815EM GMCH
R
Memory
Segments
0F0000H 0FFFFFH
4.2.1.1.
Attributes
WE RE
Comments
BIOS Area
DOS Area (00000h-9FFFh)
The DOS area is 640 KB in size is always mapped to the main memory controlled by the GMCH2-M.
4.2.1.2.
Video Buffer Area (A0000h-BFFFFh)
The 128KB graphics adapter memory region is normally mapped to a legacy video device on the PCI bus
(typically VGA controller). This memory area is not controlled by attribute bits of the PAM register.
processor-initiated cycles in this region are forwarded to the hub interface or the AGP/Internal Graphics
Device for termination. Accesses to this range are directed to either the hub interface or AGP/internal
Graphics Device based on the configuration. The configuration are specified by:
• AGP on/off configuration bit
• AGP off: GMS bits of the SMRAM register in the GMCH2-M Device #0 configuration space.
There is additional steering information coming from the Device #2* configuration registers and
from some of the VGA registers in the Graphics device.
• AGP on: PCICMD1 (PCI-PCI Command) and BCTRL (PCI-PCI Bridge Control) registers in
Device #1 configuration registers
The control is applied for accesses initiated from any of the system interface i.e. processor bus, the hub
interface, or AGP (if enabled).
For Hub interface to AGP/PCI bus accesses, only memory write operations are supported. Any AGP/PCI
initiated VGA accesses targeting the GMCH2-M will master abort.
This region is also the default region for SMM space in GFX mode. The SMRAM Control register
controls how SMM accesses to this space are treated.
4.2.1.3.
Monochrome Adapter (MDA) Range (B0000h - B7FFFh)
Legacy support requires the ability to have a second graphics controller (monochrome) in the system.
In an AGP system, accesses in the standard VGA range are forwarded to the AGP bus (depending on
configuration bits). Since the monochrome adapter may be on the hub interface/PCI (or ISA) bus,
GMCH2-M must decode cycles in the MDA range and forward them to the hub interface. This
capability is controlled by a configuration bit (MDA bit – Device 0, BEh). In addition to the memory
range B0000h to B7FFFh, the GMCH2-M decodes IO cycles at 3B4h, 3B5h, 3B8h, 3B9h, 3BAh and
3BFh and forwards them to the hub interface bus
In an internal graphics system, the GMS bits of the SMRAM register in Device #0, bits in the Device #2
PCICMD2 register and bits from some of the VGA registers control this functionality.
Datasheet
115
®
Intel 82815EM GMCH
R
4.2.1.4.
Expansion Area (C0000h-DFFFFh)
This 128 KB ISA Expansion region is divided into eight 16 KB segments. Each segment can be assigned
one of four Read/Write states: read-only, write-only, Read/Write, or disabled. Typically, these blocks
are mapped through GMCH2-M and are subtractively decoded to ISA space. Memory that is disabled is
not remapped.
4.2.1.5.
Extended System BIOS Area (E0000h-EFFFFh)
This 64 KB area is divided into four 16 KB segments. Each segment can be assigned independent read
and write attributes so it can be mapped either to main DRAM or to the hub interface. Typically, this
area is used for RAM or ROM. Memory segments that are disabled are not remapped elsewhere.
4.2.1.6.
System BIOS Area (F0000h-FFFFFh)
This area is a single 64KB segment. This segment can be assigned read and write attributes. It is by
default (after reset) Read/Write disabled and cycles are forwarded to the hub interface. By manipulating
the Read/Write attributes, the GMCH2-M can “shadow” BIOS into the main DRAM. When disabled,
this segment is not remapped.
4.3.
Extended Memory Area
This memory area covers 10 0000h (1 MB) to FFFF FFFFh (4 GB-1) address range and it is divided into
the following regions:
• Main DRAM Memory from 1 MB to the Top of Main Memory (TOM); maximum of 512 MB.
• PCI Memory space from the Top of Memory to 4 GB
4.3.1.
Main DRAM Address Range (0010_0000h to TOM)
The address range from 1 MB to the top of main memory (TOM) is mapped to main DRAM address
range. The Tom of memory is limited to 512Mb. All accesses to addresses within this range will be
forwarded to the DRAM unless a hole in this range is created.
4.3.1.1.
15MB-16MB Hole Area
A hole can be created at 15MB-16MB as controlled by the fixed hole enable (FDHC register) in Device
0 space. Accesses within this hole are forwarded to the hub interface. The range of physical DRAM
memory disabled by opening the hole is not remapped to the Top of the memory – that physical DRAM
space is not accessible. This 15MB-16MB hole is an optionally enabled ISA hole. Video accelerators
originally used this hole. It is also used by validation and customer SV teams for some of their test cards.
That is why it is being supported. There is no inherent BIOS request for the 15-16 hole.
4.3.1.2.
Extended SMRAM Address Range
The HSEG and TSEG SMM transaction address spaces reside in this extended memory area. This
memory resource is normally hidden from the system OS so that the processor has immediate access to
this memory space upon entry to SMM.
116
Datasheet
®
Intel 82815EM GMCH
R
4.3.1.3.
HSEG (High Segment)
SMM-mode processor accesses to enabled HSEG are remapped to 000A0000h-000BFFFFh. Non-SMMmode processor accesses to enabled HSEG are considered invalid are terminated immediately on the
FSB. The exception to this is non-SMM-mode Write Back cycles. They are remapped to SMM space to
maintain cache coherency. AGP and hub interface originated cycles to enabled SMM space are not
allowed. Physical DRAM behind the HSEG transaction address is not remapped and is not accessible.
4.3.1.3.1. TSEG (Top of Memory Segment)
TSEG can be up to 1MB in size and is at the top of memory (TOM). SMM-mode processor accesses to
enabled TSEG access the physical DRAM at the same address. Non-SMM-mode processor accesses to
enabled TSEG are considered invalid and are terminated immediately on the FSB. The exception is nonSMM-mode Write Back cycles. They are directed to the physical SMM space to maintain cache
coherency. AGP and Hub interface originated cycles to enabled SMM memory space are not allowed.
The size of the SMRAM space is determined by the USMM value in the SMRAM register. When the
extended SMRAM space is enabled, non-SMM processor accesses and all other accesses in this range
are forwarded to the hub interface. When SMM is enabled the amount of memory available to the
system is equal to the amount of physical DRAM minus the value in the TSEG register.
4.3.2.
PCI Memory Address Range (Top of Main Memory to 4 GB)
The address range from the top of main DRAM to 4 GB (top of physical memory space supported by the
GMCH2-M) is normally mapped via the hub interface to PCI. There are two exceptions to this rule:
Internal graphics configuration (GFX)
• Addresses decoded to the Local Memory Range are forwarded to the GFX
• Addresses decoded to the Memory Mapped Range of the internal graphics device (GFX) are
forwarded to the GFX
AGP configuration
• Addresses decoded to the AGP Memory Window defined by the MBASE, MLIMIT, PMBASE, and
PMLIMIT registers are mapped to AGP.
• Addresses decoded to the Graphics Aperture range defined by the APBASE and APSIZE registers
are mapped to the main DRAM.
There are two sub-ranges within the PCI Memory address range defined as APIC Configuration Space
and High BIOS Address Range. As an internal graphics device, the Local Memory Range and the
Memory Mapped Range of the internal Graphics Device MUST NOT overlap with these two ranges.
Similarly, as an AGP device, the AGP memory window and Graphics Aperture Window MUST NOT
overlap with these two ranges. These ranges are described in detail in the following sections.
Datasheet
117
®
Intel 82815EM GMCH
R
4.3.2.1.
APIC Configuration Space (FEC0_0000h -FECF_FFFFh, FEE0_0000hFEEF_FFFFh)
These ranges are reserved for APIC configuration space. The default Local APIC configuration space is
FEE0_0000h to FEEF_FFFFh.
Processor accesses to the Local APIC configuration space do not result in external bus activity since the
Local APIC configuration space is internal to the processor. However, a MTRR must be programmed to
make the Local APIC range uncacheable (UC). The Local APIC base address in each processor should
be relocated to the FEC0_0000h (4GB-20MB) to FECF_FFFFh range so that one MTRR can be
programmed to 64 KB for the Local and I/O APICs. The I/O APIC(s) usually reside in the I/O Bridge
portion (ICH2-M) of the chip-set or as a stand-alone component(s).
I/O APIC units will be located beginning at the default address FEC0_0000h. The first I/O APIC will be
located at FEC0_0000h. Each I/O APIC unit is located at FEC0_x000h where x is I/O APIC unit number
0 through F(hex). This address range will be normally mapped via the hub interface to PCI.
Note:
There is no provision to support an I/O APIC device on AGP.
The address range between the APIC configuration space and the High BIOS (FED0_0000h to
FFDF_FFFFh) is always mapped via the hub interface to PCI.
4.3.2.2.
High BIOS Area (FFE0_0000h -FFFF_FFFFh)
The top 2 MB of the Extended Memory Region is reserved for System BIOS (High BIOS), extended
BIOS for PCI devices, and the A20 alias of the system BIOS. Processor begins execution from the High
BIOS after reset. This region is mapped via the hub interface to PCI so that the upper subset of this
region aliases to 16 MB-256 MB range. The actual address space required for the BIOS is less than 2
MB but the minimum processor MTRR range for this region is 2 MB so that full 2 MB must be
considered. The ICH2-M supports a maximum of 1MB in the High BIOS range.
4.4.
System Management Mode (SMM) Memory Range
The GMCH2-M supports the use of main memory as System Management RAM (SMRAM) enabling the
use of System Management Mode (SMM). The GMCH2-M supports three SMM options:
• Compatible SMRAM (AB segment enabled)
• High Segment (HSEG)
• Top of Memory Segment (TSEG)
System Management RAM (SMRAM) space provides a memory area that is available for the SMI
handler's and code and data storage. This memory resource is normally hidden from the operating
system so that the processor has immediate access to this memory space upon entry to SMM. The
GMCH2-M provides three SMRAM options:
• Below 1 MB option that supports compatible SMI handlers.
• Above 1 MB option that allows new SMI handlers to execute with write-back cacheable SMRAM.
118
Datasheet
®
Intel 82815EM GMCH
R
• Optional larger write-back cacheable TSEG area of either 512 KB or 1MB in size above 1MB that
is reserved from the highest area in system DRAM memory. The above 1MB solutions require
changes to compatible SMRAM handlers code to properly execute above 1MB.
The hub interface and AGP masters are not allowed to access the SMM space. This must be insured even
for the GTLB translation.
4.4.1.
SMM Space Definition
SMM space has addressed SMM space and DRAM SMM space. The addressed SMM space is the
range of bus addresses used by the processor to access SMM space. DRAM SMM space is the range of
physical DRAM memory locations containing the SMM code.
SMM space can be accessed at one of three transaction address ranges: Compatible, HSEG and TSEG.
The Compatible and TSEG SMM space is not remapped. Therefore the addressed and DRAM SMM
space is the same address range. Since the HSEG SMM space is remapped, the addressed and DRAM
SMM space is a different address range. Note that the HSEG DRAM space is the same as the
Compatible Transaction Address space. Therefore the table below describes three unique address ranges:
• Compatible Transaction Address (Adr C)
• High Transaction Address (Adr H)
• TSEG Transaction Address (Adr T)
These abbreviations are used later in the table describing SMM Space Transaction Handling.
Table 11. SMM Space Abbreviations
SMM Space Enabled
Transaction Address Space (Adr)
DRAM Space (DRAM)
Compatible (C)
A_0000h to B_FFFFh
A_0000h to B_FFFFh
High (H)
0FEE_A0000h to 0FEE_BFFFFh
A_0000h to B_FFFFh
TSEG (T)
(TOM-TSEG_SZ) to TOM
(TOM-TSEG_SZ) to TOM
If the GMCH2-M is operating in internal graphics mode, a 512Kb or 1Mb window can optionally be
used for the display frame buffer. This window is taken after the TSEG window.
4.4.2.
SMM Space Restrictions
If any of the following conditions are violated the results of SMM accesses are unpredictable and may
cause the system to hang:
• The Compatible SMM space must not be set-up as cacheable.
• HSEG or TSEG SMM transaction address space must not overlap address space assigned to system
DRAM, the AGP aperture range, or to any “PCI” devices (including the hub interface and AGP
devices). This is a BIOS responsibility.
• When TSEG SMM space is enabled, the TSEG space must not be reported to the OS as available
DRAM. This is a BIOS responsibility.
• Any address translated through the AGP Aperture GTLB must not target DRAM from 000A_0000h
to 000F_FFFFh.
Datasheet
119
®
Intel 82815EM GMCH
R
4.4.3.
SMM Space Combinations
When HSEG SMM is enabled, the Compatible SMM space must be disabled. Processor originated
accesses to the Compatible SMM space are forwarded to AGP if VGAEN=1 (also depends on MDAP),
otherwise they are forwarded to the hub interface. AGP and the hub interface originated accesses are
never allowed to access SMM space. Only the processor is allowed to access SMM space. AGP and hub
interface originated transactions are not allowed to SMM space.
4.4.4.
Initialization and Usage of SMRAM and Graphics Local Memory
SMRAM Register Bits 7:4 control the usage of memory from Main Memory space for use as Graphics
Local Memory and SMM TSEG memory. The blocks of memory selected by these fields are NOT
accessible as general system RAM. When Bit 5 of the SMRAM register is a “1” the TSEG segment of
memory can ONLY be accessed by the processor in SMM mode (No other agent can access this
memory). Therefore, BIOS should initialize this block of memory BEFORE setting either Bit 5 or Bit 7
of the SMRAM register. The memory for TSEG is used first and then the Graphics Local Memory is
used. An example of this memory usage mechanism is:
• TOM equal 64 MB,
• TSEG selected as 512 KB in size,
• Graphics Local Memory selected as 1 MB in size
• General System RAM available in system = 62.5 MB
 General System RAM Range
00000000h to 03E7FFFFh
 TSEG Address Range
03F80000h to 03FFFFFFh
 TSEG used from
03F80000h to 03FFFFFFh
 Graphics Local Memory used from 03E80000h to 03F7FFFFh
4.5.
Memory Shadowing
Any block of memory that can be designated as read-only or write-only can be “shadowed” into
GMCH2-M DRAM memory. Typically this is done to allow ROM code to execute more rapidly out of
main DRAM. ROM is used as a read-only during the copy process while DRAM at the same time is
designated write-only. After copying, the DRAM is designated read-only so that ROM is shadowed.
processor bus transactions are routed accordingly.
4.6.
I/O Address Space
The GMCH2-M does not support the existence of any other I/O devices other than itself on the processor
bus. The GMCH2-M generates either hub interface or AGP/PCI (if enabled) bus cycles for all processor
I/O accesses. If internal graphics is enabled, the GMCH2-M routes the access to hub interface or legacy
I/O registers supported by the internal Graphics Device.
The GMCH2-M contains two internal registers in the processor I/O space, Configuration Address
Register (CONFIG_ADDRESS) and the Configuration Data Register (CONFIG_DATA). These
locations are used to implement PCI configuration space access mechanism and as described in
chapter 3.
120
Datasheet
®
Intel 82815EM GMCH
R
The I/O accesses, other than ones used for PCI configuration space access or ones that target the internal
Graphics Device (GFX) (or AGP/PCI) are forwarded to the hub interface. The GMCH2-M will not post
I/O write cycles to IDE. The PCICMD1 or PCICMD2 register can disable the routing of I/O cycles to
the AGP. The GMCH2-M never responds to I/O cycles initiated on AGP.
4.6.1.
AGP/PCI - I/O Address Mapping
The GMCH2-M can be programmed to direct non-memory (I/O) accesses to the AGP bus interface when
processor initiated I/O cycle addresses are within the AGP I/O address range. This range is controlled
via the I/O Base Address (IOBASE) and I/O Limit Address (IOLIMIT) registers in GMCH2-M Device
#1 configuration space.
Address decoding for this range is based on the following concept. The top 4 bits of the respective I/O
Base and I/O Limit registers correspond to address bits A[15:12] of an I/O address. For the purpose of
address decoding, the GMCH2-M assumes that lower 12 address bits A[11:0] of the I/O base are zero
and that address bits A[11:0] of the I/O limit address are FFFh. This forces the I/O address range
alignment to 4KB boundary and produces a size granularity of 4KB. The GMCH2-M positively decodes
I/O accesses to AGP I/O address space as defined by the following equation:
I/O_Base_Address ≤ processor I/O Cycle Address ≤ I/O_Limit_Address
The plug-and-play configuration software programs the effective size of the range and it depends on the
size of I/O space claimed by the AGP device.
The GMCH2-M also forwards accesses to the Legacy VGA I/O ranges according to the settings of
Device #1 configuration registers BCTRL (VGA Enable) and PCICMD1 (IOAE1), unless a second
adapter (monochrome) is present on the hub interface/PCI (or ISA). The MDAP configuration bit
determines the presence of a second graphics adapter. When MDAP is set, the MCH will decode legacy
monochrome IO ranges and forward them to the hub interface. The IO ranges decoded for the
monochrome adapter are 3B4h, 3B5h, 3B8h, 3B9h, 3Bah and 3BFh. The PCICMD1 register can disable
the routing of I/O cycles to the AGP.
Note that the GMCH2-M Device #1 I/O address range registers defined above are used for all I/O space
allocation for any devices requiring such a window on AGP. These devices would include the AGP
device, PCI-66MHz/3.3V agents, and multifunctional AGP devices where one or more functions are
implemented as PCI devices.
4.7.
GMCH2-M Address Decode Rules and Cross-Bridge
Address Mapping
The GMCH2-M’s address map applies globally to accesses arriving on any of the three interfaces (i.e.,
Host bus, hub interface or from the internal Graphics Device).
4.7.1.
Address Decode Rules
The GMCH2-M accepts all memory Read and Write accesses from the hub interface to both System
Memory and Graphics Memory. The hub interface accesses that fall elsewhere within the PCI memory
range will not be accepted. The GMCH2-M never responds to hub interface initiated I/O read or write
cycles.
Datasheet
121
®
Intel 82815EM GMCH
R
The GMCH2-M accepts accesses from the hub interface to the following address ranges:
• All memory Read/Write accesses to Main DRAM including PAM region (except SMM space)
• All memory Read/Write accesses to the Graphics Aperture defined by APBASE and APSIZE.
• All Hub interface memory write accesses to AGP memory range defined by MBASE, MLIMIT,
PMBASE, and PMLIMIT.
• Memory writes to VGA range on AGP if enabled.
The hub interface memory accesses that fall elsewhere within the memory range are considered invalid
and will be remapped to a translated memory address, snooped on the host bus, and dispatched to
DRAM. Reads will return all 1’s with Master Abort completion. Writes will have BE’s deasserted and
will terminate with Master Abort if completion is required. I/O cycles will not be accepted. They are
terminated with Master Abort completion packets.
4.7.2.
The Hub Interface Accesses to GMCH2-M that Cross Device
Boundaries
The Hub interface accesses are limited to 256 bytes but have no restrictions on crossing address
boundaries. A single hub interface request may therefore span device boundaries (AGP, DRAM) or cross
from valid addresses to invalid addresses (or visa versa). The GMCH2-M does not support transactions
that cross device boundaries. For reads and for writes requiring completion, the GMCH2-M will provide
separate completion status for each naturally-aligned 32 or 64 byte block. If the starting address of a
transaction hits a valid address the portion of a request that hits that target device (AGP or DRAM) will
complete normally.
The remaining portion of the access that crosses a device boundary (targets a different device than that of
the starting address) or hits an invalid address will be remapped to memory address 0h, snooped on the
host bus, and dispatched to DRAM. Reads will return all 1’s with Master Abort completion. Writes will
have BE’s deasserted and will terminate with Master Abort if completion is required.
If the starting address of a transaction hits a invalid address the entire transaction will be remapped to
memory address 0h, snooped on the host bus, and dispatched to DRAM. Reads will return all 1’s with
Master Abort completion. Writes will have BE’s deasserted and will terminate with Master Abort if
completion is required.
4.7.3.
4.7.3.1.
AGP Interface Decode Rules
Cycles Initiated Using PCI Protocol
The GMCH2-M does not support any AGP/PCI access targeting the hub interface. The GMCH2-M will
claim AGP/PCI initiated memory read and write transactions decoded to the main DRAM range or the
Graphics Aperture range. All other memory read and write requests will be master-aborted by the
AGP/PCI initiator as a consequence of GMCH2-M not responding to a transaction.
Under certain conditions, the GMCH2-M restricts access to the DOS Compatibility ranges governed by
the PAM registers by distinguishing access type and destination bus. The GMCH2-M accepts AGP/PCI
write transactions to the compatibility ranges if the PAM designates DRAM as write-able. If accesses to
a range are not write enabled by the PAM, the GMCH2-M does not respond and the cycle will result in a
master-abort. AGP/PCI read transactions to the compatibility ranges are accepted if the PAM designates
122
Datasheet
®
Intel 82815EM GMCH
R
DRAM as readable. If accesses to a range are not read enabled by the PAM, the GMCH2-M does not
respond and the cycle will result in a master-abort.
If agent on AGP/PCI issues an I/O or PCI Special Cycle transaction, the GMCH2-M will not respond
and cycle will result in a master-abort. The GMCH2-M will not accept PCI configuration cycles to the
internal GMCH2-M devices.
4.7.3.2.
Cycles Initiated Using AGP Protocol
All cycles must reference main memory i.e. main DRAM address range (excluding PAM) or Graphics
Aperture range (also physically mapped within DRAM but using different address range). AGP accesses
to the PAM region from 640K -to- 1MB are not allowed. AGP accesses to SMM space are not allowed.
AGP initiated cycles that target DRAM are not snooped on the host bus, even if they fall outside of the
AGP aperture range.
If a cycle is outside of a valid main memory range then it will terminate as follows:
Reads: Remap to memory address 0h, return data from address 0h, and set the IAAF error flag.
Writes: Remapped to memory address 0h with BE’s deasserted (effectively dropped “on the floor”) and
set the IAAF error flag.
4.7.3.3.
AGP Accesses to GMCH2-M that Cross Device Boundaries
For FRAME# accesses, when an AGP or PCI master gets disconnected it will resume at the new address
which allows the cycle to be routed to or claimed by the new target. Therefore, accesses should be
disconnected by the target on potential device boundaries. The GMCH2-M will disconnect AGP/PCI
transactions on 4KB boundaries.
AGP PIPE# and SBA accesses are limited to 256 bytes and must hit DRAM. AGP accesses are
dispatched to DRAM on naturally aligned 32 byte block boundaries. The portion of the request that hits a
valid address will complete normally. The portion of a read access that hits an invalid address will be
remapped to address 0h, return data from address 0h, and set the IAAF error flag. The portion of a write
access that hits an invalid address will be remapped to memory address 0h with BE’s deasserted
(effectively dropped “on the floor”) and set the IAAF error flag.
Datasheet
123
®
Intel 82815EM GMCH
R
4.7.4.
Legacy VGA Ranges
The legacy VGA memory range A_0000h-B_FFFFh is mapped either to the hub interface or to
AGP/PCI1 depending on the programming of the VGA Enable bit in the BCTRL configuration register
in GMCH2-M Device #1 configuration space, and the MDAP bit in the GMCH2-MCFG configuration
register in Device #0 configuration space. The same register controls mapping VGA I/O address ranges.
VGA I/O range is defined as addresses where A[9:0] are in the ranges 3B0h to 3BBh and 3C0h to 3DFh
(inclusive of ISA address aliases - A[15:10] are not decoded). The function and interaction of these two
bits is described below:
MDA Present (MDAP): This bit works with the VGA Enable bit in the BCTRL register of device 1 to
control the routing of processor initiated transactions targeting MDA compatible I/O and memory
address ranges. This bit should not be set when the VGA Enable bit is not set. If the VGA enable bit is
set, then accesses to IO address range x3BCh-x3BFh are forwarded to the hub interface. If the VGA
enable bit is not set, then IO address range accesses x3BCh-x3BFh are treated like other IO accesses the cycles are forwarded to AGP if the address is within IOBASE and IOLIMIT and ISA enable bit is not
set, otherwise they are forwarded to the hub interface. MDA resources are defined as the following:
• Memory: 0B0000h - 0B7FFFh
• I/O: 3B4h, 3B5h, 3B8h, 3B9h, 3BAh, 3BFh,
Note:
Including ISA address aliases, A[15:10] are not used in decode.
Any I/O reference that includes the I/O locations listed above, or their aliases, will be forwarded to the
hub interface even if the reference includes I/O locations not listed above.
VGA Enable: Controls the routing of processor initiated transactions targeting VGA compatible I/O
and memory address ranges. When this bit is set, GMCH2-M will forward the following processor
accesses to the AGP:
• Memory accesses in the range 0A0000h to 0BFFFFh
• I/O addresses where A[9:0] are in the ranges 3B0h to 3BBh and 3C0h to 3DFh
Note:
Inclusive of ISA address aliases - A[15:10] are not decoded.
When this bit is set, forwarding of these accesses issued by the processor is independent of the I/O
address and memory address ranges defined by the previously defined base and limit registers.
Forwarding of these accesses is also independent of the settings of the bit 2 (ISA Enable) of BCTRL if
this bit is “1”. If the VGA enable bit is set, then accesses to IO address range x3BCh-x3BFh are
forwarded to the hub interface. If the VGA enable bit is not set, then IO address range accesses x3BChx3BFh are treated like other IO accesses - the cycles are forwarded to AGP if the address is within
IOBASE and IOLIMIT and ISA enable bit is not set, otherwise they are forwarded to the hub interface.
If this bit is “0” (default) , then VGA compatible memory and I/O range accesses are not forwarded to
AGP but rather they are mapped to the hub interface unless they are mapped to AGP via I/O and memory
range registers defined above (IOBASE, IOLIMIT, MBASE, MLIMIT, PMBASE, PMLIMIT
124
Datasheet
®
Intel 82815EM GMCH
R
4.8.
Host Interface
The host interface of the GMCH2-M is optimized to support the mobile Intel Pentium® III processor,
Intel Pentium® II processor, and Intel Celeron™ processor. The GMCH2-M implements the host
address, control, and data bus interfaces within a single device. The GMCH2-M supports a 4-deep inorder queue (i.e., supports pipelining of up to 4 outstanding transaction requests on the host bus). Host
bus addresses are decoded by the GMCH2-M for accesses to system memory, PCI memory & PCI I/O
(via hub interface), PCI configuration space and Graphics memory. The GMCH2-M takes advantage of
the pipelined addressing capability of the processor to improve the overall system performance.
4.8.1.
Host Bus Device Support
The GMCH2-M supports a large subset of the transaction types that are defined for the Intel Pentium®
III processor, Intel Pentium® II processor, or Intel Celeron™ processor bus interface. However, each of
these transaction types has a multitude of response types, some of which are not supported by this
controller. All transactions are processed in the order that they are received on the processor bus.
Table 12. Summay of Transactions Supported By GMCH2-M
Transaction
Datasheet
REQa[4:0]#
REQb[4:0]#
GMCH2-M Support
Deferred Reply
00000
XXXXX
The GMCH2-M will initiate a deferred reply request for a
previously deferred transaction.
Reserved
00001
XXXXX
Reserved
Interrupt
Acknowledge
01000
00000
Interrupt acknowledge cycles are forwarded to the hub
interface.
Special
Transactions
01000
00001
See separate table in special cycles section.
Reserved
01000
0001x
Reserved
Reserved
01000
001xx
Reserved
Branch Trace
Message
01001
00000
The GMCH2-M will terminate a branch trace message
without latching data.
Reserved
01001
00001
Reserved
Reserved
01001
0001x
Reserved
Reserved
01001
001xx
Reserved
I/O Read
10000
0 0 x LEN#
I/O read cycles are forwarded to hub interface. I/O cycles
which are in the GMCH2-M configuration space will not be
forwarded to the hub interface.
I/O Write
10001
0 0 x LEN#
I/O write cycles are forwarded to hub interface. I/O cycles
which are in the GMCH2-M configuration space will not be
forwarded to hub interface.
Reserved
1100x
00xxx
Reserved
Memory Read &
Invalidate
00010
0 0 x LEN#
Host initiated memory read cycles are forwarded to DRAM
or the hub interface.
Reserved
00011
0 0 x LEN#
Reserved
125
®
Intel 82815EM GMCH
R
Transaction
REQa[4:0]#
REQb[4:0]#
GMCH2-M Support
Memory Code
Read
00100
0 0 x LEN#
Memory code read cycles are forwarded to DRAM or hub
interface.
Memory Data
Read
00110
0 0 x LEN#
Host initiated memory read cycles are forwarded to DRAM
or the hub interface.
Memory Write
(no retry)
00101
0 0 x LEN#
This memory write is a writeback cycle and cannot be
retried. The GMCH2-M will forward the write to DRAM.
Memory Write
(can be retried)
00111
0 0 x LEN#
The standard memory write cycle will be forwarded to
DRAM or hub interface.
For Memory cycles, REQa[4:3]# = ASZ#. GMCH2-M only supports ASZ# = 00 (32 bit address).
REQb[4:3]# = DSZ#. DSZ# = 00 (64 bit data bus size).
LEN# = data transfer length as follows:
LEN#
Data length
00
<= 8 bytes (BE[7:0]# specify granularity)
01
Length = 16 bytes BE[7:0]# all active
10
Length = 32 bytes BE[7:0]# all active
11
Reserved
Table 13. Host Responses Supported by the GMCH2-M
RS2#
126
RS1#
RS0#
Description
GMCH2-M Support
0
0
0
idle
0
0
1
Retry
Response
This response is generated if an access is to a resource that
cannot be accessed by the processor at this time and the logic
must avoid deadlock. Hub interface directed reads, writes, and
DRAM locked reads can be retried.
0
1
0
Deferred
Response
This response can be returned for all transactions that can be
executed ‘out of order.’ Hub interface directed reads (memory,
I/O and Interrupt Acknowledge) and writes (I/O only), and
internal Graphics device directed reads (memory and I/O) and
writes (I/O only) can be deferred.
0
1
1
Reserved
Reserved
1
0
0
Hard Failure
Not supported.
1
0
1
No Data
Response
This is for transactions where the data has already been
transferred or for transactions where no data is transferred.
Writes and zero length reads receive this response.
1
1
0
Implicit
Writeback
This response is given for those transactions where the initial
transactions snoop hits on a modified cache line.
1
1
1
Normal Data
Response
This response is for transactions where data accompanies the
response phase. Reads receive this response.
Datasheet
®
Intel 82815EM GMCH
R
4.8.2.
Special Cycles
A Special Cycle is defined when REQa [4:0] = 01000 and REQb [4:0]= xx001. The first address phase
Aa [35:3]# is undefined and can be driven to any value. The second address phase, Ab [15:8]# defines
the type of Special Cycle issued by the processor.
The following table specifies the cycle type and definition, as well as the action taken by the GMCH2-M
when the corresponding cycles are identified.
Table 14. Special Cycles
BE[7:0}#
Datasheet
Special Cycle Type
Action Taken
0000 0000
NOP
This transaction has no side-effects.
0000 0001
Shutdown
This transaction is issued when an agent detects a severe software error
that prevents further processing. This cycle is claimed by the GMCH2-M.
The GMCH2-M issues a shutdown special cycle on the hub interface.
This cycle is retried on the processor bus after it is terminated on the hub
interface via a master abort mechanism.
0000 0010
Flush
This transaction is issued when an agent has invalidated its internal
caches without writing back any modified lines. The GMCH2-M claims
this cycle and retries it.
0000 0011
Halt
This transaction is issued when an agent executes a HLT instruction and
stops program execution. This cycle is claimed by the GMCH2-M and
propagated to the hub interface as a Special Halt Cycle. This cycle is
retried on the processor bus after it is terminated on the hub interface via
a master abort mechanism.
0000 0100
Sync
This transaction is issued when an agent has written back all modified
lines and has invalidated its internal caches. The GMCH2-M claims this
cycle and retries it.
0000 0101
Flush Acknowledge
This transaction is issued when an agent has completed a cache sync
and flush operation in response to an earlier FLUSH# signal assertion.
The GMCH2-M claims this cycle and retires it.
0000 0110
Stop Clock
Acknowledge
This transaction is issued when an agent enters Stop Clock mode. This
cycle is claimed by the GMCH2-M and propagated to the hub interface as
a Special Stop Grant Cycle. This cycle is completed on the processor
bus after it is terminated on the hub interface via a master abort
mechanism.
0000 0111
SMI Acknowledge
This transaction is first issued when an agent enters the System
Management Mode (SMM). Ab[7]# is also set at this entry point. All
subsequent transactions from the processor with Ab[7]# set are treated
by the GMCH2-M as accesses to the SMM space. No corresponding
cycle is propagated to the hub interface. To exit the System Management
Mode the processor issues another one of these cycles with the Ab[7]#
bit deasserted. The SMM space access is closed by the GMCH2-M at
this point.
all others
Reserved
127
®
Intel 82815EM GMCH
R
4.9.
System Memory DRAM Interface
The GMCH2-M integrates a system DRAM controller that supports a 64-bit DRAM array. The DRAM
type supported is Synchronous (SDRAM). The GMCH2-M generates the SCSA#, SCSB#, SDQM,
SCAS#, SRAS#, SWE# and multiplexed addresses, SMA for the DRAM array. The GMCH2-M’s
DRAM interface operates at a clock frequency of 100MHz, dependent upon the system bus interface
clock frequency. The DRAM controller interface is fully configurable through a set of control registers.
Complete descriptions of these registers are given in the register description section of this document.
The GMCH2-M supports industry standard 64-bit wide SO-DIMM modules with SDRAM devices. The
2 bank select lines SBS[1:0], the 12 Address lines SMAA[11:0], and copies of 4 Address lines
SMAB[7:4]# and SMAC[7:4]# allow the GMCH2-M to support 64 bit wide SO-DIMMs using 16Mb,
64Mb, 128Mb or 256Mb technology SDRAMs. The GMCH2-M has a sufficient amount of SCS# lines
to enable the support of up to six 64-bit rows of DRAM. For write operations of less than a Qword in
size, the GMCH2-M will perform a byte-wise write. The GMCH2-M targets SDRAM with CL2 and CL3
and supports both single and double-sided SO-DIMMs. The GMCH2-M provides refresh functionality
with programmable rate (normal DRAM rate is 1 refresh/15.6 µs). The GMCH2-M can be configured via
the Page Closing Policy Bit in the GMCH2-M Configuration Register to keep multiple pages open within
the memory array. Pages can be kept open in any one row of memory. Up to 4 pages can be kept open
within that row (The GMCH2-M only supports 4 Bank SDRAMs on system DRAM interface).
4.9.1.
DRAM Organization and Configuration
The GMCH2-M supports 64-bit SDRAM configurations. In the following discussion the term row refers
to a set of memory devices that are simultaneously selected by a SCS# signal. The GMCH2-M will
support a maximum of 6 rows of memory.
The interface consists of the following pins:
Multiple copies:
SMAA[7:4], SMAB[7:4]# , SMAC[7:4]#
Single Copies:
SMD[63:0]
SDQM[7:0]
SMAA[12:8,3:0]
SBS[1:0]
SCSA[5:0]#
SCSB[5:0]#
SCAS#
SRAS#
SWE#
SCKE[5:0]
128
Datasheet
®
Intel 82815EM GMCH
R
The GMCH2-M supports SO-DIMMs with 16 bit wide SDRAM devices. Registered DIMMs or DIMMs
populated with 4 bit wide SDRAM devices are not supported. The GMCH2-M supports 3.3v standard
SDRAMs.
4.9.1.1.
Configuration Mechanism SO-DIMMs
Detection of the type of DRAM installed on the SO-DIMM is supported via Serial Presence Detect
mechanism as defined in the JEDEC 168-pin DIMM standard and 144-pin SO-DIMM standard. This
standard uses the SCL, SDA and SA[2:0] pins on the SO-DIMMs to detect the type and size of the
installed SO-DIMMs. No special programmable modes are provided on the GMCH2-M for detecting the
size and type of memory installed. Type and size detection must be done via the serial presence detection
pins. Use of Serial Presence Detection is required.
Memory Detection and Initialization
Before any cycles to the memory interface can be supported, the GMCH2-M DRAM registers must be
initialized. The GMCH2-M must be configured for operation with the installed memory types. Detection
of memory type and size is done via the System Management Bus (SMBus) interface on the ICH2-M.
This two wire bus is used to extract the DRAM type and size information from the serial presence detect
port on the DRAM SO-DIMM modules.
4.9.1.2.
DRAM Register Programming
This section provides an overview of how the required information for programming the DRAM registers
is obtained from the Serial Presence Detect ports on the SO-DIMMs. The Serial Presence Detect ports
are used to determine Refresh Rate, MA and MD Buffer Strength, Row Type (on a row by row basis),
SDRAM Timings, Row Sizes and Row Page Sizes.
The following Table 5-6 lists a subset of the data available through the on-board Serial Presence Detect
ROM on each SO-DIMM module.
Table 15. Data Bytes on DIMM Used for Programming DRAM Registers
Byte
Function
2
Memory Type (EDO, SDRAM) the GMCH2-M only supports SDRAM.
3
# of Row Addresses, not counting Bank Addresses
4
# of Column Addresses
5
# of banks of DRAM (Single or Double sided) DIMM
12
Refresh Rate
17
# Banks on each SDRAM Device
36-41
Access Time from Clock for CAS# Latency 1 through 7
42
Data Width of SDRAM Components
These bytes collectively provide enough data for BIOS to program the GMCH2-M DRAM registers.
Datasheet
129
®
Intel 82815EM GMCH
R
4.9.2.
DRAM Address Translation and Decoding
The GMCH2-M contains address decoders that translate the address received on the host bus, hub
interface, or from the internal Graphics device to an effective memory address. The GMCH2-M supports
16 and 64 Mbit SDRAM devices. The GMCH2-M supports a 2 KB page sizes only. The multiplexed row
/ column address to the DRAM memory array is provided by the SBS[1:0] and SMAA[11:0] signals and
copies. These addresses are derived from the host address bus as defined by the following table for
SDRAM devices. Row size is internally computed using the values programmed in the DRP register.
Up to 4 pages can be open at any time within any row (Only 2 active pages are supported in rows
populated with either 8 MBs or 16 MBs).
13
10
9
2
2
2
128
64
256
128
MA
16
13
9
2
32
MA
8
12
10
2
64
MA
16
12
8
2
MA
8
12
9
MA
16
12
MA
16M
8
16
MA
256
32M
1
MA
256
8M
9
MA
128
16M
11
MA
128
4M
8
(MB)
MA
64
8M
Bank
MA
64
Col
MA
2M
Row
MemS
ize
BS
16
Address Usage
Width
(Mb)
Depth
Tech
BS
Table 16. GMCH2-M DRAM Address Mux Function
1
0
12
11
10
9
8
7
6
5
4
3
2
1
0
X
11
X
X
[A]
22
21
20
19
18
17
16
15
14
13
X
11
X
X
PA
X
23
10
9
8
7
6
5
4
3
12
11
X
24
[A]
22
21
20
19
18
17
16
15
14
13
12
11
X
X
PA
X
25
10
9
8
7
6
5
4
3
12
11
X
24
[A]
22
21
20
19
18
17
16
15
14
13
12
11
X
X
PA
X
X
10
9
8
7
6
5
4
3
12
11
X
24
[A]
22
21
20
19
18
17
16
15
14
13
12
11
X
X
PA
26
25
10
9
8
7
6
5
4
3
12
11
X
24
[A]
22
21
20
19
18
17
16
15
14
13
12
11
X
X
PA
X
25
10
9
8
7
6
5
4
3
12
11
27
24
[A]
22
21
20
19
18
17
16
15
14
13
12
11
X
X
PA
26
25
10
9
8
7
6
5
4
3
12
11
26
24
[A]
22
21
20
19
18
17
16
15
14
13
12
11
X
X
PA
X
25
10
9
8
7
6
5
4
3
NOTES: MA bit 10 at RAS time uses the XOR of Address bit 12 and Address bit 23
130
Datasheet
®
Intel 82815EM GMCH
R
4.9.3.
SDRAMT Register Programming
Several DRAM timing parameters are programmable in the GMCH2-M configuration registers. Table 18
summarizes the programmable parameters.
Table 17. Programmable SDRAM Timing Parameters
Parameter
DRAMT Bit
Values (SCLKs)
RAS# Precharge (SRP)
0
2,3
RAS# to CAS# Delay (SRCD)
1
2,3
CAS# Latency (CL)
2
2,3
DRAM Cycle Time (DCT)
4
Tras = 5,6
Trc = 7,8
These parameters are controlled via the DRAMT register. In order to support different device speed
grades, CAS# Latency, RAS# to CAS# Delay, and RAS# Precharge are all programmable as either two
or three SCLKs. To provide flexibility, these are each controlled by separate register bits. That is, the
GMCH2-M can support any combination of CAS# Latency, RAS# to CAS# Delay and RAS# Precharge.
4.9.4.
SDRAM Paging Policy
The GMCH2-M can maintain up to 4 active pages in any one row; however, the GMCH2-M does not
support active pages in more than 1 row at a time.
The DRAM page closing policy (DPCP) in the GMCH2-M configuration register (GMCHCFG) controls
the page closing policy of the. This bit controls whether the GMCH2-M will precharge bank or
precharge all during the service of a page miss. When this bit is set to 0, the GMCH2-M will prechange
bank during the service of a page miss. When this bit is set to 1, the GMCH2-M will prechange all during
the service of a page miss.
4.10.
Intelâ Dynamic Video Memory Technology (D.V.M.T.)
The internal graphics device on the GMCH2-M supports Intelâ Dynamic Video Memory Technology
(D.V.M.T.). Intel® D.V.M.T. dynamically responds to application requirements by allocating the proper
amount of display and texturing memory. For more details please refer to the document entitled, “Intel
810 chipset: Great Performance for Value PCs” available at: http://deverlolper.intel.com/design/
chipsets/ 810/810white.htm.
In addition to Intel® D.V.M.T., the GMCH2-M supports Display Cache (DC). The graphics engine of the
GMCH2-M uses DC for implementing rendering buffers (e.g., Z buffers). This rendering model requires
4 MB of display cache and allows graphics rendering (performed across the graphics display cache bus)
and texture MIP map access (performed across the system memory bus) simultaneously. In using Intel®
D.V.M.T., all graphics rendering is implemented in system memory. The system memory bus is
arbitrated between texture MIP-map accesses and rendering functions.
Datasheet
131
®
Intel 82815EM GMCH
R
4.11.
Display Cache Interface
The GMCH2-M Display Cache (DC) is a single channel 32 bit wide SDRAM interface. The GMCH2-M
handles the control and timing for the display cache. The display cache interface of the GMCH2-M
generates the LCS#, LDQM[7:0], LSCAS#, LSRAS#, LWE#, LMD[31:0] and multiplexed addresses,
LMA[11:0] for the display cache DRAM array. The GMCH2-M also generates the clock LTCLK[1:0]
for write cycles as well as LOCLK for read cycle timings.
The display cache interface of the GMCH2-M supports single data rate synchronous dynamic random
access memory (SDRAM). It supports a single 32-bit wide memory channel. The interface handles the
operation of D.V.M. with DC at 100/133MHz. The DRAM controller interface is fully configurable
through a set of control registers.
Internal buffering (FIFOs) of the data to and from the display cache ensures the synchronization of the
data to the internal pipelines. The D.V.M. with DC interface clocking is divided synchronous with
respect to the core and system bus.
The startup sequencing for the local memory display cache, is as follows:
• System BIOS detects if an external AGP device is present by doing a config read to PCI. If an AGP
device is present, it becomes the display device and bit 0 of the APCONT register should be set to
0. No further initialization of internal graphics will take place. If internal graphics is the preferred
display device, bit 0 of the APCONT register should be set to 1. If no AGP device is present, the
internal graphics becomes the display device and bit 0 of the APCONT register should be set to 1.
PCI enumeration takes place at this point.
In the case where internal graphics is selected, the remaining steps still apply:
• System BIOS determines if local memory display cache is present. If present the following steps
take place:
 Local Memory Clock Frequency is determined with a reset strap (on AGP pin SBA[7])
sampled as an input during reset.
 Memory Timing Options will be determined empirically by the system BIOS. The BIOS will
start with programming slow timings (CAS Latency, RAS Pre-charge, etc.) and then trying
faster timings until it breaks. The settings, which optimize performance without compromising
functionality, will be selected.
4.11.1.
Supported DRAM Types for Display Cache Memory
1Mx16 and 2Mx32 SDRAMs are supported, but the maximum memory supported is 4MB of display
cache memory.
132
Datasheet
®
Intel 82815EM GMCH
R
4.11.2.
Memory Configurations
The following table gives a summary of the characteristics of memory configurations supported. The
GMCH2-M supports a 32-bit wide channel populated with a single row of 1Mx16 SDRAMs.
Table 18. Memory Size for each configuration:
SDRAM
SDRAM
SDRAM
# of
Address Size
DRAM
Tech.
Density
Width
Banks Bank
Row
Column
Addressing
16 Mbit
1M
16
2
11
8
Asymmetric
16 Mbit
1M
32
1
DRAM Size
4MB
The following figure shows the GMCH LMI connected to 4 MB of memory in a 32-bit SDRAM channel
configuration.
Figure 6. MCH2-M Display Cache Interface to 4MB
GMCH
LCS#
LSRAS
LSCAS
LWE#
LMA[11:0]
LMD[31:0]
LDQM[3:0]
LTCLK[1:0]
LRCLK
LOCLK
(2) 1M x 16
SDRAM
disp_ca_interface
4.11.3.
Address Translation
The GMCH2-M contains address decoders that translate the address received by the display cache into
an effective display cache address. The LMA[11:0] bits are as defined in the following table. Entries in
the table (e.g., A21(X)) imply that the GMCH2-M puts out A21 on that MA line but it is not used by the
SDRAM.
Datasheet
133
®
Intel 82815EM GMCH
R
Table 19, GMCH2-M Local Memory Address Mapping
MA
1Mx16
Row
11(BA)
A10
A10
10
A11
X
9
A21
X
8
A20
X
7
A19
A9
6
A18
A8
5
A17
A7
4
A16
A6
3
A15
A5
2
A14
A4
1
A13
A3
0
A12
A2
NOTES:
4.11.4.
Column
BA = Bank address
Display Cache Interface Timing
The GMCH2-M provides a variety of programmable wait states for DRAM read and write cycles. These
options are programmed in the display cache I/O addresses of the GMCH2-M configuration space. The
wrap type and the burst length is implied since they are not programmable and fixed. Only sequential
wrap is allowed. Burst length is fixed at two.
4.12.
4.12.1.
Internal Graphics Device
3D/2D Instruction Processing
The GMCH2-M contains an extensive set of instructions that controls various functions including 3D
rendering, BLT and STRBLT operations, display, motion compensation, and overlay. The 3D
instructions set 3D pipeline states and control the processing functions. The 2D instructions provide an
efficient method for invoking BLT and STRBLT operations.
The graphics controller executes instructions from one of two instruction buffers located in system
memory: Interrupt Ring or Low Priority Ring. Instead of writing instructions directly to the GMCH2-M’s
graphics controller, software sets up instruction packets in these memory buffers and then instructs the
GMCH2-M to process the buffers. The GMCH2-M uses DMAs to put the instructions into its FIFO and
executes them. Instruction flow in the ring buffer instruction stream can make calls to other buffers, much
like a software program makes subroutine calls. Flexibility has been built into the ring operation
permitting software to efficiently maintain a steady flow of instructions.
Batching instructions in memory ahead of time and then instructing the graphics controller to process the
instructions provides significant performance advantages over writing directly to FIFOs including: 1)
134
Datasheet
®
Intel 82815EM GMCH
R
Reduced software overhead, 2) Efficient DMA instruction fetches from graphics memory, and 3)
Software can more efficiently set up instruction packets in buffers in graphics memory (faster than
writing to FIFOs).
Figure 7. 3D/2D Pipeline Preprocessor
Low Priority Ring
(Graphics Memory)
Batch Buffers
Instruction
Batch Buff Instr
Instruction
DMA
Instruction access and decoding
2D Instructions
DMA
FIFO
Interrupt Ring
(Graphics Memory)
Batch Buffers
Instr
Parser
3D Instructions (3D state,
3D Primitives, STRBLT,
Motion Compensation)
Instruction
DMA
BLT
Engine
3D
Engine
Batch Buff Instr
Instruction
Display
Engine
Overlay
Engine
cmd_str.vsd
4.12.2.
3D Engine
The 3D engine of the GMCH2-M has been architected as a deep pipeline, where performance is
maximized by allowing each stage of the pipeline to simultaneously operate on different primitives or
portions of the same primitive. The internal graphics device of the GMCH2-M supports perspectivecorrect texture mapping, bilinear, trilinear, and anisotropic MIP map filtering, Gouraud shading, alphablending, colorkeying and chromakeying, full color specular shading, fogging and Z Buffering. These
features can be independently enabled or disabled via set of 3D instructions. In addition, the GMCH2-M
supports a Dynamic Video Memory Technology (D.V.M.T.) which allows the entire 3D rendering
process to take place in system memory; thus, alleviating the need for the display cache.
The main blocks of the pipeline are the Setup Engine, Scan Converter, Texture Pipeline, and Color
Calculator block. A typical programming sequence would be to send instructions to set the state of the
pipeline followed by rendering instructions containing 3D primitive vertex data.
4.12.3.
Buffers
The 2D, 3D and video capabilities of the internal graphics device of the GMCH2-M provide control over
a variety of graphics buffers which can be implemented either in display cache or system memory. To aid
the rendering process, the display cache of the GMCH2-M contains two hardware buffers—the Front
Buffer (display buffer) and the Back Buffer (rendering buffer). The image being drawn is not visible until
the scene is complete and the back buffer made visible (via an instruction) or copied to the front buffer
(via a 2D BLT operation). By rendering to one and displaying from the other, the possibility of image
tearing is removed. This also speeds up the display process over a single buffer.
Datasheet
135
®
Intel 82815EM GMCH
R
The 3D pipeline of the GMCH2-M operates on the Back Buffer and the Z Buffer. The pixels’ 16-bit (or
15-bit) RGB colors are stored in the back buffer. The Z-buffer can be used to store 16-bit depth values or
5-bit “destination alpha” values. The Instruction set of the GMCH2-M provides a variety of controls for
the buffers (e.g., initializing, flip, clear, etc.).
Figure 8. Data Flow for the 3D Pipeline
GMCH Graphics Pipeline
(Conceptual Representation)
System Memory
Instructions and
Data
Textures
Setup
Frame Buffer
Discard
(back Face
culling)
Primitives
Pixels
Mapping
Engine
Display Cache
Color
Calculator
Rasterize
GMCH
Interface
Depth Buffer
(Z-Buffer)
Notes:
1. Frame Buffer = Front and Back Buffers
3d_pipe1.vsd
.
4.12.4.
Setup
The setup stage of the pipeline takes the input data associated with each vertex of the line or triangle
primitive and computes the various parameters required for scan conversion. In formatting this data, the
GMCH2-M maintains sub-pixel accuracy. Data is dynamically formatted for each rendered polygon and
output to the proper processing unit. As part of the setup, the GMCH2-M removes polygons from further
processing, if they are not facing the user’s viewpoint (referred to as “ Back Face Culling”).
4.12.5.
Texturing
The internal graphics device of the GMCH2-M allows an image, pattern, or video to be placed on the
surface of a 3D polygon. Textures must be located in system memory. Being able to use textures directly
from system memory means that large complex textures can easily be handled without the limitations
imposed by the traditional approach of only using the display cache.
136
Datasheet
®
Intel 82815EM GMCH
R
The texture processor receives the texture coordinate information from the setup engine and the texture
blend information from the scan converter. The texture processor performs texture color or chroma-key
matching, texture filtering (anisotropic, bilinear, and trilinear interpolation), and YUV to RGB
conversions.
The GMCH2-M supports up to 11 Levels-of-Detail (LODs) ranging in size from 1024x1024 to 1x1
texels. (A texel is defined as a texture map pixel). Textures need not be square. Included in the texture
processor is a small cache, which provides efficient mip-mapping.
• Nearest. Texel with coordinates nearest to the desired pixel is used. (This is used if only one LOD
is present).
• Linear. A weighted average of a 2x2 area of texels surrounding the desired pixel are used. (This is
used if only one LOD is present).
• Mip Nearest. This is used if many LODs are present. The appropriate LOD is chosen and the texel
with coordinates nearest to the desired pixel are used.
• Mip Linear. This is used if many LODs are present. The appropriate LOD is chosen and a weighted
average of a 2x2 area of texels surrounding the desired pixel are used. This is also referred to as bilinear mip-mapping.
• Trilinear. Tri-linear filtering blends two mip maps of the same image to provide a smooth transition
between different mips (floor and ceiling of the calculated LOD).
• Anisotropic. This can be used if multiple LODs are present. This filtering method improves the
visual quality of texture-mapped objects when viewed at oblique angles (i.e., with a high degree of
perspective foreshortening). The improvement comes from a more accurate (anisotropic) mapping
of screen pixels onto texels -- where using bilinear or trilinear filtering can yield overly-blurred
results. Situations where anisotropic filtering demonstrates superior quality include text viewed at
an angle, lines on roadways, etc.
The GMCH2-M can store each of the above mip-maps in any of the following formats:
• 8bpt Surface Format
• 16bpt Surface Format
 RGB 565
 ARGB 1555
 ARGB 4444
 AY 88
• 8bpt (Indexed) Surface Format
 RGB 565
 ARGB 1555
 ARGB 4444
 AY 88
• 4:2:2
 YCrCb, Swap Y Format
 YCrCb, Normal
 YCrCb, UV Swap
 YCrCb, UV/Y Swap
Many texture mapping modes are supported. Perspective correct mapping is always performed. As the
map is fitted across the polygon, the map can be tiled, mirrored in either the U or V directions, or
Datasheet
137
®
Intel 82815EM GMCH
R
mapped up to the end of the texture and no longer placed on the object (this is known as clamp mode).
The way a texture is combined with other object attributes is also definable.
Texture ColorKey and ChromaKey
ColorKey and ChromaKey describe two methods of removing a specific color or range of colors from a
texture map before it is applied to an object. For “nearest” texture filter modes, removing a color simply
makes those portions of the object transparent (the previous contents of the back buffer show through).
For “ linear “ texture filtering modes, the texture filter is modified if only the non-nearest neighbor texels
match the key (range).
ColorKeying occurs with paletted textures, and removes colors according to an index (before the palette
is accessed). When a color palette is used with indices to indicate a color in the palette, the indices can be
compared against a state variable “ColorKey Index Value” and if a match occurs and ColorKey is
enabled, then this value’s contribution is removed from the resulting pixel color. The GMCH2-M defines
index matching as ColorKey.
ChromaKeying can be performed for both paletted and non-paletted textures, and removes texels that fall
within a specified color range. The ChromaKey mode refers to testing the RGB or YUV components to
see if they fall between high and low state variable values. If the color of a texel contribution is in this
range and ChromaKey is enabled, then this contribution is removed from the resulting pixel color.
Multiple Texture Composition
The GMCH2-M includes support for two simultaneous texture maps. This support greatly reduces the
need for multipass compositing techniques for effects such as diffuse light maps, specular reflection
maps, bump mapping, detail textures, gloss maps, shadows, and composited effects like dirt or tire
marks. Supporting these techniques in hardware greatly increases compositing performance by reducing
the need to read and write the frame buffer multiple times.
This multitexture support provides a superset of the “legacy” one-texture (pre-DirectX 6) texture blend
modes and a large subset of the operations defined in DirectX 6 and the OpenGL ARB multitexture
extensions.
The Multitexture Compositing Unit is capable of combining the interpolated vertex diffuse color, a
constant color value, and up to two texels per pixel in a fully programmable fashion. Up to three
operations (combinations) can be performed in a pipelined organization, with intermediate storage to
support complex equations, e.g., of the form “A*B + C*D” required for light maps and specular gloss
maps. Separate operations can be performed on color (RGB) and alpha components.
4.12.6.
2D Operation
The GMCH2-M contains BLT and STRBLT functionality, a hardware cursor, and an extensive set of 2D
registers and instructions.
GMCH2-M VGA Registers and Enhancements
The 2D registers are a combination of registers defined by IBM* when the Video Graphics Array (VGA)
was first introduced, and others that Intel® has added to support graphics modes that have color depths,
resolutions, and hardware acceleration features that go beyond the original VGA standard. The internal
graphics device of the GMCH2-M improves upon VGA by providing additional features that are used
through numerous additional registers.
138
Datasheet
®
Intel 82815EM GMCH
R
The GMCH2-M also supports an optional display cache. As an improvement on the VGA standard
display cache port-hole, the GMCH2-M also maps the entire display cache into part of a single
contiguous memory space at a programmable location, providing what is called “linear” access to the
display cache. The size of this memory can be up to 4 MB, and the base address is set via PCI
configuration registers. Alternatively, these buffers may be implemented in system memory (via
D.V.M.), thus alleviating the need for the display cache.
4.12.7.
Fixed Blitter (BLT) and Stretch Blitter (STRBLT) Engines
The GMCH2-M‘s 64-bit BLT engine provides hardware acceleration for many common Windows*
operations. The following are two primary BLT functions: Fixed Blitter (BLT) and Stretch Blitter
(STRBLT). The term BLT refers to a block transfer of pixel data between memory locations. The word
“fixed” is used to differentiate from the Stretch BLT engine. The BLT engine can be used for the
following:
• Move rectangular blocks of data between memory locations
• Data Alignment
• Perform logical operations
The internal graphics device of the GMCH2-M has instructions to invoke BLT and STRBLT operations,
permitting software to set up instruction buffers and use batch processing as described in the 3D/2D
Instruction Processing (Pipeline Preprocessor) Section. Note that these instructions replace the need to
do PIO directly to BLT and STRBLT registers, which speeds up the operation.
4.12.7.1.
Fixed BLT Engine
The rectangular block of data does not change as it is transferred between memory locations. The
allowable memory transfers are between: system memory and display cache, display cache and display
cache, and system memory and system memory. Data to be transferred can consist of regions of memory,
patterns, or solid color fills. A pattern will always be 8x8 pixels wide and may be 8, 16, or 24 bits per
pixel.
The internal graphics device of the GMCH2-M has the ability to expand monochrome data into a color
depth of 8, 16, or 24 bits. BLTs can be either opaque or transparent. Opaque transfers, move the data
specified to the destination. Transparent transfers, compare destination color to source color and write
according to the mode of transparency selected.
Data is horizontally and vertically aligned at the destination. If the destination for the BLT overlaps with
the source memory location, the GMCH2-M can specify which area in memory to begin the BLT
transfer. Use of this BLT engine accelerates the Graphical User Interface (GUI) of Microsoft* Windows.
Hardware is included for all 256 raster operations (Source, Pattern, and Destination) defined by
Microsoft*, including transparent BLT.
4.12.7.2.
Arithmetic Stretch BLT Engine
The stretch BLT function can stretch source data in the X and Y directions to a destination larger or
smaller than the source. Stretch BLT functionality expands a region of memory into a larger or smaller
region using replication and interpolation.
The stretch BLT engine also provides format conversion and data alignment. Through an algorithm
implemented in the mapping engine, object expansion and contraction can occur in the horizontal and
vertical directions.
Datasheet
139
®
Intel 82815EM GMCH
R
4.12.8.
Hardware Motion Compensation
The Motion Compensation (MC) process consists of reconstructing a new picture by predicting (either
forward, backward or bidirectionally) the resulting pixel colors from one or more reference pictures. The
GMCH2-M intercepts the DVD pipeline at Motion Compensation and implements Motion Compensation
and subsequent steps in hardware. Performing Motion Compensation in hardware reduces the processor
demand of software-based MPEG-2 decoding, and thus improves system performance.
The GMCH2-M’s implementation of Hardware Motion Compensation supports a motion smoothing
algorithm. When the system processor is not able to process the MPEG decoding stream in a timely
manner (as can happen in software DVD implementations), the Intel® 815EM chipset GMCH2-M
supports downsampled MPEG decoding. Downsampling allows for reduced spatial resolution in the
MPEG picture while maintaining a full frame rate, and thus reduces processor load while maintaining the
best video quality possible given the processor constraints.
4.12.9.
Hardware Cursor and Popup Support
The internal graphics device of the GMCH2-M allows an unlimited number of cursor patterns to be
stored in the display cache or system memory. Two sets of registers, contain the x and y position of the
cursor relative to the upper left corner of the display. The following four cursor modes are provided:
• 32x32 2 bpp AND/XOR 2-plane mode
• 64x64 2 bpp 3-color and transparency mode
• 64x64 2 bpp AND/XOR 2-plane mode
• 64x64 2 bpp 4-color mode
Intel® 815EM chipset implements a popup in addition to the cursor. The Intel® 815EM chipset popup
and cursor have the same functionality as the Intel® 815E chipset cursor. The Intel® 815EM chipset
popup is implemented by replicating the cursor logic, it is allowable for both the cursor and the popup to
be active on the screen at the same time.
4.12.10. Overlay Engine
The overlay engine provides a method of merging either video capture data (from an external PCI Video
Capture Adapter) or data delivered by the processor, with the graphics data on the screen. Supported data
formats include YUV 4:2:2, YUV 4:2:0, YUV 4:1:0, YUV 4:1:1, RGB15, and RGB16. The source data
can be mirrored horizontally or vertically or both. Overlay data comes from a buffer located in system
memory. Additionally, the overlay engine can be quadruple buffered in order to support flipping between
different overlay images. Data can either be transferred into the overlay buffer from the host or from an
external PCI adapter, such as DVD hardware or video capture hardware. Buffer swaps can be done by
the host and internally synchronized with the display VBLANK.
The internal graphics device of the GMCH2-M can accept line widths up to 720 pixels. In addition,
overlay source and destination ChromaKeying are also supported. Overlay source/destination
ChromaKeying enables blending of the overlay with the underlying graphics background. Destination
color/chroma keying can be used to handle occluded portions of the overlay window on a pixel by pixel
basis, which is actually an underlay. Source color/chroma keying is used to handle transparency based on
the overlay window on a pixel by pixel basis. This is used when “blue screening” an image in order to
overlay the image on a new background later.
140
Datasheet
®
Intel 82815EM GMCH
R
To compensate for overlay color intensity loss due to the non-linear response between display devices,
the overlay engine supports independent gamma correction. In addition, the brightness, saturation, and
contrast of the overlay may be independently varied.
4.12.11. Display
The display function contains a RAM-based Digital-to-Analog Converter (RAMDAC) that transforms
the digital data from the graphics and video subsystems to analog data for the monitor. The GMCH2-M’s
integrated 230MHz RAMDAC provides resolution support up to 1600x1200. Circuitry is incorporated to
limit the switching noise generated by the DACs. Three 8-bit DACs provide the R, G, and B signals to
the monitor. Sync signals are properly delayed to match any delays from the D-to-A conversion.
Associated with each DAC is a 256 pallet of colors. The RAMDAC can be operated in either direct or
indexed color mode. In Direct color mode, pixel depths of 15, 16, or 24 bits can be realized. Noninterlaced mode is supported. Gamma correction can be applied to the display output.
The GMCH2-M supports a wide range of resolutions, color depths, and refresh rates via a programmable
dot clock that has a maximum frequency of 230MHz.
Table 20. Partial List of Display Modes Supported
Bits Per Pixel (frequency in Hz)
Resolution
Datasheet
8-bit Indexed
16-bit
24-bit
320x200
70
70
70
320x240
70
70
70
352x480
70
70
70
352x576
70
70
70
400x300
70
70
70
512x384
70
70
70
640x400
70
70
70
640x480
60,70,72,75,85
60,70,72,75,85
60,70,72,75,85
720x480
75,85
75,85
75,85
720x576
60,75,85
60,75,85
60,75,85
800x600
60,70,72,75,85
60,70,72,75,85
60,70,72,75,85
1024x768
60, 70,72,75,85
60, 70,72,75,85
60, 70,72,75,85
1152x864
60,70,72,75,85
60,70,72,75,85
60,70,72,75,85
1280x720
60,75,85
60,75,85
60,75,85
1280x960
60,75,85
60,75,85
60,75,85
1280x1024
60,70,72,75,85
60,70,72,75,85
60,70,75,85
1600x900
60,75,85
60,75,85
1600x1200
60,70,72,75
141
®
Intel 82815EM GMCH
R
4.12.12. Digital Video Out (DVO) Port
The Digital Video Out (DVO) port is an Intel® standard. This port is a 16 bit digital port (4 control bits
and 12 data bits). The DVO Port is a scaleable low voltage interface that ranges from nominal 1.1V to
1.8V. It consists of a single interface that can serve as either a TV-Out or Digital-Out interface.
TV-Out interface is an output of PC graphics controller to one of many possible connections to analog
video standard (NTSC or PAL) that can be connected to a TV monitor, VCR or other video devices. This
is accomplished by using a TV Encoder device that encodes the digital pixel data generated by the
graphics controller into a usable video signal. Depending on the capabilities of the TV Encoder, this may
be NTSC or PAL TV format.
The Digital-Out interface is an output of the PC graphics controller to the Digital TFT (Thin-FilmTransistor) Flat Panel or Digital CRT monitor. This is accomplished by using a TMDS (Transition
Minimized Differential Signaling) device that transmits the digital pixel data generated by the graphics
controller into a TMDS interface.
4.12.12.1. VCH interface
For mobile applications, this DVO port can be connected to the VCH chip that has integrated LVDS and
panel fitting logic. The VCH provides direct flat panel interface via LVDS (or CMOS) signaling and can
be set to bypass the DVO stream (DVOr) to support other DVO devices to the platform. When used with
the VCH, a LTVCLKIN/STALL signal is used for panel fitting flow control.
This DVO port is used to stream the primary and overlay surfaces to an LCD Display, TV, or digital
display. This is accomplished by providing the interface to the discrete TV Encoder, discrete TMDS
Transmitter, or integrated TV Encoder & TMDS Transmitter, all of which are external to the GMCH2M. Simultaneous operation of the LCD Panel and the DVO bypass is not allowed. However, switching
between TV Display and LCD Display modes is possible at any time, and will not require the device to
be powered down.
142
Datasheet
®
Intel 82815EM GMCH
R
Figure 9. Digital Video Out Port Mobile Application Block Diagram With VCH
815EM
FW82807AA
DVO bypass
Data [11:0]
DVOr signals
3rd Party TV Encoder
or Flat Panel Transmitter
or Combo Chip
Figure 10. Digital Video Out Port Block Diagram Without VCH
815EM
Analog RGB
Monitor
DVO Port
DVOrCLKIN
Differential Clock
Differential Data [11:0]
DVOrHsync
DVOrVsync
DVOrBlank#
3rd Party
TV Encoder
or
Flat Panel Transmitter
or
Combo Chip
Analog RGB
NTSC/PAL
Digital (TMDS)
LCD or
Digital CRT
INTRPT#
DVOP Blk Diag
4.12.12.2. DVO Port Data Format
The Intel® DVO port data bus width is 12-bits wide, the data is output in a 12-bit package at each edge of
a pixel clock, and uses the LTVBLANK signal to determine when active pixel data is available.
Intel® DVO port overlay input data can be YUV progressive. In this case the overlay should be set to
frame mode. Each overlay buffer contains a frame, and the overlay is flipped between them. The overlay
does not do anything special, since the data matches the typical non-interlaced CRT screens. In other
cases, where the input data is an interlaced field type data, the overlay can be set to field mode, and
interlaced scanning is bi-linearly filtered into progressive data before it is displayed.
Connecting the GMCH2-M to a flat panel transmitter is demonstrated below. For more details, refer to
The 815EM Mobile Platform Recommended Design and Debug Practices.
Datasheet
143
®
Intel 82815EM GMCH
R
The 815EM supports a variety of Flat Panel display modes and refresh rates that require up to a 65MHz
dot clock over this interface. The following table shows some of the display modes supported by the
815EM.
Table 21. Partial List of Flat Panel Modes Supported
Bits Per Pixel (frequency in Hz)
Resolution
8-bit Indexed
16-bit
24-bit
320x2001
60
60
60
1
60
60
60
1
352x480
60
60
60
352x4801
60
60
60
1
60
60
60
1
60
60
60
1
60
60
60
1
640x350
60
60
60
640x4001
60
60
60
640x480
320x240
352x576
400x300
512x384
60
60
60
1
60
60
60
1
720x576
60
60
60
800x600
60
60
60
1024x768
60
60
60
1400x1050
60
60
N/A
720x480
1
These resolutions are supported via centering.
Table 22. Partial List of TV-Out Modes Supported
Resolution
320x2001
320x240
1
352x480
1
352x576
1
400x300
144
Colors
NTSC
PAL
256
Yes
Yes
16M
Yes
Yes
64k
Yes
Yes
256
Yes
Yes
16M
Yes
Yes
64k
Yes
Yes
256
Yes
Yes
16M
Yes
Yes
64k
Yes
Yes
256
Yes
Yes
16M
Yes
Yes
64k
Yes
Yes
256
Yes
Yes
Datasheet
®
Intel 82815EM GMCH
R
Resolution
Colors
NTSC
PAL
16M
Yes
Yes
64k
Yes
Yes
640x400
256
Yes
Yes
640x480
256
Yes
Yes
64k
Yes
Yes
16M
Yes
Yes
256
Yes
Yes
64k
Yes
Yes
16M
Yes
Yes
256
Yes
Yes
64k
Yes
Yes
16M
Yes
Yes
16
Yes
Yes
256
Yes
Yes
32k
Yes
Yes
64k
Yes
Yes
16M
Yes
Yes
1
1
720x480
720x5761
800x600
These resolutions are supported via centering.
4.12.12.3. DVO Port I2C Functionality
I2C bus is a standard 2-wire communications bus/protocol. There are two I2C compatible buses in the
GMCH2-M for DDC and DVO connectivities. The I2C buses are used to collect EDID from Digital
Display Panel, and to detect and configure registers in the TV Encoder, or TMDS Transmitter chips. The
Intel® DVO port controls the video front-end devices through the I2C bus using LTVDA and LTVCK
pins. The Display Data Channel or DDC uses I2C bus via DDDA and DDCK pins.
4.12.13. DDC (Display Data Channel)
DDC is a standard defined by VESA. Its purpose is to allow communication between the host system and
display. Both configuration and control information can be exchanged allowing plug-and-play systems to
be realized. Support for DDC 2B is implemented.
4.13.
System Reset for the GMCH2-M
Refer to the Intelâ 815 Chipset Platform Design Guide (Power Sequencing section) for details.
Datasheet
145
®
Intel 82815EM GMCH
R
4.14.
System Clock Description
4.14.1.
External Clock Sources
The GMCH2-M device shall receive the following clock frequencies and references:
• 100/66 MHz host clock, used to run core logic (host, 2D, 3D).
• 66 MHz AGP/Hub Interface.
• 48 MHz reference - used to derive video dot clocks for Display and LCD/TVOut interfaces.
• 40 MHz Tvclkin from the External DVO device
The following table describes the clock alignments for the external clock sources.
Table 23. Supported Frequencies and Corresponding Phase Alignments
4.14.2.
Host
Interface
Hub / AGP
Interface
System
Memory
Local
Memory
(HCLK)
(BCLK)
(SCLK)
(LCLK)
66 Mhz
66 Mhz
100 Mhz
100 Mhz
66 Mhz
100 Mhz
100 Mhz
133Mhz
100 Mhz
133Mhz
Core
External Phase
Alignment
(RCLK)
100 Mhz
HCLK and BCLK 180
degrees out of phase
100 Mhz
HCLK and SCLK 180
degrees out of phase
Internal Clock Sources
The internal graphics device shall source the following interface frequencies and references:
1. 100 MHz for the system memory subsystem.
2. 133/100 MHz for the local memory subsystem.
3. 100 MHz for the graphics core.
4. 266/133/66 MHz for AGP in 4X/2X/1X modes respectively.
5. 66 MHz for the hub interface.
6. 20-112 MHz – LCD/TVout interface.
4.15.
146
Power Management
Datasheet
®
Intel 82815EM GMCH
R
4.15.1.
Specifications Supported
The platform is compliant with the following specifications:
• APM Rev 1.2
• ACPI Rev 1.0
• PCI Power Management Rev 1.0
• PC 99 System Design Guide, Rev 1.0
This chapter includes general description of the ACPI power states, transition rules among its states,
power states Intel® 815EM chipset supports, entering and exiting C2/C3/S1/S3 state. Intel® 815EM
chipset supports several levels of power management.
4.16.
General Description of ACPI Power States
Intel® 815EM chipset will support the ACPI compliant power states. The following table describes in
general the ACPI power states. It is for informational purposes and should not be used by designers or
validations as part of the behavioral description.
Table 24. General Description of ACPI Power States
Datasheet
G0/S0/C0
Full on: Processor is fully operating. Individual devices may be shutdown to save power.
Different processor operating levels are described in system level state transition table
below. Within the C0 state, the ICH2-M can throttle the STPCLK# signal to further reduce
power consumption.
G0/S0/C1
Auto-Halt: Processor has executed an AutoHalt instruction and is not executing code. The
processor snoops the bus and maintains cache coherency.
G0/S0/C2
Stop-Grant for Desktop or Quick-start for Mobile: The STPCLK# signal goes active to the
processor. The processor performs a Stop-Grant cycle, halts its instruction stream, and
remain in that state until the STPCLK# signal goes inactive. In the Stop-Grant state, the
processor snoops the bus and maintains cache coherency. The Quick-start state is lower
power version, but there are restrictions on what interrupt signals can go active while
STPCLK# is active. Note: Some mobile systems may use the Stop-Grant state rather than
the Quick-start state.
G0/S0/C3
Stop-Clock: This is only for mobile systems. The STPCLK# signal goes active to the
processor. The processor performs a Stop-Grant cycle and halts its instruction stream.
ICH2-M then asserts STP_CPU# that forces the clock generator to stop the processor
clock. This is also used in the mobile system for Intel SpeedStep Technology support.
G1/S1(Desktop)
Stop-Grant: Similar to G0/S0/C2 state. ICH2-M also has the option to assert CPUSLP#
signal to the processor to further reduce power consumption.
G1/S1 (Mobile)
Power-On Suspend (POS): In this state, all clocks are stopped except the 32.768 KHz
clock. The system context is maintained in the system memory. Power is on to processor,
PCI, memory controller, memory, and all other critical circuits.
G1/S3
Suspend-to-RAM (STR): The system context is maintained in system memory, but power
is shut to non-critical circuits. Memory is retained and refreshes continue. All clocks are
shut except RTC.
G1/S4
Suspend-to-Disk (STD): The system context is maintained on the disk. All power is shut
except for the logic required to resume. Externally appears same as S5, but may have
different wake events.
147
®
Intel 82815EM GMCH
R
4.17.
G2/S5
Soft Off (SOFF): System context not maintained. All power is shut except for logic
required to restart. A full boot is required when waking.
G3
Mechanical Off (MOFF): System context not maintained. All power is shut except for the
RTC. No wake events are possible. When system power returns, transition will depend on
the state just prior to the entry to G3.
Power State Transition Rules at Platform Level
The following table describes in general power states transition rules based on various trigger events.
The intermediate transitions and states are not listed in the table. It is for informational purposes and
should not be used by designers or validations as part of the behavioral description.
Table 25. State Transition Rules at Platform Level
Present State
G0/S0/C0
Next State
processor halt instruction
G0/S0/C1
ACPI defined Level 2 Read
G0/S0/C2
ACPI defined Level 3 Read
G0/S0/C3
SLP_EN bit set
G1/Sx or G2/S5 (specified by SLP_TYP)
Power Button Override
G2/S5
Mechanical Off/Power Failure
G3
Any Enabled Break Event
G0/S0/C0
STPCLK# goes asserted
G0/S0/C2
Power Button Override
G2/S5
Power Failure
G3
Any Enabled Break Event
G0/S0/C0
STPCLK# goes inactive and was in C1
G0/S0/C1
Power Button Override
G2/S5
Power Failure
G3
Any Enabled Break Event
G0/S0/C0
Power Button Override
G2/S5
Power Failure
G3
G1/S1,G1/S3,
Any Enabled Break Event
G0/S0/C0
G1/S4
Power Button Override
G2/S5
Power Failure
G3
Any Enabled Break Event
G0/S0/C0
Power Failure
G3
Power Returns
G0/S0/C0 reboot or G2/S5 till power button
pressed / other wake event.
G0/S0/C1
G0/S0/C2
G0/S0/C3
G2/S5
G3
148
Transition Trigger
Datasheet
®
Intel 82815EM GMCH
R
4.18.
ACPI Support
Advanced Configuration and Power Management Interface. ACPI primarily describes and runs
motherboard devices. It is completely controlled by the operating system that OS drivers directly power
down PCI/AGP devices. System or SMI BIOS plays a part of waking the system, however. Device
drivers save and restore state while bus drivers change the physical power state of the device.
The Intel® 815EM chipset power management architecture is designed to allow a single system to
support multiple suspend modes and to switch between those modes as required. A suspended system
can be resumed via a number of different events. It will then return to full operation where it can
continue processing or be placed into another suspend mode (potentially a lower power mode than it
resumed from).
Intel® 815EM chipset supports the minimum requirements for ACPI support. Intel® 815EM chipset must
support the minimum requirements for both system logic and for graphics controllers, as well as be
capable of controlling monitors minimum functions. The transition sequences of entering and exiting
C2/C3/S1/S3 states are described in respective sections below.
4.18.1.
Full on (C0 State)
C0 state is normal operation mode that processor executes instructions. According to typical usage
models running battery life benchmark programs, processor is in C0 about 20% of time. To save power,
the following is recommended:
If IOQ and Deferred Queue are empty, Intel® 815EM chipset shuts down GTL buffer sense amps. When
there is a need to snoop, the logic wakes up on detection of signal ADS#.
4.18.2.
Stop Grant or Quick Start (C2 State)
Stop-Grant for Desktop or Quick-start for Mobile: The STPCLK# signal goes active to the processor.
The processor performs a Stop-Grant cycle, halts its instruction stream, and remain in that state until the
STPCLK# signal goes inactive. In the Stop-Grant state, the processor snoops the bus and maintains cache
coherency. The Quick-start state is lower power version, but there are restrictions on what interrupt
signals can go active while STPCLK# is active. Note: Some mobile systems may use the Stop-Grant state
rather than the Quick-start state.
4.18.3.
Stop Clock (C3 State)
According to typical usage models running battery life programs, processor could be in C3 deep sleep
mode 64% of time. The ICH2-M performs stop clock functions with the processor directly. The
involvement of Intel® 815EM chipset is limited to the following functions: As a result of STPCLK#
assertion, the processor responds with a Stop Grant special cycle which Intel® 815EM chipset passes
down to Hub Interface to ICH2-M.
Similarly, as a result of Halt instruction, the processor performs Halt special cycle that Intel® 815EM
chipset passes to Hub Interface to ICH2-M. Intel® 815EM chipset does not prevent snoops while in the
Stop Grant State. There are a lot of power savings from Platform viewpoint.
Datasheet
149
®
Intel 82815EM GMCH
R
4.18.4.
C3 Support AGP Port Signal
This signal is generated either by the AGP Graphics Chip in AGP mode or by the Intel® 815EM chipset
in graphics mode, but not by both. This is an input to the ICH2-M I/O Controller Hub.
In AGP mode:
• Asserted (active low): When asserted, AGPBUSY# indicates that the AGP device is currently busy
and requests that the system not transition to the C3 state. However, assertion of AGPBUSY# does
not guarantee that the system will not enter the C3 state or perform an Intel SpeedStep™
technology transition. If system is in C3 state, then the assertion of AGPBUSY# is used to request
that the system exit from the C3 state.
• The AGP GC must assert AGPBUSY# whenever the AGP GC has a pending request to use the
AGP interface. The AGP GC must assert AGPBUSY# regardless of which protocol it intends to
use on the AGP interface: SBA, PIPE#, or PCI.
• AGPBUSY# may only be asserted by the AGP GC when AGPBUSY# is in the D0 state and should
not be asserted in the D1, D2, or D3 states.
• Deasserted (high): When deasserted, AGPBUSY# indicates that the AGP device is not busy and has
no need to use the AGP interface.
In Graphics mode:
• Asserted (active low): When asserted, AGPBUSY# indicates the internal graphics unit is requesting
snoop or having interrupt request to be serviced. Therefore, it requests that the system not transition
to the C3 state. However, assertion of AGPBUSY# does not guarantee that the system will not enter
the C3 state or perform an Intel SpeedStep™ technology transition. If system is in C3 state, then
the assertion of AGPBUSY# is used to request that the system exit from the C3 state.
• Deasserted (high): When deasserted, AGPBUSY# indicates the internal graphics unit has no
pending snoop request nor graphics interrupt request.
AGPBUSY#, with a minimum of 100 ns pulse width, is considered by the rest of the system to be
asynchronous to the AGP clock.
AGPBUSY# must have a 10-KΩ pull-up on the platform.
Note:
This signal is OD since it can be driven by both the 815EM Memory Controller and the AGP GC.
Note:
Please refer to Accelerated Graphics Port (AGP) BUSY and STOP signals for 815EM platform rev 0.1
reference number: SC2984.
4.18.5.
Power-on-suspend (POS) (S1 State)
S1 state is different in Desktop versus Mobile. S1-Desktop requires processor in C2 State, main memory
and all clocks are continuously running. S1-Mobile requires processor in C3 state, main memory in self
refresh mode, and all clock inputs are off so only leakage power for PLLs. S1-Mobile is called POS,
power on suspend state. Only RTC is on, all other clocks on Platform are off. All devices are in D1 or
lower state. The graphics device will be in D3 hot. Intel® 815EM chipset has to retain registers to
continue operation from where it was suspended. The ACPI OS may transition the system to S3 state if
not exiting from S1.
150
Datasheet
®
Intel 82815EM GMCH
R
When the system resumes from POS, ICH2-M can optionally resume without resetting the system, can
reset the processor only, or can reset the entire system. When no reset is performed, the ICH2-M only
needs to wait for the clock synthesizer and processor PLLs to lock before the system is resumed. This
takes typically 20ms.
4.18.6.
Suspend-to-RAM (STR) (S3 State)
S3 state is also called STR, suspend to RAM state. Power is removed from most of the system
components during STR, except the DRAM. Power is supplied to the host bridge (for DRAM Suspend
Refresh) and RTC. ICH2-M provides control signals and 32 kHz Suspend Clock (SUSCLK) to allow for
DRAM refresh and to turn off the clock synthesizer and other power planes. ICH2-M will reset the
system on resume from STR.
4.18.7.
Suspend to DISK (STD) S4 State
Power is removed from most of the system components during STD. Power is maintained to the RTC
and Suspend Well logic in the ICH2-M. ICH2-M will reset the system on resume from STD. This state
can also be called the Soft Off (SOFF) state. The difference depends on whether the system state is
restored by software to a pre-suspend condition or if the system is rebooted.
4.18.8.
Graphics Controller Requirements
PC9x implies that D0 and D3 are obligatory for graphics controllers. D0, D2, and D3 are obligatory for
monitors. Intel® 815EM chipset also permits D1 for monitors. Intel® 815EM chipset does not implement
a D1 graphics controller state.
4.18.8.1.
The D0 State
This is the normal on state for the Intel® 815EM chipset graphics functions. The Intel® 815EM chipset
graphics functions enter this state after initialization by the Video BIOS.
4.18.8.2.
The D3 State
During D3, graphics configuration space must remain accessible when power is on, but memory and I/O
space access are disabled. During D3, the driver has turned off the DAC and Intel® 815EM chipset does
not pulse HSYNC and VSYNC in this state.
When the OS decides to put the Intel® 815EM chipset graphics functions into D3, it calls the Intel®
815EM chipset graphics driver and tells it to place the device into the D3 state. The driver saves the
device context. Device context consists of the Intel® 815EM chipset graphics mode as well as local and
non-local video memory context.
Datasheet
151
®
Intel 82815EM GMCH
R
This page left intentionally blank
152
Datasheet
®
Intel 82815EM GMCH
R
5.
Pinout and Package Information
5.1.
GMCH2-M Pinout
Figure 11 and Figure 12 show the ball footprint of the GMCH2-M. These figures represent the ballout by
ball number provides an alphabetical signal listing of the ballout.
GMCH2-M is ball compatible to the Intel® 815 chipset GMCH except for the following ballouts in the
following table.
GMCH2-M ballout is subject to change prior to ballout freeze.
®
®
Table 26. Ballout differences between Intel 815 Chipset GMCH and Intel 815EM Chipset GMCH2M
PIN #
Note:
Datasheet
Intel® 815 Chipset
GMCH
Intel® 815EM Chipset
GMCH2-M
E7
VSS
AGPBUSY#
Y17
VSS
INTRPT#
AA6*
V_1.8
RESERVED
AC18
LTVCLKIN/INT#
LTVCLKIN/STALL
AA6 in Intel® 815EM chipset needs to be connected to V_1.8 for backward compatibility to Intel 815
chipset
153
®
Intel 82815EM GMCH
R
Figure 11. GMCH2-M Pinout (Top View-Left Side)
1
2
3
4
5
6
7
8
9
10
11
12
13
A
SMD_17
SMD_49
SMD_16
SMD_48
SDQM_7
SDQM_3
SDQM_2
SCSB_5#
SMAC_7#
SMAC_5#
SMAA_7
SMAA_5
SMAA_11
B
SMD_50
VSUS_3.3
SMD_56
SMD_23
VSUS_3.3
SDQM_6
SMAA_12
VSUS_3.3
SCSB_4#
SMAC_4#
VSUS_3.3
SMAA_4
SBS_0
C
SMD_18
SMD_25
VSS
SMD_55
SMD_22
VSS
SCKE_5
SCKE_4
VSS
SMAC_6#
SMAA_6
VSS
SMAA_9
D
VSS
SMD_57
SMD_26
SMD_24
SMD_54
SMD_21
SCKE_3
SCKE_0
SCSB_3#
SCSB_2#
SBS_1
SMAA_8
SMAA_0
E
SMD_51
VSUS_3.3
SMD_58
SMD_27
VSS
SMD_53
AGPBUSY#
SCKE_1
SCKE_2
VSS
SMAA_10
VSS
SCSA_4#
F
VSS
SMD_19
VSS
SMD_59
SMD_28
SMD_60
SCLK
SCSB_1#
SCSB_0#
VSUS_3.3
VSS
SMAA_2
VSS
G
ADS#
SMD_52
SMD_20
SMD_29
SMD_61
VSUS_3.3
SRCOMP
VSUS_3.3
VSS
NC
H
RS_2#
VSUS_3.3
RESET#
SMD_62
VSUS_3.3
VSS
VSUS_3.3
J
DRDY#
VSS
DBSY#
SMD_63
VSS
SMD_30
VCCDPLL
K
HIT#
RS_0#
HTRDY#
VSS
SMD_31
V_1.8
VSSDPLL
L
RS_1#
VSS
HITM#
HLOCK#
HREQ_3#
VSS
VSS
VSS
VSS
M
HREQ_0#
HREQ_2#
DEFER#
VSS
BPRI#
V_1.8
VSS
VSS
VSS
N
HREQ_1#
VSS
HREQ_4#
BNR#
HA_7#
VSS
VSS
VSS
VSS
P
HA_4#
HA_11#
HA_14#
VSS
HA_8#
V_1.8
VSS
VSS
VSS
R
HA_9#
VSS
HA_6#
HA_3#
HA_16#
VSS
VSS
VSS
VSS
T
HA_12#
HA_5#
HA_13#
VSS
HA_15#
V_1.8
VSS
VSS
VSS
U
HA_10#
VSS
HA_28#
HA_21#
HA_25#
GTLREF0
VSS
V
HA_31#
HA_22#
HA_19#
VSS
HA_17#
VSS
V_1.8
W
HA_20#
VSS
HA_23#
HA_24#
HA_30#
V_1.8
VSS
Y
HA_29#
HA_18#
HA_27#
VSS
HA_26#
VSS
V_1.8
VSS
V_1.8
VSS
AA
HD_0#
VSS
HD_6#
HD_15#
CPURST#
RESERVED
HCLK
V_1.8
VSS
GTLREF1
V_1.8
VSS
V_1.8
AB
HD_4#
HD_1#
HD_5#
VSS
HD_23#
HD_19#
HD_31#
HD_34#
HD_37#
HD_42#
HD_41#
HD_48#
HD_55#
AC
HD_8#
VSS
HD_17#
HD_7#
VSS
HD_25#
VSS
HD_22#
VSS
HD_44#
VSS
HD_63#
VSS
AD
HD_10#
HD_12#
HD_13#
HD_3#
HD_30#
HD_16#
HD_33#
HD_29#
HD_43#
HD_39#
HD_27#
HD_47#
HD_59#
AE
HD_18#
VSS
HD_11#
VSS
HD_21#
VSS
HD_35#
VSS
HD_36#
VSS
HD_49#
VSS
HD_57#
AF
HD_14#
HD_2#
HD_9#
HD_20#
HD_24#
HD_26#
HD_32#
HD_28#
HD_38#
HD_45#
HD_51#
HD_40#
HD_52#
1
2
3
4
5
6
7
8
9
10
11
12
13
154
Datasheet
®
Intel 82815EM GMCH
R
Figure 12. GMCH2-M Pinout (Top View-Right Side)
14
15
16
17
18
19
20
21
22
23
24
25
26
SMAB_7#
SMAB_5#
SMAA_3
SCSA_1#
SDQM_4
SMD_47
SMD_45
SMD_42
SMD_39
SMD_37
SMD_35
SMD_33
SMD_32
A
VSUS_3.3
SMAB_4#
SMAA_1
SCSA_5#
SMD_12
VSUS_3.3
SMD_44
SMD_41
VSUS_3.3
SMD_36
SMD_34
VSUS_3.3
VSS
B
SMAB_6#
VSS
SRAS#
SDQM_5
VSS
SMD_46
SMD_43
VSS
SMD_38
SMD_1
VSS
V_1.8
HL_10
C
SCSA_2#
SCSA_0#
SDQM_0
SMD_15
SCAS#
SMD_10
SMD_7
SMD_40
SMD_2
SMD_0
HL_9
HL_8
HL_7
D
SCSA_3#
VSS
SWE#
VSS
SMD_11
SMD_9
VSS
SMD_4
VSSBA
VCCBA
V_1.8
HL_6
HL_5
E
VSUS_3.3
SDQM_1
VSS
VSUS_3.3
SMD_13
SMD_8
SMD_6
SMD_3
HLCLK
V_1.8
HL_4
VSS
HLPSTRB#
F
VSS
SMD_14
VSUS_3.3
SMD_5
VSS
V_1.8
VSS
HL_3
HLPSTRB
V_1.8
G
HLZCOMP
HLREF
VSS
GCBE_0#
HL_0
HL_2
HL_1
H
GAD_5
GAD_3
GAD_1
VSS
AGPREF
VSS
GRCOMP
J
VDDQ
VSS
GAD_8
GAD_7
VSS
GAD_2
GAD_0
K
VSS
VSS
VSS
VDDQ
VSS
ADSTB0#
GAD_4
VSS
GAD_6
L
VSS
VSS
VSS
GAD_12
ADSTB0
VDDQ
GAD_10
GAD_9
GAD_11
M
VSS
VSS
VSS
GCBE_1#
GAD_14
VSS
GAD_13
VDDQ
GAD_15
N
VSS
VSS
VSS
GTRDY#
LRCLK
GIRDY#
VSS
GSTOP#
GDEVSEL#
P
VSS
VSS
VSS
VDDQ
LOCLK
VSS
GPAR
VSS
GFRAME#
R
VSS
VSS
VSS
VSS
GAD_17
GAD_19
GAD_21
GCBE_2#
GAD_16
T
VDDQ
GAD_23
ADSTB1
VDDQ
GAD_18
VDDQ
GAD_20
U
VSS
GAD_25
VSS
ADSTB1#
GAD_22
GAD_24
GAD_26
V
VDDQ
GAD_27
GAD_29
VSS
GAD_28
VSS
GAD_30
W
Y
INTRPT#
V_1.8
VSSDA
IWASTE
GAD_31
SBA_6
SBSTB
VDDQ
SBA_7
GCBE_3#
VSS
V_1.8
VSS
V_1.8
DDDA
V_1.8
LTVDA
VCCDA
SBA_4
VSS
SBSTB#
VSS
SBA_5
AA
HD_53#
HD_56#
V_1.8
LTVHSYNC
DDCK
LTVBLANK#
V_1.8
LTVCK
SBA_0
SBA_2
WBF#
SBA_1
SBA_3
AB
HD_58#
VSS
LTVVSYNC
VSS
LTVCLKIN/
VSS
LTVDATA_8
VSS
V_1.8
ST_2
ST_1
VSS
PIPE#
AC
V_1.8
LTVDATA_7
LTVDATA_1
RED
IREF
ST_0
GGNT#
RBF#
AD
GREEN
BLUE
DCLKREF
VSSDACA
GREQ#
AE
AF
STALL
HD_46#
HD_60#
LTVDATA_0
LTVDATA_3
LTVDATA_5
1
VSS
HD_50#
VSS
LTVDATA_2
VSS
LTVCLKOUT_0
VSS
LTVDATA_1
0
HD_54#
HD_62#
HD_61#
LTVDATA_1
LTVDATA_4
LTVCLKOUT_1
LTVDATA_6
LTVDATA_9
VSYNC
HSYNC
VSSDACA
VCCDACA2
VCCDACA1
14
15
16
17
18
19
20
21
22
23
24
25
26
Datasheet
155
®
Intel 82815EM GMCH
R
Table 27. Alphabetical Pin Assignment (by Signal Name)
Signal Name
Ball #
Signal Name
Ball #
Signal Name
Ball #
Signal Name
Ball #
ADS#
G1
GAD_20
U26
HA_12#
T1
HD_16#
AD6
ADSTB0
M22
GAD_21
T24
HA_13#
T3
HD_17#
AC3
ADSTB0#
L23
GAD_22
V24
HA_14#
P3
HD_18#
AE1
ADSTB1
U22
GAD_23
U21
HA_15#
T5
HD_19#
AB6
ADSTB1#
V23
GAD_24
V25
HA_16#
R5
HD_20#
AF4
AGPBUSY#
E7
GAD_25
V21
HA_17#
V5
HD_21#
AE5
AGPREF
J24
GAD_26
V26
HA_18#
Y2
HD_22#
AC8
BLUE
AE23
GAD_27
W21
HA_19#
V3
HD_23#
AB5
BNR#
N4
GAD_28
W24
HA_20#
W1
HD_24#
AF5
BPRI#
M5
GAD_29
W22
HA_21#
U4
HD_25#
AC6
CPURST#
AA5
GAD_30
W26
HA_22#
V2
HD_26#
AF6
DBSY#
J3
GAD_31
Y21
HA_23#
W3
HD_27#
AD11
DCLKREF
AE24
GCBE_0#
H23
HA_24#
W4
HD_28#
AF8
DDCK
AB18
GCBE_1#
N21
HA_25#
U5
HD_29#
AD8
DDDA
AA18
GCBE_2#
T25
HA_26#
Y5
HD_30#
AD5
DEFER#
M3
GCBE_3#
Y26
HA_27#
Y3
HD_31#
AB7
DRDY#
J1
GDEVSEL#
P26
HA_28#
U3
HD_32#
AF7
GAD_00
K26
GFRAME#
R26
HA_29#
Y1
HD_33#
AD7
GAD_01
J22
GGNT#
AD25
HA_30#
W5
HD_34#
AB8
GAD_02
K25
GIRDY#
P23
HA_31#
V1
HD_35#
AE7
GAD_03
J21
GPAR
R24
HCLK
AA7
HD_36#
AE9
GAD_04
L24
GRCOMP
J26
HD_00#
AA1
HD_37#
AB9
GAD_05
J20
GREEN
AE22
HD_01#
AB2
HD_38#
AF9
GAD_06
L26
GREQ#
AE26
HD_02#
AF2
HD_39#
AD10
GAD_07
K23
GSTOP#
P25
HD_03#
AD4
HD_40#
AF12
GAD_08
K22
GTLREF0
U6
HD_04#
AB1
HD_41#
AB11
GAD_09
M25
GTLREF1
AA10
HD_05#
AB3
HD_42#
AB10
GAD_10
M24
GTRDY#
P21
HD_06#
AA3
HD_43#
AD9
GAD_11
M26
HA_03#
R4
HD_07#
AC4
HD_44#
AC10
GAD_12
M21
HA_04#
P1
HD_08#
AC1
HD_45#
AF10
GAD_13
N24
HA_05#
T2
HD_09#
AF3
HD_46#
AD14
GAD_14
N22
HA_06#
R3
HD_10#
AD1
HD_47#
AD12
GAD_15
N26
HA_07#
N5
HD_11#
AE3
HD_48#
AB12
GAD_16
T26
HA_08#
P5
HD_12#
AD2
HD_49#
AE11
GAD_17
T22
HA_09#
R1
HD_13#
AD3
HD_50#
AE15
GAD_18
U24
HA_10#
U1
HD_14#
AF1
HD_51#
AF11
GAD_19
T23
HA_11#
P2
HD_15#
AA4
HD_52#
AF13
156
Datasheet
®
Intel 82815EM GMCH
R
Signal Name
Ball #
Signal Name
Ball #
Signal Name
Ball #
Signal Name
Ball #
HD_53#
AB14
IREF
AD23
SBA_5
AA26
SMAA_03
A16
HD_54#
AF14
IWASTE
Y20
SBA_6
Y22
SMAA_04
B12
HD_55#
AB13
LOCLK
R22
SBA_7
Y25
SMAA_05
A12
HD_56#
AB15
LRCLK
P22
SBS_0
B13
SMAA_06
C11
HD_57#
AE13
LTVBLANK#
AB19
SBS_1
D11
SMAA_07
A11
HD_58#
AC14
LTVCK
AB21
SBSTB
Y23
SMAA_08
D12
HD_59#
AD13
LTVDA
AA20
SBSTB#
AA24
SMAA_09
C13
HD_60#
AD15
LTVCLKOUT_0 AE19
SCAS#
D18
SMAA_10
E11
HD_61#
AF16
LTVCLKOUT_1 AF19
SCKE_0
D8
SMAA_11
A13
HD_62#
AF15
LTVCLKIN
AC18
SCKE_1
E8
SMAA_12
B7
HD_63#
AC12
LTVDATA_00 AD16
SCKE_2
E9
SMAB_4#
B15
HIT#
K1
LTVDATA_01 AF17
SCKE_3
D7
SMAB_5#
A15
HITM#
L3
LTVDATA_02 AE17
SCKE_4
C8
SMAB_6#
C14
HL_00
H24
LTVDATA_03 AD17
SCKE_5
C7
SMAB_7#
A14
HL_01
H26
LTVDATA_04 AF18
SCLK
F7
SMAC_4#
B10
HL_02
H25
LTVDATA_05 AD18
SCSA_0#
D15
SMAC_5#
A10
HL_03
G24
LTVDATA_06 AF20
SCSA_1#
A17
SMAC_6#
C10
HL_04
F24
LTVDATA_07 AD20
SCSA_2#
D14
SMAC_7#
A9
HL_05
E26
LTVDATA_08 AC20
SCSA_3#
E14
SMD_00
D23
HL_06
E25
LTVDATA_09 AF21
SCSA_4#
E13
SMD_01
C23
HL_07
D26
LTVDATA_10 AE21
SCSA_5#
B17
SMD_02
D22
HL_08
D25
LTVDATA_11 AD21
SCSB_0#
F9
SMD_03
F21
HL_09
D24
LTVHSYNC
AB17
SCSB_1#
F8
SMD_04
E21
HL_10
C26
LTVVSYNC
AC16
SCSB_2#
D10
SMD_05
G20
HLCLK
F22
NC
G10
SCSB_3#
D9
SMD_06
F20
HLOCK#
L4
PIPE#
AC26
SCSB_4#
B9
SMD_07
D20
HLPSTRB
G25
RBF#
AD26
SCSB_5#
A8
SMD_08
F19
HLPSTRB#
F26
RED
AD22
SDQM_0
D16
SMD_09
E19
HLREF
H21
RESERVED
AA6
SDQM_1
F15
SMD_10
D19
HLZCOMP
H20
RESET#
H3
SDQM_2
A7
SMD_11
E18
HREQ_0#
M1
RS_0#
K2
SDQM_3
A6
SMD_12
B18
HREQ_1#
N1
RS_1#
L1
SDQM_4
A18
SMD_13
F18
HREQ_2#
M2
RS_2#
H1
SDQM_5
C17
SMD_14
G18
HREQ_3#
L5
SBA_0
AB22
SDQM_6
B6
SMD_15
D17
HREQ_4#
N3
SBA_1
AB25
SDQM_7
A5
SMD_16
A3
HSYNC
AF23
SBA_2
AB23
SMAA_00
D13
SMD_17
A1
HTRDY#
K3
SBA_3
AB26
SMAA_01
B16
SMD_18
C1
INTRPT#
Y17
SBA_4
AA22
SMAA_02
F12
SMD_19
F2
Datasheet
157
®
Intel 82815EM GMCH
R
Signal Name
Ball #
Signal Name
Ball #
Signal Name
Ball #
Signal Name
Ball #
SMD_20
G3
SMD_58
E3
VCCDACA1
AF26
VSS
AE20
SMD_21
D6
SMD_59
F4
VCCDACA2
AF25
VSS
AE4
SMD_22
C5
SMD_60
F6
VCCDPLL
J7
VSS
AE6
SMD_23
B4
SMD_61
G5
VDDQ
K20
VSS
AE8
SMD_24
D4
SMD_62
H4
VDDQ
L21
VSS
B26
SMD_25
C2
SMD_63
J4
VDDQ
M23
VSS
C12
SMD_26
D3
SRAS#
C16
VDDQ
N25
VSS
C15
SMD_27
E4
SRCOMP
G7
VDDQ
R21
VSS
C18
SMD_28
F5
ST_0
AD24
VDDQ
U20
VSS
C21
SMD_29
G4
ST_1
AC24
VDDQ
U23
VSS
C24
SMD_30
J6
ST_2
AC23
VDDQ
U25
VSS
C3
SMD_31
K5
SWE#
E16
VDDQ
W20
VSS
C6
SMD_32
A26
V_1.8
AA11
VDDQ
Y24
VSS
C9
SMD_33
A25
V_1.8
AA13
VSS
AA12
VSS
D1
SMD_34
B24
V_1.8
AA15
VSS
AA14
VSS
E10
SMD_35
A24
V_1.8
AA17
VSS
AA16
VSS
E12
SMD_36
B23
V_1.8
AA19
VSS
AA2
VSS
E15
SMD_37
A23
V_1.8
AA8
VSS
AA23
VSS
E17
SMD_38
C22
V_1.8
AB16
VSS
AA25
VSS
E20
SMD_39
A22
V_1.8
AB20
VSS
AA9
VSS
E5
SMD_40
D21
V_1.8
AC22
VSS
AB4
VSS
F1
SMD_41
B21
V_1.8
AD19
VSS
AC11
VSS
F11
SMD_42
A21
V_1.8
C25
VSS
AC13
VSS
F13
SMD_43
C20
V_1.8
E24
VSS
AC15
VSS
F16
SMD_44
B20
V_1.8
F23
VSS
AC17
VSS
F25
SMD_45
A20
V_1.8
G22
VSS
AC19
VSS
F3
SMD_46
C19
V_1.8
G26
VSS
AC2
VSS
G17
SMD_47
A19
V_1.8
K6
VSS
AC21
VSS
G21
SMD_48
A4
V_1.8
M6
VSS
AC25
VSS
G23
SMD_49
A2
V_1.8
P6
VSS
AC5
VSS
G9
SMD_50
B1
V_1.8
T6
VSS
AC7
VSS
H22
SMD_51
E1
V_1.8
V7
VSS
AC9
VSS
H6
SMD_52
G2
V_1.8
W6
VSS
AE10
VSS
J2
SMD_53
E6
V_1.8
Y18
VSS
AE12
VSS
J23
SMD_54
D5
V_1.8
Y7
VSS
AE14
VSS
J25
SMD_55
C4
V_1.8
Y9
VSS
AE16
VSS
J5
SMD_56
B3
VCCBA
E23
VSS
AE18
VSS
K21
SMD_57
D2
VCCDA
AA21
VSS
AE2
VSS
K24
158
Datasheet
®
Intel 82815EM GMCH
R
Signal Name
Ball #
Signal Name
Ball #
Signal Name
Ball #
Signal Name
Ball #
VSS
K4
VSS
N16
VSS
T12
VSSDACA
AE25
VSS
L11
VSS
N2
VSS
T13
VSSDACA
AF24
VSS
L12
VSS
N23
VSS
T14
VSSDPLL
K7
VSS
L13
VSS
N6
VSS
T15
VSUS_3.3
B11
VSS
L14
VSS
P11
VSS
T16
VSUS_3.3
B14
VSS
L15
VSS
P12
VSS
T21
VSUS_3.3
B19
VSS
L16
VSS
P13
VSS
T4
VSUS_3.3
B2
VSS
L2
VSS
P14
VSS
U2
VSUS_3.3
B22
VSS
L22
VSS
P15
VSS
U7
VSUS_3.3
B25
VSS
L25
VSS
P16
VSS
V20
VSUS_3.3
B5
VSS
L6
VSS
P24
VSS
V22
VSUS_3.3
B8
VSS
M11
VSS
P4
VSS
V4
VSUS_3.3
E2
VSS
M12
VSS
R11
VSS
V6
VSUS_3.3
F10
VSS
M13
VSS
R12
VSS
W2
VSUS_3.3
F14
VSS
M14
VSS
R13
VSS
W23
VSUS_3.3
F17
VSS
M15
VSS
R14
VSS
W25
VSUS_3.3
G19
VSS
M16
VSS
R15
VSS
W7
VSUS_3.3
G6
VSS
M4
VSS
R16
VSS
Y10
VSUS_3.3
G8
VSS
N11
VSS
R2
VSS
Y4
VSUS_3.3
H2
VSS
N12
VSS
R23
VSS
Y6
VSUS_3.3
H5
VSS
N13
VSS
R25
VSS
Y8
VSUS_3.3
H7
VSS
N14
VSS
R6
VSSBA
E22
VSYNC
AF22
VSS
N15
VSS
T11
VSSDA
Y19
WBF#
AB24
Datasheet
159
®
Intel 82815EM GMCH
R
5.2.
GMCH2-M Package Dimensions
This specification outlines the mechanical dimensions for the Intel® 815EM chipset GMCH2-M. The
package is a 544 ball grid array (BGA).
Figure 13. GMCH2-M GMCH BGA Package Dimensions (Top and Side Views)
160
Datasheet
®
Intel 82815EM GMCH
R
®
Figure 14. Intel 815EM Chipset GMCH2-M BGA Package Dimensions (Bottom View)
Datasheet
161
®
Intel 82815EM GMCH
R
This page left intentionally blank.
162
Datasheet
®
Intel 82815EM GMCH
R
6.
Testability
6.1.
XOR Chain
This is the primary test mode for checking the IO buffer connectivity. There a total of 6 XOR chains each
containing less than 60 XOR gates. The XOR gates are physically located in the IO buffers. This can be
invoked with the use of reset straps.
6.2.
All Z
On a system board in order to be able to apply vectors to XOR chains other chips on the board must be
tristated to allow for this vector application. This is a feature that enables all GMCH2-M outputs to be
tristated when ICH2-M is in the XOR chain mode. This mode can also be activated using the assigned
reset strap.
Datasheet
163