64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet March 2005 Document Number: 306754-001 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 64-Bit Intel® Xeon™ processor MP with up to 8MB L3 cache 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. I2C is a two-wire communications bus/protocol developed by Philips. SMBus is a subset of the I2C 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. Intel, Pentium, Intel Xeon, Intel NetBurst, Intel SpeedStep and the Intel logo are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United States and other countries. *Other names and brands may be claimed as the property of others. Copyright © 2005, Intel Corporation. 2 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Contents 1 Introduction....................................................................................................................... 11 1.1 1.2 1.3 2 Electrical Specifications.................................................................................................... 17 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10 2.11 2.12 2.13 3 Front Side Bus and GTLREF .............................................................................. 17 2.1.1 Front Side Bus Clock and Processor Clocking....................................... 18 2.1.2 Front Side Bus Clock Select (BSEL[1:0]) ............................................... 19 2.1.3 Phase Lock Loop (PLL) Power and Filter............................................... 19 Voltage Identification (VID).................................................................................. 20 Cache Voltage Identification (CVID).................................................................... 23 Reserved, Unused, and TESTHI Pins.................................................................24 Mixing Processors ............................................................................................... 24 Front Side Bus Signal Groups............................................................................. 25 GTL+ Asynchronous and AGTL+ Asynchronous Signals ................................... 27 Test Access Port (TAP) Connection.................................................................... 27 Maximum Ratings................................................................................................ 27 Processor DC Specifications............................................................................... 28 2.10.1 Flexible Motherboard (FMB) Guidelines................................................. 28 2.10.2 VCC and VCACHE Overshoot Specification.......................................... 32 2.10.3 Die Voltage Validation ............................................................................ 33 AGTL+ Front Side Bus Specifications .................................................................37 Front Side Bus AC Specifications ....................................................................... 37 Processor AC Timing Waveforms ....................................................................... 42 Front Side Bus Signal Quality Specifications ................................................................... 53 3.1 4 Terminology......................................................................................................... 12 References .......................................................................................................... 14 State of Data ....................................................................................................... 15 Front Side Bus Signal Quality Specifications and Measurement Guidelines ...... 53 3.1.1 Ringback Guidelines .............................................................................. 53 3.1.2 Overshoot/Undershoot Guidelines ......................................................... 56 3.1.3 Overshoot/Undershoot Magnitude ......................................................... 56 3.1.4 Overshoot/Undershoot Pulse Duration................................................... 57 3.1.5 Activity Factor......................................................................................... 57 3.1.6 Reading Overshoot/Undershoot Specification Tables............................ 57 3.1.7 Determining if a System Meets Over/Undershoot Specifications........... 58 Mechanical Specifications ................................................................................................ 61 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 Package Mechanical Drawing ............................................................................. 62 Processor Component Keep-Out Zones ............................................................. 65 Package Loading Specifications ......................................................................... 65 Package Handling Guidelines ............................................................................. 66 Package Insertion Specifications ........................................................................ 66 Processor Mass Specifications ........................................................................... 66 Processor Materials............................................................................................. 66 Processor Markings............................................................................................. 67 Processor Pin-Out Coordinates........................................................................... 68 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 3 5 Pin Listing......................................................................................................................... 69 5.1 6 Signal Definitions ............................................................................................................. 87 6.1 7 7.2 8.3 8.4 Power-On Configuration Options ...................................................................... 105 Clock Control and Low Power States................................................................ 105 8.2.1 Normal State ........................................................................................ 106 8.2.2 HALT or Enhanced Power Down State ................................................ 106 8.2.3 Stop-Grant State .................................................................................. 107 8.2.4 Enhanced HALT Snoop State or HALT Snoop State, Stop Grant Snoop State ....................................................................... 108 Enhanced Intel SpeedStep® Technology.......................................................... 108 System Management Bus (SMBus) Interface ................................................... 109 8.4.1 Processor Information ROM (PIROM).................................................. 110 8.4.2 Scratch EEPROM ................................................................................ 112 8.4.3 PIROM and Scratch EEPROM Supported SMBus Transactions ......... 113 8.4.4 SMBus Thermal Sensor ....................................................................... 113 8.4.5 Thermal Sensor Supported SMBus Transactions ................................ 114 8.4.6 SMBus Thermal Sensor Registers ....................................................... 115 8.4.7 SMBus Thermal Sensor Alert Interrupt ................................................ 118 8.4.8 SMBus Device Addressing................................................................... 118 8.4.9 Managing Data in the PIROM .............................................................. 120 Boxed Processor Specifications..................................................................................... 127 9.1 9.2 9.3 4 Package Thermal Specifications ......................................................................... 97 7.1.1 Thermal Specifications ........................................................................... 97 7.1.2 Thermal Metrology ............................................................................... 100 Processor Thermal Features............................................................................. 100 7.2.1 Thermal Monitor ................................................................................... 100 7.2.2 Thermal Monitor 2 ................................................................................ 101 7.2.3 On-Demand Mode................................................................................ 102 7.2.4 PROCHOT# Signal Pin ........................................................................ 103 7.2.5 FORCEPR# Signal Pin ........................................................................ 103 7.2.6 THERMTRIP# Signal Pin ..................................................................... 103 7.2.7 TCONTROL and Fan Speed Reduction............................................... 103 7.2.8 Thermal Diode...................................................................................... 104 Features ......................................................................................................................... 105 8.1 8.2 9 Signal Definitions ................................................................................................ 87 Thermal Specifications..................................................................................................... 97 7.1 8 Processor Pin Assignments ................................................................................ 69 5.1.1 Pin Listing by Pin Name ......................................................................... 69 5.1.2 Pin Listing by Pin Number ...................................................................... 78 Introduction ....................................................................................................... 127 Mechanical Specifications ................................................................................. 128 9.2.1 Boxed Processor Heatsink Dimensions ............................................... 128 9.2.2 Boxed Processor Heatsink Weight....................................................... 135 9.2.3 Boxed Processor Retention Mechanism and Heatsink Supports......... 135 Thermal Specifications...................................................................................... 135 9.3.1 Boxed Processor Cooling Requirements ............................................. 135 9.3.2 Boxed Processor Contents .................................................................. 136 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 10 Debug Tools Specifications............................................................................................137 10.1 Logic Analyzer Interface (LAI) ...........................................................................137 10.1.1 Mechanical Considerations ..................................................................137 10.1.2 Electrical Considerations......................................................................137 2-1 2-2 2-3 2-4 2-5 2-6 2-7 2-8 2-9 2-10 2-11 2-12 2-13 2-14 2-15 2-16 2-17 2-18 2-19 2-20 2-21 2-22 3-1 3-2 3-3 3-4 3-5 4-1 4-2 4-3 4-4 4-5 4-6 7-1 7-2 7-3 8-1 8-2 9-1 9-2 9-3 9-4 On-Die Front Side Bus Termination .................................................................... 17 Phase Lock Loop (PLL) Filter Requirements ...................................................... 20 Processor Load Current vs. Time........................................................................ 30 VCC Static and Transient Tolerance................................................................... 32 VCC and VCACHE Overshoot Example Waveform............................................ 33 Electrical Test Circuit........................................................................................... 42 TCK Clock Waveform.......................................................................................... 43 Differential Clock Waveform................................................................................ 43 Differential Clock Crosspoint Specification.......................................................... 44 Front Side Bus Common Clock Valid Delay Timing Waveform........................... 44 Source Synchronous 2X (Address) Timing Waveform........................................ 45 Source Synchronous 4X (Data) Timing Waveform ............................................. 46 TAP Valid Delay Timing Waveform ..................................................................... 47 Test Reset (TRST#), Async GTL+ Input, and PROCHOT# Timing Waveform ... 47 THERMTRIP# Power Down Sequence............................................................... 47 SMBus Timing Waveform.................................................................................... 48 SMBus Valid Delay Timing Waveform ................................................................ 48 Voltage Sequence Timing Requirements............................................................ 49 VIDPWRGD Timing Requirements ..................................................................... 50 FERR#/PBE# Valid Delay Timing ....................................................................... 50 VID Step Timings ................................................................................................ 51 VID Step Times and VCC Waveforms ................................................................ 51 Low-to-High Front Side Bus Receiver Ringback Tolerance ................................ 54 High-to-Low Front Side Bus Receiver Ringback Tolerance ................................ 54 Low-to-High Receiver Ringback Tolerance for PWRGOOD and TAP Signals ... 55 High-to-Low Receiver Ringback Tolerance for PWRGOOD and TAP Signals ... 56 Maximum Acceptable Overshoot/Undershoot Waveform ................................... 60 Processor Package Assembly Sketch.................................................................61 Processor Package Drawing (Sheet 1 of 2) ........................................................ 63 Processor Package Drawing (Sheet 2 of 2) ........................................................ 64 Processor Topside Markings............................................................................... 67 Processor Bottom-Side Markings........................................................................ 67 Processor Pin-Out Coordinates, Top View.......................................................... 68 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Thermal Profile .. 99 Case Temperature (TCASE) Measurement Location .......................................100 Thermal Monitor 2 Frequency and Voltage Ordering ........................................102 Stop Clock State Machine .................................................................................107 Logical Schematic of SMBus Circuitry ..............................................................110 Passive Processor Thermal Solution (3U and larger) .......................................128 Top Side Board Keep-Out Zones (Part 1) .........................................................129 Top Side Board Keep-Out Zones (Part 2) .........................................................130 Bottom Side Board Keep-Out Zones.................................................................131 Figures 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 5 9-5 9-6 9-7 Board Mounting-Hole Keep-Out Zones ............................................................. 132 Thermal Solution Volumetric ............................................................................. 133 Recommended Processor Layout and Pitch ..................................................... 134 1-1 2-1 2-2 2-3 2-4 2-5 2-6 2-7 2-8 2-9 2-10 2-11 2-12 2-13 2-14 2-15 2-16 2-17 2-18 2-19 2-20 2-21 2-22 2-23 2-24 2-25 2-26 2-27 2-28 3-1 3-2 3-3 3-4 3-5 3-6 4-1 4-2 4-3 5-1 5-2 6-1 7-1 7-2 8-1 Features of the 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache .. 12 Core Frequency to Front Side Bus Multiplier Configuration ................................ 18 BSEL[1:0] Frequency Table for BCLK[1:0] ......................................................... 19 Voltage Identification (VID) Definition ................................................................. 22 Cache Voltage Identification (CVID) Definition ................................................... 23 Front Side Bus Pin Groups ................................................................................. 25 Signal Description Table ..................................................................................... 26 Signal Reference Voltages.................................................................................. 26 Processor Absolute Maximum Ratings ............................................................... 28 Voltage and Current Specifications..................................................................... 29 VCC Static and Transient Tolerance................................................................... 31 VCC and VCACHE Overshoot Specification....................................................... 32 Front Side Bus Differential BCLK Specifications................................................. 33 BSEL[1:0], VID[5:0], and CVID[3:0] DC Specifications ....................................... 34 VIDPWRGD DC Specifications ........................................................................... 34 AGTL+ Signal Group DC Specifications ............................................................. 35 PWRGOOD and TAP Signal Group DC Specifications ...................................... 35 GTL+ and AGTL+ Asynchronous Signal Group DC Specifications .................... 36 SMBus Signal Group DC Specifications ............................................................. 36 AGTL+ Bus Voltage Definitions........................................................................... 37 Front Side Bus Differential Clock Specifications ................................................. 38 Front Side Bus Common Clock AC Specifications.............................................. 38 Front Side Bus Source Synchronous AC Specifications ..................................... 38 Miscellaneous Signals AC Specifications ........................................................... 39 Front Side Bus AC Specifications (Reset Conditions) ........................................ 40 TAP Signal Group AC Specifications .................................................................. 40 VIDPWRGD and Other Voltage Sequence AC Specifications ............................ 41 VID Signal Group AC Timing Specifications ....................................................... 41 SMBus Signal Group AC Specifications ............................................................. 41 Ringback Specifications for AGTL+ and GTL+ Asynchronous Signal Groups.... 54 Ringback Specifications for PWRGOOD and TAP Signal Groups...................... 55 Source Synchronous (667MHz) AGTL+ Signals Over/Undershoot Tolerance ... 59 Source Synchronous (333 MHz) AGTL+ Signals Over/Undershoot Tolerance .. 59 Common Clock (166 MHz) AGTL+ Signals Overshoot/Undershoot Tolerance .. 59 GTL+ Asynchronous, PWRGOOD, TAP Signals Over/Undershoot Tolerance... 60 Processor Loading Specifications ....................................................................... 65 Package Handling Guidelines ............................................................................. 66 Processor Materials ............................................................................................ 66 Pin Listing by Pin Name ...................................................................................... 69 Pin Listing by Pin Number ................................................................................... 78 Signal Definitions ................................................................................................ 87 Processor Thermal Specifications....................................................................... 98 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Thermal Profile .. 99 Power-On Configuration Option Pins ................................................................ 105 Tables 6 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 8-2 8-3 8-4 8-5 8-6 8-7 8-8 8-9 8-10 8-11 8-12 8-13 8-14 8-15 8-16 8-17 8-18 Processor Information ROM Format .................................................................110 Read Byte SMBus Packet .................................................................................113 Write Byte SMBus Packet .................................................................................113 Write Byte SMBus Packet .................................................................................114 Read Byte SMBus Packet .................................................................................114 Send Byte SMBus Packet .................................................................................114 Receive Byte SMBus Packet.............................................................................114 ARA SMBus Packet ..........................................................................................114 SMBus Thermal Sensor Command Byte Bit Assignments................................115 Thermal Value Register Encoding.....................................................................116 SMBus Thermal Sensor Status Register...........................................................117 SMBus Thermal Sensor Configuration Register ...............................................117 SMBus Thermal Sensor Conversion Rate Registers ........................................118 Thermal Sensor SMBus Addressing .................................................................119 Memory Device SMBus Addressing..................................................................119 Offset 78h Definitions ........................................................................................124 128 Byte ROM Checksum Values.....................................................................125 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 7 Revision History Version Number -001 8 Description • Initial Release Date March 2005 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Product Features 64-bit Processor with Intel® Extended Memory 64 Technology1 Available at 2.66, 2.83, 3.00, 3.16, 3.33, and 3.66 GHz Multi-processing (4 sockets and above) server support Binary compatible with applications running on previous members of the Intel® IA-32 microprocessor line Intel NetBurst® microarchitecture Hyper-Threading Technology — Hardware support for multi-threaded applications 667 MHz System bus with data-bus Error Correcting Code (ECC) — Dual independent bus architecture supports two processors per bus segment — Bandwidth up to 5.33 GB/second — Split Transaction Bus with Modified Enhanced Defer improves throughput Rapid Execution Engine: Arithmetic Logic Units (ALUs) run at twice the processor core frequency Hyper-Pipelined Technology Advance Dynamic Execution — Very deep out-of-order execution — Enhanced branch prediction Level 1 Execution Trace Cache stores 12 K microops and removes decoder latency from main execution loops 1MB Advanced Transfer L2 Cache (on-die, full speed Level 2 cache) with 8-way associativity and Error Correcting Code (ECC) 4-MB or 8-MB L3 Cache (on-die, full speed Level 3 cache) with 8-way associativity and Error Correcting Code (ECC) Enables system support of up to 1024 GB of physical memory Streaming SIMD Extensions 2 (SSE2 and SSE3) — 144 new instructions for double-precision floating point operations, media/video streaming, and secure transactions Enhanced floating point and multimedia unit for enhanced video, audio, encryption, and 3D performance Power Management capabilities — Enhanced Intel Speed-Step® technology — System Management mode — Multiple low-power states Advanced System Management Features — System Management Bus — Processor Information ROM (PIROM) — OEM Scratch EEPROM — Thermal Monitor, Thermal Monitor 2 (TM2) — Machine Check Architecture (MCA) Execute Disable Bit2 when used with operating systems helps protect against a certain class of malicious virus attacks. — Includes 16-KB Level 1 data cache NOTES: 1. 2. 64-bit Intel® Xeon™ processors with Intel® EM64T requires a computer system with a processor, chipset, BIOS, OS, device drivers and applications enabled for Intel EM64T. Processor will not operate (including 32-bit operation) without an Intel EM64T-enabled BIOS. Performance will vary depending on your hardware and software configurations. Intel EM64T-enabled OS, BIOS, device drivers and applications may not be available. Check with your vendor for more information. Execute Disable Bit requires operating system support. See http://www.intel.com/business/bss/infrastructure/security/xdbit.htm for more information on how to implement this feature. The 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache is designed for high-performance multiprocessor server applications for mid-tier enterprise serving and server consolidation. Based on the Intel NetBurst® microarchitecture and the new Hyper-Threading Technology, it is binary compatible with previous Intel Architecture (IA-32) processors. The addition of Intel® EM64T provides 64-bit computing and 40-bit addressing provides up to 1 Terabyte of direct memory addressability. The 64-bit Intel® Xeon™ processor MP with 4MB L3 cache is scalable to four processors and beyond in a multiprocessor system providing exceptional performance for applications running on advanced operating systems such as Microsoft Windows* 2003 Server, and Linux* operating systems. The 64-bit Intel® Xeon™ processor MP with 8MB L3 cache delivers compute power at unparalleled value and flexibility for internet infrastructure and departmental server applications, including application servers, databases, and business intelligence. The Intel NetBurst microarchitecture with Hyper-Threading technology and Intel EM64T deliver outstanding performance and headroom for peak internet server workloads, resulting in faster response times, support for more users, and improved scalability. § 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 9 10 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 1 Introduction The 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache is a 64-bit multi-processor capable server processor based on improvements to the Intel NetBurst® microarchitecture. It maintains the tradition of compatibility with IA-32 software and includes features found in the Intel® Xeon™ processor such as Hyper Pipelined Technology, a Rapid Execution Engine, and an Execution Trace Cache. Hyper Pipelined Technology includes a multi-stage pipeline, allowing the processor to reach much higher core frequencies. The 667 MHz front side bus is a quad-pumped bus running off a 166 MHz system clock making 5.3 GB per second data transfer rates possible. The Execution Trace Cache is a level 1 (L1) cache that stores decoded micro-operations, which removes the decoder from the main execution path, thereby increasing performance. In addition, the 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache includes the Intel® Extended Memory 64 Technology, providing additional address capability. In addition, enhanced thermal and power management capabilities are implemented, including Thermal Monitor, Thermal Monitor 2 (TM2), and Enhanced Intel SpeedStep® technology. Thermal Monitor and Thermal Monitor 2 provide efficient and effective cooling in high temperature situations. Enhanced Intel SpeedStep technology allows trade-offs to be made between performance and power consumption. This may lower average power consumption (in conjunction with OS support). The 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache supports Hyper-Threading Technology. This feature allows a single, physical processor to function as two logical processors. While some execution resources such as caches, execution units, and buses are shared, each logical processor has its own architectural state with its own set of general-purpose registers, control registers to provide increased system responsiveness in multitasking environments, and headroom for next generation multi-threaded applications. More information on Hyper-Threading Technology can be found at http://www.intel.com/technology/hyperthread. Support for Intel's Execute Disable Bit functionality has been added which can prevent certain classes of malicious “buffer overflow” attacks when combined with a supporting operating system. Execute Disable Bit allows the processor to classify areas in memory by where application code can execute and where it cannot. When a malicious worm attempts to insert code in the buffer, the processor disables code execution, preventing damage or worm propagation. Other features within the Intel NetBurst microarchitecture include Advanced Dynamic Execution, Advanced Transfer Cache, enhanced floating point and multi-media unit, and Streaming SIMD Extensions 2 (SSE2). The Advanced Dynamic Execution improves speculative execution and branch prediction internal to the processor. The Advanced Transfer Cache is a 1 MB on-die level 2 (L2) cache with increased bandwidth. The floating point and multi-media units include 128-bit wide registers and a separate register for data movement. SSE2 instructions provide highly efficient double-precision floating point, SIMD integer, and memory management operations. In addition, Streaming SIMD Extensions 3 (SSE3) instructions have been added to further extend the capabilities of Intel processor technology. Other processor enhancements include core frequency improvements and microarchitectural improvements. The 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache supports Intel® Extended Memory 64 Technology (Intel® EM64T) as an enhancement to Intel’s IA-32 architecture. This enhancement allows the processor to execute operating systems and applications written to take advantage of the 64-bit extension technology. The processor supports 40-bit addressing, data bus 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 11 Introduction ECC protection (single-bit error correction with double-bit error detection), and the bus protocol addition of the Deferred Phase. Further details can be found in the 64-bit Extension Technology Software Developer’s Guide at http://developer.intel.com/technology/64bitextensions/. The 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache is intended for high performance multi-processor server systems with support for up to two processors on a 667 MHz front side bus. The 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache will be available with 4 MB or 8 MB of on-die level 3 (L3) cache. All versions of the processor will include manageability features. Components of the manageability features include an OEM EEPROM and Processor Information ROM which are accessed through an SMBus interface and contain information relevant to the particular processor and system in which it is installed. Table 1-1. Features of the 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Processor # of Symmetric Agents # of Supported Symmetric Agents Per FSB L2 Advanced Transfer Cache Integrated L3 Cache FSB Frequency 1-255 1-2 1 MB 4 MB or 8 MB 667 MHz The processor is packaged in a 604-pin Flip-Chip Micro Pin Grid Array (FC-mPGA4) package and utilizes a surface-mount Zero Insertion Force (ZIF) mPGA604 socket. The processor uses a scalable system bus referred to as the “Front Side Bus” (FSB) in this document. The FSB utilizes a split-transaction, deferred reply and modified enhanced deferred phase protocol that improves bandwidth and throughput by reducing the number of cycles needed to return data from a deferred response. The front side bus uses Source-Synchronous Transfer (SST) of address and data to improve performance. The processor transfers data four times per bus clock (4X data transfer rate). Along with the 4X data bus, the address bus can deliver addresses two times per bus clock and is referred to as a ‘double-clocked’, ‘double-pumped’, or the 2X address bus. In addition, the Request Phase completes in one clock cycle. Working together, the 4X data bus and 2X address bus provide a data bus bandwidth of up to 5.3 GB per second. Finally, the front side bus is also used to deliver interrupts. 1.1 Terminology A ‘#’ symbol after a signal name refers to an active low signal, indicating that a signal is in the asserted state when driven to a low level. For example, when RESET# is low (i.e. when RESET# is asserted), a reset has been requested. Conversely, when NMI is high (i.e. when NMI is asserted), a nonmaskable interrupt request has occurred. In the case of signals where the name does not imply an active state but describes part of a binary sequence (such as address or data), the ‘#’ symbol implies that the signal is inverted. For example, D[3:0] = ‘HLHL’ refers to a hex ‘A’, and D[3:0]# = ‘LHLH’ also refers to a hex ‘A’ (H= High logic level, L= Low logic level). “Front side bus” refers to the interface between the processor, system core logic (i.e. the chipset components), and other bus agents. The front side bus supports multiprocessing and cache coherency. For this document, “front side bus” is used as the generic term for the processor system bus. Commonly used terms are explained here for clarification: • 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache — The entire product, including processor core substrate and integrated heat spreader (IHS). 12 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Introduction • Enhanced Intel SpeedStep technology — Enhanced Intel SpeedStep technology is the next generation implementation of the Geyserville technology which extends power management capabilities of servers. • FC-mPGA4 — The processor is available in a Flip-Chip Micro Pin Grid Array 4 package, consisting of a processor core mounted on a pinned substrate with an integrated heat spreader (IHS). This packaging technology employs a 1.27 mm [0.05 in] pitch for the substrate pins. • Front Side Bus (FSB) — The electrical interface that connects the processor to the chipset. Also referred to as the processor system bus or the system bus. All memory and I/O transactions as well as interrupt messages pass between the processor and chipset over the FSB. • Functional Operation — Refers to the normal operating conditions in which all processor specifications, including DC, AC, system bus, signal quality, mechanical, and thermal, are satisfied. • Integrated Heat Spreader (IHS) — A component of the processor package used to enhance the thermal performance of the package. Component thermal solutions interface with the processor at the IHS surface. • mPGA604 — The processor mates with the system board through this surface mount, 604-pin, zero insertion force (ZIF) socket. • OEM — Original Equipment Manufacturer. • Processor core — The processor’s execution engine. All AC timing and signal integrity specifications are to the pads of the processor core. • Processor Information ROM (PIROM) — A memory device located on the processor and accessible via the System Management Bus (SMBus) which contains information regarding the processor’s features. This device is shared with the Scratch EEPROM, is programmed during manufacturing, and is write-protected. • Scratch EEPROM (Electrically Erasable, Programmable Read-Only Memory) — A memory device located on the processor and addressable via the SMBus which can be used by the OEM to store information useful for system management. • SMBus — System Management Bus. A two-wire interface through which simple system and power management related devices can communicate with the rest of the system. It is based on the principals of the operation of the I2C* two-wire serial bus from Phillips Semiconductor. Note: I2C is a two-wire communications bus/protocol developed by Philips. SMBus is a subset of the I2C bus/protocol and was developed by Intel. Implementations of the I2C bus/protocol or the SMBus bus/protocol may require licenses from various entities, including Philips Electronics N.V. and North American Philips Corporation. • Storage Conditions — Refers to a non-operational state. The processor may be installed in a platform, in a tray, or loose. Processors may be sealed in packaging or exposed to free air. Under these conditions, processor pins should not be connected to any supply voltages, have any I/Os biased, or receive any clocks. • Symmetric Agent - A symmetric agent is a processor which shares the same I/O subsystem and memory array, and runs the same operating system as another processor in a system. Systems using symmetric agents are known as Symmetric MultiProcessing (SMP) systems. The 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache should only be used in SMP systems which have two or fewer symmetric agents per front side bus. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 13 Introduction 1.2 References Material and concepts available in the following documents may be beneficial when reading this document. Some document references in this specification are generic and should be referenced back to the specific documents listed here. Document AP-485 Intel® Processor Identification and the CPUID Instruction Intel Order Number Notes 241618 2 ATX/ATX12V Power Supply Design Guidelines 3 Chassis Strength and Stiffness Measurement and Improvement Guidelines for Direct Chassis Attach Solutions 1 IA-32 Intel® Architecture and Intel® Extended Memory 64 Technology Software Developer's Manual Documentation Changes 252046 2 IA-32 Intel® Architecture Optimization Reference Manual 248966 2 IA-32 Intel® Architecture Software Developer's Manual • Volume 1: Basic Architecture 253665 • Volume 2A: Instruction Set Reference, A-M 253666 • Volume 2B: Instruction Set Reference, N-Z 253667 • Volume 3: System Programming Guide 253668 2 Intel® Extended Memory 64 Technology Software Developer’s Manual • Volume 1 300834 • Volume 2 300835 mPGA604 Socket Design Guidelines 254239 MPS Power Supply: A Server System Infrastructure (SSI) Specification For Midrange Chassis Power Supplies 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Specification Update 2 1 4 289741 1 64-bit Intel® Xeon™ processor MP with up to 8MB L3 Cache Mechanical Models 1 64-bit Intel® Xeon™ processor MP with up to 8MB L3 Cache Cooling Solution Mechanical Models 1 64-bit Intel® Xeon™ processor MP with up to 8MB L3 Cache Thermal Test Vehicle and Cooling Solution Thermal Models 1 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Thermal/Mechanical Design Guide 306749 1 Prescott New Instructions Software Development Guide 252490 2 System Management Bus (SMBus) Specification 5 Voltage Regulator Module (VRM) and Enterprise Voltage Regulator Down (EVRD) 10.1 Design Guidelines 302732 1 Voltage Regulator Module (VRM) and Enterprise Voltage Regulator Down (EVRD) 10.2 Design Guidelines 306760 1 Voltage Regulator Module (VRM) 10.2L Design Guidelines 306761 1 VRM 9.1 DC-DC Converter Design Guidelines 306760 2 NOTES: 1. See the Intel® XeonTM Processor Message of the Week for latest document version and order numbers. Contact your Intel representative to receive the latest revisions of these documents. 2. This collateral is available publicly at http://developer.intel.com. 3. This document is available at http://www.formfactors.org. 14 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Introduction 4. This document is available at http://www.ssiforum.org. 5. This document is available at http://www.smbus.org. 1.3 State of Data The data contained within this document is subject to change. It is the most accurate information available by the publication date of this document. Electrical AC / DC specifications are based on I/O buffer behavior in the production processor. Mechanical data is based on the final FC-mPGA4 package design. For processor stepping information, refer to the 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Specification Update. § 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 15 Introduction 16 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 2 Electrical Specifications 2.1 Front Side Bus and GTLREF Most processor front side bus (FSB) signals use Assisted Gunning Transceiver Logic (AGTL+) signaling technology. This technology provides improved noise margins and reduced ringing through low voltage swings and controlled edge rates. AGTL+ buffers are open-drain and require pull-up resistors to provide the high logic level and termination. AGTL+ output buffers differ from GTL+ buffers with the addition of an active pMOS pull-up transistor to “assist” the pull-up resistors during the first clock of a low-to-high voltage transition. Platforms implement a termination voltage level for AGTL+ signals defined as VTT. Because platforms implement separate power planes for each processor, separate VCC and VTT supplies are necessary. This configuration allows for improved noise tolerance as processor frequency increases. Speed enhancements to data and address busses have caused signal integrity considerations and platform design methods to become even more critical than with previous processor families. The AGTL+ inputs require a reference voltage (GTLREF) which is used by the receivers to determine if a signal is a logical 0 or a logical 1. GTLREF must be generated on the motherboard (see Table 2-19 for GTLREF specifications). The on-die termination resistors are a selectable feature and can be enabled or disabled via the ODTEN signal. For end bus agents, on-die termination resistors are enabled to control reflections on the transmission line. For the middle bus agent, on-die termination RTT resistors must be disabled. Intel chipsets will also provide ontermination, thus eliminating the need to terminate the bus on the motherboard for most AGTL+ signals. Processor wired-OR signals may also include additional on-die resistors (RL) to further ensure proper noise margin and signal integrity. RL is not configurable and is always enabled for these signals. See Table 2-6 for a list of these signals. Figure 2-1 illustrates the active on-die termination. Figure 2-1. On-Die Front Side Bus Termination End Agent Middle Agent V TT R TT Signal Signal RL R TT - On-die termination resistors for AGTL+ signals R L - Additional on-die resistance implemented for proper noise margin and signal integrity (wired-OR signals only) 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 17 Electrical Specifications Note: 2.1.1 Some AGTL+ signals do not include on-die termination (RTT) and must be terminated on the motherboard. See Table 2-6 for details regarding these signals. Front Side Bus Clock and Processor Clocking BCLK[1:0] directly controls the front side bus interface speed as well as the core frequency of the processor. The processor core frequency is a multiple of the BCLK[1:0] frequency. The processor bus ratio multiplier will be set at its default ratio during manufacturing. The default setting generates the maximum speed for the processor. It is possible to override this setting using software. Refer to the Processor BIOS Writers Guide for details. This will permit operation at a speed lower than the processor’s tested frequency. The processor core frequency is configured during reset by using values stored internally during manufacturing. The stored values set the highest bus fraction at which the particular processor can operate. If lower speeds are desired, the appropriate bus ratio multiplier can be configured by driving the A[21:16]# pins at reset. For details of operation at core frequencies lower than the maximum rated processor speed, refer to the Processor BIOS Writers Guide. The bus ratio multipliers supported are shown in Table 2-1. Other combinations will not be validated or supported by Intel. For a given processor, only the ratios which result in a core frequency equal to or less than the frequency marked on the processor are supported. Table 2-1. Core Frequency to Front Side Bus Multiplier Configuration Core Frequency to Front Side Bus Multiplier Core Frequency 1/16 2.66 GHz L H L L L L 1/17 2.83 GHz L H L L L H 1/18 3.00 GHz L H L L H L 1/19 3.16 GHz L H L L H H 1/20 3.33 GHz L H L H L L 1/22 3.66 GHz L H L H H L A21# A20# A19# A18# A17# A16# NOTES: 1. Individual processors operate only at or below the frequency marked on the package. 2. Listed frequencies are not necessarily committed production frequencies. 3. For valid core frequencies of the processor, refer to the Processor Specification Update. 4. As described in Section 1.1, “H” refers to a high logic level (i.e. signal asserted) and “L” refers to a low logic level (i.e. signal deasserted). The processor uses a differential clocking implementation. 18 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Electrical Specifications 2.1.2 Front Side Bus Clock Select (BSEL[1:0]) The BSEL[1:0] signals are used to select the frequency of the processor input clock (BCLK[1:0]). Table 2-2 defines the possible combinations of the signals and the frequency associated with each combination. The required frequency is determined by the processor, chipset, and clock synthesizer. All processors must operate at the same front side bus frequency. The processor operates at a 667 MHz front side bus frequency (selected by a 166 MHz BCLK[1:0] frequency). Individual processors operate at the front side bus frequency specified by BSEL[1:0]. For more information about these pins, refer to Section 6.1. Table 2-2. BSEL[1:0] Frequency Table for BCLK[1:0] 2.1.3 BSEL1 BSEL0 Function 0 0 RESERVED 0 1 RESERVED 1 0 RESERVED 1 1 166 MHz Phase Lock Loop (PLL) Power and Filter VCCA, VCCIOPLL, and VCCA_CACHE are power sources required by the PLL clock generators on the processor. These are analog PLLs and they require low noise power supplies for minimum jitter. These supplies must be low pass filtered from VTT. The AC low-pass requirements, with input at VTT, are as follows: • • • • < 0.2 dB gain in pass band < 0.5 dB attenuation in pass band < 1 Hz > 34 dB attenuation from 1 MHz to 66 MHz > 28 dB attenuation from 66 MHz to core frequency The filter requirements are illustrated in Figure 2-2. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 19 Electrical Specifications Figure 2-2. Phase Lock Loop (PLL) Filter Requirements 0.2 dB 0 dB -0.5 dB forbidden zone -28 dB forbidden zone -34 dB DC 1 Hz fpeak 1 MHz 66 MHz fcore high frequency band passband NOTES: 1. Diagram not to scale. 2. No specification for frequencies beyond fcore (core frequency). 3. fpeak, if existent, should be less than 0.05MHz. 4. fcore represents the maximum core frequency supported by the platform. 2.2 Voltage Identification (VID) The VID[5:0] pins supply the encodings that determine the voltage to be supplied by the VCC (the core voltage for the processor) voltage regulator. The VID specification for the processor is defined by the Voltage Regulator Module (VRM) and Enterprise Voltage Regulator-Down (EVRD) 10.1 Design Guidelines, the Voltage Regulator Module (VRM) and Enterprise Voltage Regulator-Down (EVRD) 10.2 Design Guidelines, and the Voltage Regulator Module (VRM) 10.2L Design Guidelines. Please refer to these documents for all VRM and VRD design issues. The voltage set by the VID signals is the maximum VCC voltage allowed by the processor. VID signals are open drain outputs, which must be pulled up to VTT. Please refer to Table 2-13 for the DC specifications for these signals. A minimum VCC voltage is provided in Table 2-9 and changes with frequency. This allows processors running at a higher frequency to have a relaxed minimum VCC voltage specification. The specifications have been set such that one voltage regulator can work with all supported frequencies. 20 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Electrical Specifications Individual processor VID values may be calibrated during manufacturing such that two devices at the same core speed may have different default VID settings. Furthermore, any processor, even those on the same processor front side bus, can drive different VID settings during normal operation. The processor uses six voltage identification pins, VID[5:0], to support automatic selection of power supply voltages. Table 2-3 specifies the voltage level corresponding to the state of VID[5:0]. A ‘1’ in this table refers to a high voltage level and a ‘0’ refers to a low voltage level. If the processor socket is empty (i.e. VID[5:0] = x11111), or the voltage regulation circuit cannot supply the voltage that is requested, the processor’s voltage regulator must disable itself. The processor provides the ability to operate while transitioning to an adjacent VID and its associated processor core voltage (VCC). This will represent a DC shift in the load line. It should be noted that a low-to-high or high-to-low voltage state change may result in as many VID transitions as necessary to reach the target core voltage. Transitions above the specified VID are not permitted. Table 2-9 includes VID step sizes and DC shift ranges. Minimum and maximum voltages must be maintained as shown in Table 2-10 and Figure 2-4. The VRM or VRD utilized must be capable of regulating its output to the value defined by the new VID. DC specifications for VID transitions are included in Table 2-9 and Table 2-10. Power source characteristics must be guaranteed to be stable whenever the supply to the voltage regulator is stable. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 21 Electrical Specifications Table 2-3. Voltage Identification (VID) Definition VID5 VID4 VID3 VID2 VID1 VID0 VID (V) VID5 VID4 VID3 VID2 VID1 VID0 VID (V) 0 0 1 0 1 0 0.8375 0 1 1 0 1 0 1.2125 1 0 1 0 0 1 0.8500 1 1 1 0 0 1 1.2250 0 0 1 0 0 1 0.8625 0 1 1 0 0 1 1.2375 1 0 1 0 0 0 0.8750 1 1 1 0 0 0 1.2500 0 0 1 0 0 0 0.8875 0 1 1 0 0 0 1.2625 1 0 0 1 1 1 0.9000 1 1 0 1 1 1 1.2750 0 0 0 1 1 1 0.9125 0 1 0 1 1 1 1.2875 1 0 0 1 1 0 0.9250 1 1 0 1 1 0 1.3000 0 0 0 1 1 0 0.9375 0 1 0 1 1 0 1.3125 1 0 0 1 0 1 0.9500 1 1 0 1 0 1 1.3250 0 0 0 1 0 1 0.9625 0 1 0 1 0 1 1.3375 1 0 0 1 0 0 0.9750 1 1 0 1 0 0 1.3500 0 0 0 1 0 0 0.9875 0 1 0 1 0 0 1.3625 1 0 0 0 1 1 1.0000 1 1 0 0 1 1 1.3750 0 0 0 0 1 1 1.0125 0 1 0 0 1 1 1.3875 1 0 0 0 1 0 1.0250 1 1 0 0 1 0 1.4000 0 0 0 0 1 0 1.0375 0 1 0 0 1 0 1.4125 1 0 0 0 0 1 1.0500 1 1 0 0 0 1 1.4250 0 0 0 0 0 1 1.0625 0 1 0 0 0 1 1.4375 1 0 0 0 0 0 1.0750 1 1 0 0 0 0 1.4500 0 0 0 0 0 0 1.0875 0 1 0 0 0 0 1.4625 1 1 1 1 1 1 VRM off 1 0 1 1 1 1 1.4750 0 1 1 1 1 1 VRM off 0 0 1 1 1 1 1.4875 1 1 1 1 1 0 1.1000 1 0 1 1 1 0 1.5000 0 1 1 1 1 0 1.1125 0 0 1 1 1 0 1.5125 1 1 1 1 0 1 1.1250 1 0 1 1 0 1 1.5250 0 1 1 1 0 1 1.1375 0 0 1 1 0 1 1.5375 1 1 1 1 0 0 1.1500 1 0 1 1 0 0 1.5500 0 1 1 1 0 0 1.1625 0 0 1 1 0 0 1.5625 1 1 1 0 1 1 1.1750 1 0 1 0 1 1 1.5750 0 1 1 0 1 1 1.1875 0 0 1 0 1 1 1.5875 1 1 1 0 1 0 1.2000 1 0 1 0 1 0 1.6000 22 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Electrical Specifications 2.3 Cache Voltage Identification (CVID) The CVID[3:0] pins supply the encodings that determine the voltage to be supplied by the VCACHE (the L3 cache voltage for the processor) voltage regulator. The CVID specification for the processor is defined by the VRM 9.1 DC-DC Converter Design Guidelines, Voltage Regulator Module (VRM) and Enterprise Voltage Regulator-Down (EVRD) 10.1 Design Guidelines, Voltage Regulator Module (VRM) and Enterprise Voltage Regulator-Down (EVRD) 10.2 Design Guidelines, and Voltage Regulator Module (VRM)10.2L Design Guidelines. The voltage set by the CVID pins is the maximum VCACHE voltage allowed by the processor. A minimum VCACHE voltage is provided in Table 2-9. Processors with the same front side bus frequency, internal cache sizes, and stepping will have consistent CVID values. The processor uses four voltage identification pins (CVID[3:0]) to support automatic selection of power supply voltages. Table 2-4 specifies the voltage level corresponding to the state of CVID[3:0]. A ‘1’ in this table refers to a high voltage level and a ‘0’ refers to a low voltage level. If the processor socket is empty, or the voltage regulation circuit cannot supply the voltage that is requested, the processor’s voltage regulator must disable itself. Table 2-4. Cache Voltage Identification (CVID) Definition CVID3 CVID2 CVID1 CVID0 CVID (V) 1 1 1 1 Off 1 1 1 0 1.100 1 1 0 1 1.125 1 1 0 0 1.150 1 0 1 1 1.175 1 0 1 0 1.200 1 0 0 1 1.225 1 0 0 0 1.250 0 1 1 1 1.275 0 1 1 0 1.300 0 1 0 1 1.325 0 1 0 0 1.350 0 0 1 1 1.375 0 0 1 0 1.400 0 0 0 1 1.425 0 0 0 0 1.450 NOTE: The voltage regulator will have a fifth VID input and, for VRM 10.2-compliant regulators, a sixth VID input as well. The extra input(s) should be tied to a high voltage on the motherboard for correct operation. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 23 Electrical Specifications 2.4 Reserved, Unused, and TESTHI Pins All RESERVED pins must be left unconnected. Connection of these pins to VCC, VSS, or to any other signal (including each other) can result in component malfunction or incompatibility with future processors. See Section 5 for a pin listing for the processor and the location of all RESERVED pins. For reliable operation, always terminate unused inputs or bidirectional signals to their respective deasserted states. On-die termination has been included on the processor to allow signals to be terminated within the processor silicon. Most unused AGTL+ inputs may be left as no-connects since AGTL+ termination is provided on the processor silicon. See Table 2-6 for details on AGTL+ signals that do not include on-die termination. Unused active-high inputs should be connected through a resistor to ground (VSS). Unused outputs may be left unconnected. However, this may interfere with some TAP functions, complicate debug probing, and prevent boundary scan testing. A resistor must be used when tying bidirectional signals to power or ground. When tying any signal to power or ground, a resistor will also allow for system testability. For unused AGTL+ input or I/O signals, use pull-up resistors of the same value as the on-die termination resistors (RTT). See Table 2-15. Most TAP signals, GTL+ asynchronous inputs, and GTL+ asynchronous outputs do not include ondie termination (see Table 2-6 for those signals which do not have on-die termination). Inputs and used outputs must be terminated on the system board. Unused outputs may be terminated on the system board or left connected. Note that leaving unused outputs unterminated may interfere with some TAP functions, complicate debug probing, and prevent boundary scan testing. The TESTHI pins should be tied to VTT using a matched resistor, where a matched resistor has a resistance value within ±20% of the impedance of the board transmission line traces. For example, if the trace impedance is 50 W, then a value between 40 W and 60 W is required. The TESTHI pins may use individual pull-up resistors or be grouped together as detailed below. Please note that utilization of boundary scan test will not be functional if pins are connected together. A matched resistor should be used for each group: • TESTHI[3:0] • TESTHI[6:5] • TESTHI4 — cannot be grouped with other TESTHI signals 2.5 Mixing Processors Intel supports and validates multi-processor configurations in which all processors operate with the same front side bus frequency and internal cache sizes. Intel does not support or validate operation of processors with different cache sizes. Mixing different processor steppings but the same model (as per the CPUID instruction) is supported. Details on CPUID are provided in the Processor BIOS Writers Guide document and the AP-485 Intel® Processor Identification and the CPUID Instruction application note. 24 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Electrical Specifications 2.6 Front Side Bus Signal Groups The front side bus signals are grouped by buffer type as listed in Table 2-5. The buffer type indicates which AC and DC specifications apply to the signals. AGTL+ input signals have differential input buffers that use GTLREF as a reference level. In this document, the term “AGTL+ Input” refers to the AGTL+ input group as well as the AGTL+ I/O group when receiving. Similarly, “AGTL+ Output” refers to the AGTL+ output group as well as the AGTL+ I/O group when driving. AGTL+ asynchronous outputs can become active anytime and include an active pMOS pull-up transistor to assist during the first clock of a low-to-high voltage transition. Implementing a source synchronous data bus requires specifying two sets of timing parameters. One set is for common clock signals which are dependent upon the rising edge of BCLK0 (ADS#, HIT#, HITM#, etc.). The second set is for the source synchronous signals that are relative to their respective strobe lines (data and address) as well as the rising edge of BCLK0. Asynchronous signals are present (A20M#, IGNNE#, etc.) and can become active at any time during the clock cycle. Table 2-5 identifies signals as common clock, source synchronous, and asynchronous. Table 2-5. Front Side Bus Pin Groups (Sheet 1 of 2) Signals1 Signal Group Type AGTL+ Common Clock Input Synchronous to BCLK[1:0] BPRI#, BR[3:1]#, DEFER#, ID[7:0]#, IDS#, OOD#, RESET#, RS[2:0]#, RSP#, TRDY# AGTL+ Common Clock I/O Synchronous to BCLK[1:0] ADS#, AP[1:0]#, BINIT#, BNR#, BPM[5:0]#, BR0#, DBSY#, DP[3:0]#, DRDY#, HIT#, HITM#, LOCK#, MCERR# Signals AGTL+ Source Synchronous I/O Synchronous to associated strobe Associated Strobe REQ[4:0]#, A[37:36,16:3]# ADSTB0# A[39:38,35:17]# ADSTB1# D[15:0]#, DEP[1:0]#, DBI0# DSTBP0#, DSTBN0# D[31:16]#, DEP[3:2]#, DBI1# DSTBP1#, DSTBN1# D[47:32]#, DEP[5:4]#, DBI2# DSTBP2#, DSTBN2# D[63:48]#, DEP[7:6]#, DBI3# DSTBP3#, DSTBN3# AGTL+ Strobe Input/Output Synchronous to BCLK[1:0] ADSTB[1:0]#, DSTBP[3:0]#, DSTBN[3:0]# AGTL+ Asynchronous Output Asynchronous FERR#/PBE#, IERR#, PROCHOT# GTL+ Asynchronous Input Asynchronous A20M#, FORCEPR#, IGNNE#, INIT#, LINT0/ INTR, LINT1/NMI, SMI#, STPCLK# GTL+ Asynchronous Output Asynchronous THERMTRIP# TAP Input Synchronous to TCK TCK, TDI, TMS TAP Input Asynchronous TRST# TAP Output Synchronous to TCK TDO 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 25 Electrical Specifications Table 2-5. Front Side Bus Pin Groups (Sheet 2 of 2) Signal Group Signals1 Type Front Side Bus Clock Input Clock BCLK[1:0] SMBus Synchronous to SM_CLK SM_ALERT#, SM_CLK, SM_DAT, SM_EP_A[2:0], SM_TS_A[1:0], SM_WP Power/Other BOOT_SELECT, BSEL[1:0], COMP0, CVID[3:0], GTLREF[3:0], ODTEN, PWRGOOD, RESERVED, SKTOCC#, SLEW_CTRL, SM_VCC, TEST_BUS, TESTHI[6:0], VCACHE, VCC, VCCA, VCCA_CACHE, VCC_CACHE_SENSE, VCCIOPLL, VCCPLL, VCCSENSE, VID[5:0], VIDPWRGD, VSS, VSSA, VSSA_CACHE, VSS_CACHE_SENSE, VSSSENSE, VTT, VTTEN Power/Other NOTES: 1. Refer to Section 6.1 for signal descriptions. Table 2-6. Signal Description Table Signals with RTT1 A[39:3]#, ADS#, ADSTB[1:0]#, AP[1:0]#, BINIT#, BNR#, BOOT_SELECT2, BPRI#, D[63:0]#, DBI[3:0]#, DBSY#, DEFER#, DEP[7:0]#, DP[3:0]#, DRDY#, DSTBN[3:0]#, DSTBP[3:0]#, HIT#, HITM#, ID[7:0]#, IDS#, LOCK#, MCERR#, OOD#, REQ[4:0]#, RS[2:0]#, RSP#, TRDY# Signals with RL BINIT#, BNR#, HIT#, HITM#, MCERR# NOTES: 1. Signals not included in the “Signals with RTT” list require termination on the baseboard. Please refer to Table 2-5 for the signal type and Table 2-13 to Table 2-18 for the corresponding DC specifications. 2. The BOOT_SELECT pin is not terminated to RTT. It has a 500-5000 Ω internal pullup. The ODTEN signals enables or disables RTT. Those signals affected by ODTEN still present RTT termination to the signal’s pin when the processor is placed in tri-state mode. Furthermore, the following signals are not affected when the processor is placed in tri-state mode: BSEL[1:0], CVID[3:0], SKTOCC#, SM_ALERT#, SM_CLK, SM_DAT, SM_EP_A[2:0], SM_TS_A[1:0], SM_WP, TEST_BUS, TESTHI[6:0], VID[5:0], and VTTEN. Table 2-7. Signal Reference Voltages GTLREF A20M#, A[39:3]#, ADS#, ADSTB[1:0]#, AP[1:0]#, BINIT#, BNR#, BPM[5:0]#, BPRI#, BR[3:0]#, D[63:0]#, DBI[3:0]#, DBSY#, DEFER#, DEP[7:0]#, DP[3:0]#, DRDY#, DSTBN[3:0]#, DSTBP[3:0]#, FORCEPR#, HIT#, HITM#, ID[7:0]#, IDS#, IGNNE#, INIT#, LINT0/INTR, LINT1/NMI, LOCK#, MCERR#, ODTEN, OOD#, REQ[4:0]#, RESET#, RS[2:0]#, RSP#, SMI#, STPCLK#, TRDY# VTT / 2 BOOT_SELECT, PWRGOOD1, TCK1, TDI1, TMS1, TRST#1, VIDPWRGD NOTES: 1. These signals also have hysteresis added to the reference voltage. See Table 2-16 for more information. 26 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Electrical Specifications 2.7 GTL+ Asynchronous and AGTL+ Asynchronous Signals The processor does not utilize CMOS voltage levels on any signals that connect to the processor silicon. As a result, inputs signals such as A20M#, FORCEPR#, IGNNE#, INIT#, LINT0/INTR, LINT1/NMI, SMI#, and STPCLK# utilize GTL buffers. Legacy output THERMTRIP# utilizes a GTL+ output buffer. All of these asynchronous signals follow the same DC requirements as GTL+ signals; however, the outputs are not driven high (during the logical 0-to-1 transition) by the processor. FERR#/PBE#, IERR#, and PROCHOT# have now been defined as AGTL+ asynchronous signals as they include an active pMOS device. GTL+ asynchronous and AGTL+ asynchronous signals do not have setup or hold time specifications in relation to BCLK[1:0]. However, all of the GTL+ asynchronous and AGTL+ asynchronous signals are required to be asserted/deasserted for at least six BCLKs in order for the processor to recognize the proper signal state, except during power-on configuration (see Table 2-24 for the proper specifications at RESET). See Table 2-17 and Table 2-23 for the DC and AC specifications for the GTL+ asynchronous and AGTL+ asynchronous signal groups. 2.8 Test Access Port (TAP) Connection Due to the voltage levels supported by other components in the TAP logic, Intel recommends that the processor(s) be first in the TAP chain, followed by any other components within the system. Use of a translation buffer to connect to the rest of the chain is recommended unless one of the other components is capable of accepting an input of the appropriate voltage. Similar considerations must be made for TCK, TMS, TRST#, TDI, and TDO. Two copies of each signal may be required, each driving a different voltage level. 2.9 Maximum Ratings Table 2-8 specifies absolute maximum and minimum ratings. Within functional operation limits, functionality and long-term reliability can be expected. At conditions outside functional operation condition limits, but within absolute maximum and minimum ratings, neither functionality nor long-term reliability can be expected. If a device is returned to conditions within functional operation limits after having been subjected to conditions outside these limits, but within the absolute maximum and minimum ratings, the device may be functional, but with its lifetime degraded depending on exposure to conditions exceeding the functional operation condition limits. At conditions exceeding absolute maximum and minimum ratings, neither functionality nor longterm reliability can be expected. Moreover, if a device is subjected to these conditions for any length of time then, when returned to conditions within the functional operating condition limits, it will either not function, or its reliability will be severely degraded. Although the processor contains protective circuitry to resist damage from static electric discharge, precautions should always be taken to avoid high static voltages or electric fields. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 27 Electrical Specifications Table 2-8. Processor Absolute Maximum Ratings Symbol Parameter Min Max Unit VCC Processor core supply voltage with respect to VSS -0.3 1.55 V VCACHE Processor L3 cache voltage with respect to VSS -0.3 1.55 V VTT Front side bus termination voltage with respect to VSS -0.3 1.55 V TCASE Processor case temperature See Section 7 See Section 7 °C TSTORAGE Processor storage temperature -40 85 °C Notes1,2 3, 4 NOTES: 1. For functional operation, all processor electrical, signal quality, mechanical, and thermal specifications must be satisfied. 2. Overshoot and undershoot voltage guidelines for input, output, and I/O signals are outlined in Section 3. Excessive overshoot or undershoot on any signal will likely result in permanent damage to the processor. 3. Storage temperature is applicable to storage conditions only. In this scenario, the processor must not receive a clock, and no pins can be connected to a voltage bias. Storage within these limits will not affect the longterm reliability of the device. For functional operation, please refer to the processor case temperature specifications. 4. This rating applies to the processor and does not include any packaging or trays. 2.10 Processor DC Specifications The following notes apply: • The processor DC specifications in this section are defined at the processor core silicon and not at the package pins unless noted otherwise. • The notes associated with each parameter are part of the specification for that parameter. • Unless otherwise noted, all specifications in the tables apply to all frequencies and cache sizes. • Unless otherwise noted, all the specifications in the tables are based on estimates and simulations. These specifications will be updated with characterized data from silicon measurements at a later date. See Section 6 for the pin signal definitions. Most of the signals on the processor front side bus are in the AGTL+ signal group. The DC specifications for these signals are listed in Table 2-15. Table 2-9 through Table 2-18 list the DC specifications for the processor and are valid only while meeting specifications for case temperature, clock frequency, and input voltages. 2.10.1 Flexible Motherboard (FMB) Guidelines The FMB guidelines are estimates of the maximum values that the processor will have over certain time periods. The values are only estimates as actual specifications for future processors may differ. The processor may or may not have specifications equal to the FMB value in the foreseeable future. System designers should meet the FMB values to ensure that their systems will be compatible with future releases of the processor. 28 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Electrical Specifications Table 2-9. Symbol Voltage and Current Specifications Parameter Core Freq Min Typ Max Refer to VID Unit 1.3875 V Notes 1,2,3, VCC VCC for processor core VID Transition VID step size during transition All freq. ± 12.5 mV 18 Total allowable DC load line shift from VID steps All freq. 450 mV 19 VCC for processor L3 cache All freq. 1.275 V VCACHE FMB Table 2-10 1.125 CVID 4,5,7 17 11,12, VTT FSB termination voltage (DC specification) All freq. 1.176 1.20 1.224 V VTT FSB termination voltage (AC specification) All freq. 1.140 1.20 1.260 V SM_VCC SMBus supply voltage All freq. 3.135 3.300 3.465 V 13 ICC ICC for processor core FMB 91 A 7,10 ICC_TDC Thermal Design Current (TDC) FMB 86 A 20 ICACHE ICC for processor L3 cache All freq 24 A ITT FSB termination current, post power good All freq. 4 A 11,15,21 ITT FSB mid-agent current, post power good All freq. 1.3 A 11,16,21 ISM_VCC ICC for SMBus supply All freq. 122.5 mA 11 ISGnt_CORE ICC Stop-Grant Core All freq. 56 A 6,9 ISGnt_CACHE ICC Stop-Grant Cache All freq. 23 A 6,9 ITCC ICC TCC active All freq. ICC A 8 ICC VCCA ICC for PLL pin All freq. 60 mA ICC VCCIOPLL ICC for I/O PLL pin All freq. 60 mA ICC VCCA_CACHE ICC for L3 cache PLL pin All freq. 60 mA ICC GTLREF ICC per GTLREF pin All freq. 200 µA 100 13 11,12, 13,14 NOTES: 1. These voltages and frequencies are targets only. A variable voltage source should exist on systems in the event that a different voltage is required. See Section 2.2 and Table 2-3 for more information. 2. The voltage specification requirements are measured across the VCCSENSE and VSSSENSE pins using an oscilloscope set to a 100 MHz bandwidth and probes that are 1.5 pF maximum capacitance and 1 MΩ minimum impedance at the processor socket. The maximum length of ground wire on the probe should be less than 5 mm. Ensure external noise from the system is not coupled into the scope probe. 3. Refer to Table 2-10 for the minimum, typical, and maximum VCC allowed for a given current. The processor should not be subjected to any VCC and ICC combination wherein VCC exceeds VCC_MAX for a given current. 4. Moreover, VCC should never exceed the VID voltage. Failure to adhere to this specification can shorten the processor lifetime. 5. VCC_MIN and VCC_MAX are defined at the frequency’s associated ICC_MAX on the VCC load line. 6. The current specified is also for the HALT State. 7. FMB is the Flexible Motherboard guideline. These guidelines are for estimation purposes only. See Section 2.10.1 for further details on FMB guidelines. 8. The maximum instantaneous current the processor will draw while the thermal control circuit (TCC) is active as indicated by the assertion of PROCHOT# is the same as the maximum ICC for the processor. 9. The core and cache portions of Stop-Grant current is specified at VCC and VCACHE max. 10.ICC_MAX is specified at the relative VCC_MAX point on the VCC load line. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 29 Electrical Specifications 11.These parameters are based on design characterization and are not tested. 12.VTT must be provided via a separate voltage source and must not be connected to VCC. 13.These specifications are measured at the package pin. 14.Baseboard bandwidth is limited to 20 MHz. 15.This specification refers to a single processor with RTT enabled. 16.This specification refers to a single processor with RTT disabled. 17.The voltage specification requirements are measured across the VCC_CACHE_SENSE and VSS_CACHE_SENSE pins at the socket with a 100 MHz bandwidth oscilloscope, 1.5 pF maximum probe capacitance, and 1 MΩ minimum impedance. The maximum length of ground wire on the probe should be less than 5 mm. Ensure external noise from the system is not coupled into the scope probe. 18.This specification represents the VCC reduction due to each VID transition. See Section 2.2. 19.This specification refers to the total reduction of the load line due to VID transitions below the specified VID. 20.ICC_TDC is the sustained (DC equivalent) current that the processor is capable of drawing indefinitely and should be used for the voltage regulator temperature assessment. The voltage regulator is responsible for monitoring its temperature and asserting the necessary signal to inform the processor of a thermal excursion. The processor is capable of drawing ICC_TDC indefinitely. Refer to Figure 2-3 for further details on the average processor current draw over various time durations. This parameter is based on design characterization and is not tested. 21.ITT may draw up to 5A prior to power good assertion. Figure 2-3. Processor Load Current vs. Time 155 Sustained Current (A) 150 145 140 135 130 125 0.01 0.1 1 10 100 1000 Time Duration (s) NOTES: 1. Processor or voltage regulator thermal protection circuitry should not trip for load currents greater than ICC_TDC. 2. Not 100% tested. Specified by design characterization. 30 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Electrical Specifications Table 2-10. VCC Static and Transient Tolerance ICC [A] VCC_MAX [V] VCC_TYP [V] VCC_MIN [V] Notes 0 VID - 0.000 VID - 0.020 VID - 0.040 1,2,3 5 VID - 0.006 VID - 0.026 VID - 0.046 1,2,3 10 VID - 0.013 VID - 0.033 VID - 0.053 1,2,3 15 VID - 0.019 VID - 0.039 VID - 0.059 1,2,3 20 VID - 0.025 VID - 0.045 VID - 0.065 1,2,3 25 VID - 0.031 VID - 0.051 VID - 0.071 1,2,3 30 VID - 0.038 VID - 0.058 VID - 0.078 1,2,3 35 VID - 0.044 VID - 0.064 VID - 0.084 1,2,3 40 VID - 0.050 VID - 0.070 VID - 0.090 1,2,3 45 VID - 0.056 VID - 0.076 VID - 0.096 1,2,3 50 VID - 0.063 VID - 0.083 VID - 0.103 1,2,3 55 VID - 0.069 VID - 0.089 VID - 0.109 1,2,3 60 VID - 0.075 VID - 0.095 VID - 0.115 1,2,3 65 VID - 0.081 VID - 0.101 VID - 0.121 1,2,3 70 VID - 0.087 VID - 0.108 VID - 0.128 1,2,3 75 VID - 0.094 VID - 0.114 VID - 0.134 1,2,3 80 VID - 0.100 VID - 0.120 VID - 0.140 1,2,3 85 VID - 0.106 VID - 0.126 VID - 0.146 1,2,3 90 VID - 0.113 VID - 0.133 VID - 0.153 1,2,3 95 VID - 0.119 VID - 0.139 VID - 0.159 1,2,3 100 VID - 0.125 VID - 0.145 VID - 0.165 1,2,3 NOTES: 1. The VCC_MIN and VCC_MAX load lines represent static and transient limits. 2. This table is intended to aid in reading discrete points on Figure 2-4. 3. The load lines specify voltage limits at the die measured at the VCCSENSE and VSSSENSE pins. Voltage regulation feedback for voltage regulator circuits must be taken from processor VCC and VSS pins. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 31 Electrical Specifications Figure 2-4. VCC Static and Transient Tolerance Icc [A] 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 VID - 0.000 VID - 0.020 VCC Maximum VID - 0.040 Vcc [V] VID - 0.060 VID - 0.080 VID - 0.100 VID - 0.120 VCC Typical VID - 0.140 VCC Minimum VID - 0.160 VID - 0.180 NOTES: 1. The VCC_MIN and VCC_MAX load lines represent static and transient limits. 2. Refer to Table 2-9 for processor VID information for VCC. 3. The load lines specify voltage limits at the die measured at the VCCSENSE and VSSSENSE pins. Voltage regulation feedback for voltage regulator circuits must also be taken from processor VCCSENSE and VSSSENSE pins. 2.10.2 VCC and VCACHE Overshoot Specification The processor can tolerate short transient overshoot events where VCC exceeds the VID voltage or where VCACHE exceeds the CVID voltage when transitioning from a high-to-low current load condition. This overshoot cannot exceed VID + VOS_MAX, or CVID + VOS_MAX. (VOS_MAX is the maximum allowable overshoot above VID or CVID). These specifications apply to the processor die voltage as measured across the VCCSENSE and VSSSENSE pins for Vcc, and VCC_CACHE_SENSE and VSS_CACHE_SENSE pins for CVID. Table 2-11. VCC and VCACHE Overshoot Specification Symbol 32 Parameter VOS_MAX Magnitude of VCC overshoot above VID or VCACHE overshoot above CVID TOS_MAX Time duration of VCC overshoot above VID or VCACHE overshoot above CVID Min Max Units Figure 0.025 V 2-5 5 µs 2-5 Notes 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Electrical Specifications Voltage Figure 2-5. VCC and VCACHE Overshoot Example Waveform VOS: Peak maximum overshoot identified during validation TOS(OS time above VID or CVID): Time from initial VID or CVID crossing to final crossing of VID or CVID Time Overshoot Waveform 2.10.3 VID/CVID Reference Die Voltage Validation Overshoot events from application testing on the processor must meet the specifications in Table 2-11 when measured across the VCCSENSE and VSSSENSE pins. Overshoot events that are < 10 ns in duration may be ignored. These measurements of processor die level overshoot should be taken with a 100 MHz bandwidth limited oscilloscope. Table 2-12. Front Side Bus Differential BCLK Specifications Symbol Parameter Min Typ Max Unit Figure Notes VL Input Low Voltage -0.150 0.000 N/A V 2-8 VH Input High Voltage 0.660 0.700 0.850 V 2-8 VCROSS(abs) Absolute Crossing Point 0.250 N/A 0.550 V 2-8, 2-9 1,7 VCROSS(rel) Relative Crossing Point 0.250 + 0.5* (VHavg - 0.700) N/A 0.550 + 0.5* (VHavg - 0.700) V 2-8, 2-9 2,7,8 ∆ VCROSS Range of Crossing Point N/A N/A 0.140 V 2-8, 2-9 VOV Overshoot N/A N/A + 0.300 V 2-8 3 VUS Undershoot - 0.300 N/A N/A V 2-8 4 VRBM Ringback Margin 0.200 N/A N/A V 2-8 5 VTM Threshold Margin VCROSS-0.100 VCROSS+0.100 V 2-8 6 NOTES: 1. Crossing voltage is defined as the instantaneous voltage value when the rising edge of BCLK0 is equal to the falling edge of BCLK1. 2. VHavg is the statistical average of the VH measured by the oscilloscope. 3. Overshoot is defined as the absolute value of the maximum voltage. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 33 Electrical Specifications 4. Undershoot is defined as the absolute value of the minimum voltage. 5. Ringback Margin is defined as the absolute voltage difference between the maximum Rising Edge Ringback and the maximum Falling Edge Ringback. 6. Threshold Region is defined as a region entered around the crossing point voltage in which the differential receiver switches. It includes input threshold hysteresis. 7. The crossing point must meet the absolute and relative crossing point specifications simultaneously. 8. VHavg can be measured directly using “Vtop” on Agilent scopes and “High” on Tektronix scopes. Table 2-13. BSEL[1:0], VID[5:0], and CVID[3:0] DC Specifications Symbol Parameter Max Unit Notes RON Buffer On Resistance 60 Ω 1 IOL Max Pin Current 8 mA ILO Output Leakage Current 200 µA 2 Voltage Tolerance 3.3 * 1.05 V 3 VTOL NOTES: 1. These parameters are not tested and are based on design simulations. 2. Leakage to VSS with pin held at 2.5V. 3. Represents the maximum allowable termination voltage. Table 2-14. VIDPWRGD DC Specifications Symbol 34 Parameter Min Max Unit Figure VIL Input Low Voltage 0.0 0.30 V 2-19 VIH Input High Voltage 0.90 VTT V 2-19 Notes 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Electrical Specifications Table 2-15. AGTL+ Signal Group DC Specifications Symbol Parameter Min Max Unit Notes VIL Input Low Voltage 0.0 GTLREF - (0.10 * VTT) V 1,5 VIH Input High Voltage GTLREF + (0.10 * VTT) VTT V 2,3,5 VOH Output High Voltage 0.90 * VTT VTT V 3,5 IOL Output Low Current N/A VTT / (0.50 * Rtt_min + RON_min || RL) mA 7 ILI Input Leakage Current N/A ± 200 µA 6 ILO Output Leakage Current N/A ± 200 µA 8 RON Buffer On Resistance 8 12 W 4 NOTES: 1. VIL is defined as the voltage level at a receiving agent that will be interpreted as a logical low value. 2. VIH is defined as the voltage level at a receiving agent that will be interpreted as a logical high value. 3. VIH and VOH may experience excursions above VCC. However, input signal drivers must comply with the signal quality specifications in Section 3. 4. Refer to Processor Signal Integrity Models for I/V characteristics. 5. The VTT referred to in these specifications refers to the instantaneous VTT. 6. Leakage to VSS with pin held at VTT. 7. The maximum output current is based on maximum current handling capability of the buffer and is not specified into the test load. 8. Leakage to VTT with pin held at 300 mV. Table 2-16. PWRGOOD and TAP Signal Group DC Specifications Parameter Min Max Unit Notes1 Input Hysteresis 200 350 mV 6 VT+ Input Low to High Threshold Voltage 0.5 * (VTT + VHYS_MIN) 0.5 * (VTT + VHYS_MAX) V 4 VT- Input High to Low Threshold Voltage 0.5 * (VTT - VHYS_MAX) 0.5 * (VTT - VHYS_MIN) V 4 VOH Output High Voltage N/A VTT V 2,4 IOL Output Low Current 45 mA 5 ILI Input Leakage Current ±200 µA ILO Output Leakage Current ±200 µA RON Buffer On Resistance 12 Ω Symbol VHYS 8 3 NOTES: 1. All outputs are open drain. 2. TAP signal group must meet system signal quality specification in Section 3. 3. Refer to the Processor Signal Integrity Models for I/V characteristics. 4. The VTT referred to in these specifications refers to instantaneous VTT. 5. The maximum output current is based on maximum current handling capability of the buffer and is not specified into the test load. 6. VHYS represents the amount of hysteresis, nominally centered about 0.5 * VTT for all TAP inputs. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 35 Electrical Specifications Table 2-17. GTL+ and AGTL+ Asynchronous Signal Group DC Specifications Symbol Parameter Min Max Unit Notes VIL Input Low Voltage 0 GTLREF - (10% * VTT) V 2 VIH Input High Voltage GTLREF + (10% * VTT) VTT V 3,4,6 VOH Output High Voltage VTT V 1,4,6 IOL Output Low Current 50 mA 7 ± 200 µA 8 ± 200 µA 9 12 Ω 5 ILI Input Leakage Current ILO Output Leakage Current Ron Buffer On Resistance N/A 8 NOTES: 1. All outputs are open-drain. 2. VIL is defined as the voltage range at a receiving agent that will be interpreted as a logical low value. 3. VIH is defined as the voltage range at a receiving agent that will be interpreted as a logical high value. 4. VIH and VOH may experience excursions above VTT. However, input signal drivers must comply with the signal quality specifications in Section 3. 5. Refer to the Processor Signal Integrity Models for I/V characteristics. 6. The VTT referred to in these specifications refers to instantaneous VTT. 7. The maximum output current is based on maximum current handling capability of the buffer and is not specified into the test load. 8. Leakage to VSS with pin held at VTT 9. Leakage to VTT with pin held at 300 mV. Table 2-18. SMBus Signal Group DC Specifications Symbol Parameter Min Max Unit VIL Input Low Voltage -0.30 0.30 * SM_VCC V VIH Input High Voltage 0.70 * SM_VCC 3.465 V VOL Output Low Voltage 0 0.400 V IOL Output Low Current N/A 3.0 mA ILI Input Leakage Current N/A ± 10 µA ILO Output Leakage Current N/A ± 10 µA CSMB SMBus Pin Capacitance 15.0 pF Notes 1,2 3 NOTES: 1. These parameters are based on design characterization and are not tested. 2. All DC specifications for the SMBus signal group are measured at the processor pins. 3. Platform designers may need this value to calculate the maximum loading of the SMBus and to determine maximum rise and fall times for SMBus signals. 36 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Electrical Specifications 2.11 AGTL+ Front Side Bus Specifications Termination resistors are not required for most AGTL+ signals because they are integrated into the processor silicon. Valid high and low levels are determined by the input buffers which compare a signal’s voltage with a reference voltage called GTLREF. Table 2-19 lists the GTLREF specifications. GTLREF should be generated on the system board using high-precision voltage divider circuits. Table 2-19. AGTL+ Bus Voltage Definitions Symbol Parameter Min Typ Max Units Notes 0.98 * (0.63 * VTT) 0.63 * VTT 1.02 * (0.63 * VTT) V 1,2,6 GTLREF Bus Reference Voltage RTT Termination Resistance (pull-up) 40.5 45 49.5 Ω 3 RL Termination Resistance (pull-down) 360 450 540 Ω 4 COMP0 COMP Resistance 49.4 49.9 50.4 Ω 5 SLEW_CTRL SLEW_CTRL Resistance 49.4 49.9 50.4 Ω 7 NOTES: 1. The tolerances for this specification have been stated generically to enable system designers to calculate the minimum values across the range of VTT. 2. GTLREF is generated from VTT on the baseboard by a voltage divider of 1% resistors. 3. RTT is the on-die termination resistance measured at VOL of the AGTL+ output driver. 4. RL is the on-die termination resistance for improved noise margin and signal integrity. 5. The COMP0 resistor is provided by the baseboard with 1% resistors. 6. The VTT referred to in these specifications refers to instantaneous VTT. 7. The SLEW_CTRL resistor is provided by the baseboard with 1% resistors. 2.12 Front Side Bus AC Specifications The processor front side bus timings specified in this section are defined at the processor core silicon and are thus not measurable at the processor pins. See Section 6 for processor pin signal definitions. Table 2-20 through Table 2-27 list the AC specifications associated with the processor front side bus. All AGTL+ timings are referenced to GTLREF for both ‘0’ and ‘1’ logic levels unless otherwise specified. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 37 Electrical Specifications Table 2-20. Front Side Bus Differential Clock Specifications T# Parameter Min Nom Max Unit FSB Clock Frequency 165.78 166.72 MHz T1: BCLK[1:0] Period 5.9982 6.0320 ns T2: BCLK[1:0] Period Stability Figure Notes1 2-8 2 175 ps 3,4 T3: BCLK[1:0] Rise Time 175 700 ps 5 T4: BCLK[1:0] Fall Time 175 700 ps 5 NOTES: 1. Unless otherwise noted, all specifications in this table apply to all processor core frequencies based on a 166 MHz BCLK[1:0]. 2. The period specified here is the average period. A given period may vary from this specification as governed by the period stability specification (T2). Min period specification is based on -300 PPM deviation from a 6 ns period. Max period specification is based on the summation of +300 PPM deviation from a 6 ns period and a +0.5% maximum variance due to spread spectrum clocking. 3. For the clock jitter specification, refer to the applicable clock driver design specification. 4. In this context, period stability is defined as the worst case timing difference between successive crossover voltages. In other words, the largest absolute difference between adjacent clock periods must be less than the period stability. 5. Rise and fall times are measured single-ended between 245 mV and 455 mV of the clock swing. . Table 2-21. Front Side Bus Common Clock AC Specifications T# Parameter Min Max Unit Figure Notes 1, 2 T10: Common Clock Output Valid Delay -0.125 1.470 ns 2-10 3 T11: Common Clock Input Setup Time 0.810 N/A ns 2-10 4 T12: Common Clock Input Hold Time 0.355 N/A ns 2-10 4 T13: RESET# Pulse Width 1.00 10.00 ms 2-18 5,6,7 NOTES: 1. These parameters are based on design characterization and are not tested. 2. All common clock AC timings for AGTL+ signals are referenced to the Crossing Voltage (VCROSS) of the BCLK[1:0] at rising edge of BCLK0. All common clock AGTL+ signal timings are referenced at GTLREF at the processor core. 3. Valid delay timings for these signals are specified into the test circuit described in Figure 2-6 and with GTLREF at 0.63 * VTT ± 2%. 4. Specification is for a minimum swing defined between VIL_MAX to VIH_MIN. This assumes an edge rate of 0.9 V/ns to 1.2 V/ns. 5. RESET# can be asserted (active) asynchronously, but must be deasserted synchronously. 6. This should be measured after VTT and BCLK[1:0] become stable. 7. Maximum specification applies only while PWRGOOD is asserted. . Table 2-22. Front Side Bus Source Synchronous AC Specifications (Sheet 1 of 2) 38 T# Parameter Min T20: Source Sync. Output Valid Delay (first data/address only) -0.150 T21: TVBD Source Sync. Data Output Valid Before Data Strobe T22: TVAD Source Sync. Data Output Valid After Data Strobe Typ Notes Max Unit Figure 1.400 ns 2-11,2-12 4 0.400 ns 2-12 4,7 0.400 ns 2-12 4,8 1,2,3 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Electrical Specifications Table 2-22. Front Side Bus Source Synchronous AC Specifications (Sheet 2 of 2) T# Parameter Min Typ Max Unit Figure Notes 1,2,3 T23: TVBA Source Sync. Address Output Valid Before Address Strobe 1.120 ns 2-11 4,7 T24: TVAA Source Sync. Address Output Valid After Address Strobe 1.120 ns 2-11 4,8 T25: TSUSS Source Sync. Input Setup Time to Strobe 0.100 ns 2-11,2-12 5 T26: THSS Source Sync. Input Hold Time to Strobe 0.100 ns 2-11,2-12 5 T27: TSUCC Source Sync. Input Setup Time to BCLK[1:0] 0.910 ns 2-11,2-12 6 T29: TFASS First Address Strobe to Second Address Strobe 1/2 BCLKs 2-11 9 T30: TFDSS: First Data Strobe to Subsequent Strobes n/4 BCLKs 2-12 10,11 12 T31: Data Strobe ‘n’ (DSTBN#) Output Valid Delay 5.100 6.650 ns 2-12 T32: Address Strobe Output Valid Delay 1.350 2.900 ns 2-11 NOTES: 1. Not 100% tested. These parameters are based on design characterization. 2. All source synchronous AC timings are referenced to their associated strobe at GTLREF. Source synchronous data signals are referenced to the falling edge of their associated data strobe. Source synchronous address signals are referenced to the rising and falling edge of their associated address strobe. All source synchronous AGTL+ signal timings are referenced at GTLREF at the processor core. 3. Unless otherwise noted, these specifications apply to both data and address timings. 4. Valid delay timings for these signals are specified into the test circuit described in Figure 2-6 and with GTLREF at 0.63 * VTT ± 2%. 5. Specification is for a minimum swing defined between VIL_MAX to VIH_MIN. AC timings are specified as GTLREF ± (0.06 * VTT). This assumes an edge rate of 0.9 V/ns to 1.2 V/ns. 6. All source synchronous signals must meet the specified setup time to BCLK as well as the setup time to each respective strobe. 7. This specification represents the minimum time the data or address will be valid before its strobe. 8. This specification represents the minimum time the data or address will be valid after its strobe. 9. The rising edge of ADSTB# must come approximately 1/2 BCLK period after the falling edge of ADSTB#. 10.For this timing parameter, n = 1, 2, and 3 for the second, third, and last data strobes respectively. 11.The second data strobe (falling edge of DSTBn#) must come approximately 1/4 BCLK period after the first falling edge of DSTBp#. The third data strobe (falling edge of DSTBp#) must come approximately 1/2 BCLK period after the first falling edge of DSTBp#. The last data strobe (falling edge of DSTBn#) must come approximately 3/4 BCLK period after the first falling edge of DSTBp#. 12.This specification applies only to DSTBN[3:0]# and is measured to the second falling edge of the strobe. Table 2-23. Miscellaneous Signals AC Specifications (Sheet 1 of 2) T# Parameter T35: GTL+ asynchronous and AGTL+ asynchronous input pulse width Min Max 6 T36: PWRGOOD to RESET# deassertion time 1 T37: BCLK valid before PWRGOOD active 10 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Unit Figure BCLKs 10 Notes 1,2,5 7 ms 2-18 BCLKs 2-18 3 39 Electrical Specifications Table 2-23. Miscellaneous Signals AC Specifications (Sheet 2 of 2) T# Parameter T38: PROCHOT#, FORCEPR# pulse width Min Max Figure µs 2-14 4 500 ms 2-15 6 500 T39: THERMTRIP# assertion until VCC and VCACHE removal Notes Unit T40: FERR# valid delay from STPCLK# deassertion 0 5 BCLKs 2-20 T41: VCC to PWRGOOD assertion time 1 500 ms 2-18 1,2,5 NOTES: 1. All AC timings for the GTL+ asynchronous signals are referenced to the BCLK0 rising edge at Crossing Voltage (VCROSS). All GTL+ asynchronous signal timings are referenced at GTLREF. PWRGOOD is referenced to the BCLK0 rising edge at 0.5 * VTT. 2. These signals may be driven asynchronously. 3. Refer to the PWRGOOD definition for more details regarding the behavior of the signal. 4. Length of assertion for PROCHOT# does not equal TCC activation time. Time is required after the assertion or deassertion of PROCHOT# for the processor to enable or disable the TCC. Additionally, time is allocated after the assertion or deassertion of PROCHOT# for the processor to complete current instruction execution. This specification applies to the PROCHOT# signal when asserted by the processor and the FORCEPR# signal when asserted by the system. 5. Refer to Section 8.2 for additional timing requirements for entering and leaving low power states. 6. Intel recommends the VTT power supply also be removed upon assertion of THERMTRIP#. 7. A minimum pulse width of 500us is recommended when FORCEPR# is asserted by the system. Table 2-24. Front Side Bus AC Specifications (Reset Conditions) T# Parameter Min Max Unit Figure Notes T47: Reset Configuration Signals (A[21:16]#) Setup Time 1 ms T45: Reset Configuration Signals (A[39:22]#, A[15:3]#, BR[3:0]#, INIT#, SMI#) Setup Time 4 BCLKs 2-18 1 T46: Reset Configuration Signals (A[39:3]#, BR[3:0]#, INIT#, SMI#) Hold Time 2 28 BCLKs 2-18 2 Min Max Unit Figure 1 NOTES: 1. Before the clock that de-asserts RESET# 2. After the clock that de-asserts RESET#. Table 2-25. TAP Signal Group AC Specifications T# Parameter Notes 1,7 T55: TCK Period 13.3 ns 2-7 2 T61: TDI, TMS Setup Time 1.5 ns 2-13 3,6 T62: TDI, TMS Hold Time 3.0 ns 2-13 3,6 T63: TDO Clock to Output Delay 0.5 ns 2-13 4 TCK 2-14 5 T64: TRST# Assert Time 2 10.0 NOTES: 1. Not 100% tested. These parameters are based on design characterization. 2. This specification is based on the capabilities of the ITP-XDP debug port tool, not on processor silicon. 3. Referenced to the rising edge of TCK. 4. Referenced to the falling edge of TCK. 5. TRST# must be held asserted for 2 TCK periods to be guaranteed that it is recognized by the processor. 40 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Electrical Specifications 6. Specification for a minimum swing defined between TAP VT- to VT+. This assumes a minimum edge rate of 0.5 V/ns. 7. It is recommended that TMS be asserted while TRST# is being deasserted. Table 2-26. VIDPWRGD and Other Voltage Sequence AC Specifications T# Parameter Min Max T70: VIDPWRGD rise time T71: VTT to VIDPWRGD delay time 1 T72: VTT to VIDPWRGD deassertion time Unit Figure Notes 1 150 ns 2-19 10 ms 2-19 1 ms 2-19 T73: VIDPWRGD to VCACHE delay time 0 ms 2-18 T74: VCACHE to VCC delay time 1 ms 2-18 2 3 NOTES: 1. Rise time is measured between 10% and 90% points on the waveform. 2. Specification refers to the time between VIDPWRGD = VTT - 20% and VIDPWRGD = VIL. 3. VCACHE to VCC delay time is measured from (0.5 * CVID) to (0.5 * VID). Table 2-27. VID Signal Group AC Timing Specifications T# Parameter Min Max Unit Figure T80: VID Step Time 5 µs 2-21, 2-22 T81: VID Dwell Time 50 µs 2-21, 2-22 T82: VID Down Transition to Valid VCC (min) 0 µs 2-21 T82: VID Up Transition to Valid VCC (min) 50 µs 2-21 T82: VID Down Transition to Valid VCC (max) 50 µs 2-22 T82: VID Up Transition to Valid VCC (max) 0 µs 2-22 Notes Table 2-28. SMBus Signal Group AC Specifications T# Parameter Max Unit 10 100 KHz T91: SM_CLK Period 10 100 µs T92: SM_CLK High Time 4.0 N/A µs 2-16 T93: SM_CLK Low Time 4.7 N/A µs 2-16 T94: SMBus Rise Time 0.02 1.0 µs 2-16 4 T95: SMBus Fall Time 0.02 0.3 µs 2-16 4 T96: SMBus Output Valid Delay 0.1 4.5 µs 2-17 T97: SMBus Input Setup Time 250 N/A ns 2-16 T98: SMBus Input Hold Time 300 N/A ns 2-16 T99: Bus Free Time 4.7 N/A µs 2-16 T100: Hold Time after Repeated Start Condition 4.0 N/A µs 2-16 T101: Repeated Start Condition Setup Time 4.7 N/A µs 2-16 T102: Stop Condition Setup Time 4.0 N/A µs 2-16 T90: SM_CLK Frequency 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Figure Notes 1,2 Min 3,5 41 Electrical Specifications NOTES: 1. These parameters are based on design characterization and are not tested. 2. All AC timings for the SMBus signals are referenced at VIL_MAX or VIL_MIN and measured at the processor pins. Refer to Figure 2-16. 3. Minimum time allowed between request cycles. 4. Rise time is measured from (VIL_MAX - 0.15V) to (VIH_MIN + 0.15V). Fall time is measured from (0.9 * SM_VCC) to (VIL_MAX - 0.15V). DC parameters are specified in Table 2-18. 5. Following a write transaction, an internal write cycle time of 10ms must be allowed before starting the next transaction 2.13 Processor AC Timing Waveforms The following figures are used in conjunction with the AC timing tables (Table 2-21 through Table 2-28). Note: For Figure 2-7 through Figure 2-19, the following apply: 1. All common clock AC timings for AGTL+ signals are referenced to the Crossing Voltage (VCROSS) of the BCLK[1:0] at rising edge of BCLK0. All common clock AGTL+ signal timings are referenced at GTLREF at the processor core. 2. All source synchronous AC timings for AGTL+ signals are referenced to their associated strobe (address or data) at GTLREF. Source synchronous data signals are referenced to the falling edge of their associated data strobe. Source synchronous address signals are referenced to the rising and falling edge of their associated address strobe. All source synchronous AGTL+ signal timings are referenced at GTLREF at the processor silicon. 3. All AC timings for AGTL+ strobe signals are referenced to BCLK[1:0] at VCROSS. All AGTL+ strobe signal timings are referenced at GTLREF at the processor silicon. 4. All AC timings for the TAP signals are referenced to the TCK at 0.5 * VTT at the processor pins. All TAP signal timings (TMS, TDI, etc.) are referenced at 0.5 * VTT at the processor pins. 5. All AC timings for the SMBus signals are referenced to the SM_CLK at 0.5 * SM_VCC at the processor pins. All SMBus signal timings (SM_DAT, SM_CLK, etc.) are referenced at VIL_MAX or VIL_MIN at the processor pins. Figure 2-6. Electrical Test Circuit VTT VTT RLOAD = 50 ohms 45 ohms, 156 ps/in, 550 mils AC Timings specified at this point 42 L = 2.4nH C = 1.2pF 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Electrical Specifications Figure 2-7. TCK Clock Waveform CLK V1 Tp Tp = T55: Period V1: TCK is referenced to 0.5 * VTT Figure 2-8. Differential Clock Waveform Overshoot BCLK1 VH Rising Edge Ringback Crossing Voltage Threshold Region Crossing Voltage Ringback Margin Falling Edge Ringback, BCLK0 VL Undershoot Tp Tp = T1: BCLK[1:0] period 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 43 Electrical Specifications Figure 2-9. Differential Clock Crosspoint Specification 650 Crossing Point (mV) 600 550 550 mV 500 450 550 + 0.5 (VHavg - 700) 400 250 + 0.5 (VHavg - 700) 350 300 250 250 mV 200 660 670 680 690 700 710 720 730 740 750 760 770 780 790 800 810 820 830 840 850 VHavg (mV) Figure 2-10. Front Side Bus Common Clock Valid Delay Timing Waveform T0 T1 T2 BCLK1 BCLK0 TP Common Clock Signal (@ driver) valid valid TQ Common Clock Signal (@ receiver) TR valid TP = T10: Common Clock Output Valid Delay TQ = T11: Common Clock Input Setup TR = T12: Common Clock Input Hold Time 44 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Electrical Specifications Figure 2-11. Source Synchronous 2X (Address) Timing Waveform T1 TP/4 TP/2 T2 3TP/4 BCLK1 BCLK0 TQ ADSTB# (@ driver) TR TH A# (@ driver) TJ valid TH TJ valid TK TS ADSTB# (@ receiver) A# (@ receiver) valid valid TM TN TP = T1: BCLK[1:0] Period TH = T23: Source Sync. Address Output Valid Before Address Strobe TJ = T24: Source Sync. Address Output Valid After Address Strobe TK = T27: Source Sync. Input Setup Time to BCLK TM = T25: Source Sync. Input Setup Time TN = T26: Source Sync. Input Hold Time TQ = T29: First Address Strobe to Second Address Strobe TS = T20: Source Sync. Output Valid Delay TR = T32: Address Strobe Output Valid Delay 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 45 Electrical Specifications Figure 2-12. Source Synchronous 4X (Data) Timing Waveform T0 T1 TP/4 TP/2 T2 3TP/4 BCLK1 BCLK0 DSTBp# (@ driver) TH TD DSTBn# (@ driver) TA TB TA TB D# (@ driver) TJ TC DSTBp# (@ receiver) DSTBn# (@ receiver) D# (@ receiver) TE TG TE TG TP = T1: BCLK[1:0] Period TA = T21: Source Sync. Data Output Valid Delay Before Data Strobe TB = T22: Source Sync. Data Output Valid Delay After Data Strobe TC = T27: Source Sync. Input Setup Time to BCLK TD = T31: Data Strobe 'n' (DSTBN#) Output Valid Delay TE = T25: Source Sync. Input Setup Time TG = T26: Source Sync. Input Hold Time TH = T30: First Data Strobe to Subsequent Strobes TJ = T20: Source Sync. Data Output Valid Delay 46 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Electrical Specifications Figure 2-13. TAP Valid Delay Timing Waveform TCK V Tx Ts Th V Valid Signal Tx = T63: TDO Clock to Output Delay Ts = T61: TDI, TMS Setup Time Th = T62: TDI, TMS Hold Time V = 0.5 * VTT Figure 2-14. Test Reset (TRST#), Async GTL+ Input, and PROCHOT# Timing Waveform V Tq T = T64 (TRST# Pulse Width), V = 0.5 * VTT q T38 (PROCHOT# Pulse Width), V = GTLREF Figure 2-15. THERMTRIP# Power Down Sequence TA THERMTRIP# VCC, V CACHE TA = T39: THERMTRIP# assertion until VCC and V CACHE removal 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 47 Electrical Specifications Figure 2-16. SMBus Timing Waveform t t LOW tF R t HD;STA Clk t HD;STA t t HD;DAT t SU;DAT HIGH t SU;STA t SU;STO Data t BUF P STOP S S START START t LOW = T93 t HD;STA = T100 t SU;STA = T101 t HIGH = T92 t HD;DAT = T98 t SU;STD = T102 tR = T94 t BUF = T95 t SU;DAT = T97 tF P STOP = T99 Figure 2-17. SMBus Valid Delay Timing Waveform SM_CLK TAA DATA VALID SM_DAT DATA OUTPUT TAA = T96 48 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Electrical Specifications Figure 2-18. Voltage Sequence Timing Requirements Ta Tb R e s e t C o n fig u ra tio n S ig n a ls Tc B C L K [1 :0 ] Td V TT V ID [5 :0 ] C V ID [3 :0 ] B S E L [1 :0 ] V ID P W R G D Te V CACH E Ti V CC Tf PW RGOOD Tg Th R ESET# T a = T 4 5 / T 4 7 : R e s e t c o n fig u ra tio n s ig n a ls s e tu p tim e T b = T 4 6 : R e s e t c o n fig u ra tio n s ig n a ls h o ld tim e T c = T 3 7 : B C L K V a lid B e fo re P W R G O O D A c tiv e T d = T 7 1 : V T T to V ID P W R G D d e la y tim e T e = T 7 3 : V ID P W R G D to V C A C H E d e la y tim e T f = T 4 1 : V C C to P W R G O O D a s s e rtio n tim e T g = T 3 6 : P W R G O O D to R E S E T # d e -a s s e rtio n tim e T h = T 1 3 : R E S E T # p u ls e w id th T i = T 7 4 : V C A C H E to V C C d e la y tim e 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 49 Electrical Specifications Figure 2-19. VIDPWRGD Timing Requirements 90% 10% 80% Ta VTT VIH VIL VIL VIDPWRGD Tb Tc Ta = T71 VTT to VIDPWRGD delay time Tb = T70 VIDPWRGD rise time Tc = T72 VTT to VIDPWRGD deassertion time Figure 2-20. FERR#/PBE# Valid Delay Timing BCLK SG Ack system bus STPCLK# Ta FERR#/ PBE# FERR# undefined PBE# undefined FERR# Ta = T40: FERR# valid delay from STPCLK# de-assertion Note: FERR#/PBE# is also undefined from STPCLK# assertion until the stop grant acknowledge is driven on the processor system bus. FERR#/PBE# is also undefined for a period of Ta from STPCLK# de-assertion. Inside these undefined regions the PBE# signal is driven. FERR# is driven at all other times. 50 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Electrical Specifications Figure 2-21. VID Step Timings VID n ... n-1 m m+1 Tc VCC(max) Ta Tb Td VCC(min) Ta Tb Tc Td = = = = T84: VID Down to Valid VCC(max) T82: VID Down to Valid VCC(min) T85: VID Up to Valid VCC(max) T83: VID Up to Valid VCC(min) Figure 2-22. VID Step Times and VCC Waveforms n-1 n-2 n-3 n-6 = VIDTM2 n-4 n-5 n-5 n-4 n-3 Tb n-2 n n-1 Ta VID n VCC(max) VCC(max,n-3) Te Tc VCC(min) VCC(max,n-4) Td VCC(min,n-3) Tf VCC(min,n-4) Ta Tb Tc Td Te Tf = = = = = = T80: VID Step Time T81: Thermal Monitor 2 Dwell Time T84: VID Down to Valid VCC(max) T82: VID Down to Valid VCC(min) T85: VID Up to Valid VCC(max) T83: VID Up to Valid VCC(min) Note: This waveform illustrates an example of an Intel Thermal Monitor 2 transition or an Intel Enhanced SpeedStep transition that is six VID steps down from the current state and six steps back up. Any arbitrary up or down transition can be generalized from this waveform. § 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 51 Electrical Specifications 52 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 3 Front Side Bus Signal Quality Specifications Source synchronous data transfer requires the clean reception of data signals and their associated strobes. Ringing below receiver thresholds, non-monotonic signal edges, and excessive voltage swings will adversely affect system timings. Ringback and signal non-monotonicity cannot be tolerated since these phenomena may inadvertently advance receiver state machines. Excessive signal swings (overshoot and undershoot) are detrimental to silicon gate oxide integrity, and can cause device failure if absolute voltage limits are exceeded. Overshoot and undershoot can also cause timing degradation due to the build up of inter-symbol interference (ISI) effects. For these reasons, it is crucial that the designer work towards a solution that provides acceptable signal quality across all systematic variations encountered in volume manufacturing. This section documents signal quality metrics used to derive topology and routing guidelines through simulation. All specifications are specified at the processor core (pad measurements). Specifications for signal quality are for measurements at the processor core only and are only observable through simulation. The same is true for all front side bus AC timing specifications in Section 2.12. Therefore, proper simulation of the processor front side bus is the only way to verify proper timing and signal quality. 3.1 Front Side Bus Signal Quality Specifications and Measurement Guidelines 3.1.1 Ringback Guidelines Many scenarios have been simulated to generate a set of AGTL+ layout guidelines. Table 3-1 provides the signal quality specifications for the AGTL+ and GTL+ asynchronous signal groups. Table 3-2 demonstrates the signal quality specification for the TAP signal group. These specifications are for use in simulating signal quality at the processor pads. Maximum allowable overshoot and undershoot specifications for a given duration of time are detailed in Table 3-3 through Section 3-6. front side bus ringback tolerance for AGTL+ and GTL+ asynchronous signal groups are shown in Figure 3-1 (low-to-high transitions) and Figure 3-2 (high-to-low transitions). The TAP signal group includes hysteresis on the input buffers and thus has relaxed ringback requirements when compared to the other buffer types. Figure 3-3 shows the front side bus ringback tolerance for low-to-high transitions and Figure 3-4 for high-to-low transitions. The hysteresis values Vt+ and Vt- can be found in Table 2-16. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 53 Front Side Bus Signal Quality Specifications Table 3-1. Ringback Specifications for AGTL+ and GTL+ Asynchronous Signal Groups Signal Group Transition Maximum Ringback (with Input Diodes Present) Unit Figure AGTL+, Async GTL+ Low → High GTLREF + (0.06 * VTT) V 3-1 1,2,3,4,5,6 AGTL+, Async GTL+ High → Low GTLREF - (0.06 * VTT) V 3-2 1,2,3,4,5,6 Notes NOTES: 1. All signal integrity specifications are measured at the processor core (pads). 2. Unless otherwise noted, all specifications in this table apply to all processor frequencies. 3. Specifications are for the edge rate of 1.8- 2.2V/ns. 4. All values specified by design characterization. 5. Please see Section 3.1.3 for maximum allowable overshoot. 6. The total ringback tolerance is 6% of VTT (0.06 * VTT). This consists of 4% AC and 2% DC components. Ringback between GTLREF + (0.06 * VTT) and GTLREF - (0.06 * VTT) is not supported. Figure 3-1. Low-to-High Front Side Bus Receiver Ringback Tolerance VTT + 6% * VTT Noise Margin GTLREF – 6% * VTT VSS Figure 3-2. High-to-Low Front Side Bus Receiver Ringback Tolerance VTT + 6% * VTT GTLREF - 6% * VTT Noise Margin VSS 54 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Front Side Bus Signal Quality Specifications Table 3-2. Ringback Specifications for PWRGOOD and TAP Signal Groups Signal Group Transition Maximum Ringback (with Input Diodes Present) Unit Figure PWRGOOD and TAP Low → High Vt+(max) to Vt-(max) V 3-3 1,2,3,4 PWRGOOD and TAP High → Low Vt-(min) to Vt+(min) V 3-4 1,2,3,4 Notes NOTES: 1. All signal integrity specifications are measured at the processor core (pads). 2. Unless otherwise noted, all specifications in this table apply to all processor frequencies. 3. All values specified by design characterization. 4. Please see Section 3.1.3 for maximum allowable overshoot. Figure 3-3. Low-to-High Receiver Ringback Tolerance for PWRGOOD and TAP Signals VTT Threshold Region to switch receiver to a logic 1. Vt+ (max) Vt+ (min) 0.5 * VTT Vt- (max) Allowable Ringback Vss 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 55 Front Side Bus Signal Quality Specifications Figure 3-4. High-to-Low Receiver Ringback Tolerance for PWRGOOD and TAP Signals VTT Allowable Ringback Vt+ (min) 0.5 * VTT Vt- (max) Vt- (min) Threshold Region to switch receiver to a logic 0. Vss 3.1.2 Overshoot/Undershoot Guidelines Overshoot (or undershoot) is the absolute value of the maximum voltage above or below VSS. The overshoot/undershoot specifications limit transitions beyond VTT or VSS due to the fast signal edge rates. The processor can be damaged by single and/or repeated overshoot or undershoot events on any input, output, or I/O buffer if the charge is large enough (i.e., if the over/undershoot is great enough). Determining the impact of an overshoot/undershoot condition requires knowledge of the magnitude, the pulse duration, and the activity factor (AF). Permanent damage to the processor is the likely result of excessive overshoot/undershoot. When performing simulations to determine the impact of overshoot and undershoot effects, ESD diodes must be properly characterized. ESD protection diodes do not act as voltage clamps and will not provide overshoot or undershoot protection. ESD diodes modeled within Intel signal integrity models do not clamp undershoot or overshoot and will yield correct simulation results. If other signal integrity models are being used to characterize the processor front side bus, care must be taken to ensure that ESD models do not clamp extreme voltage levels. Intel signal integrity models also contain I/O capacitance characterization. Therefore, removing the ESD diodes from a signal integrity model will impact results and may yield excessive overshoot/undershoot. 3.1.3 Overshoot/Undershoot Magnitude Magnitude describes the maximum potential difference between a signal and its voltage reference level. For the processor, both are referenced to VSS. It is important to note that overshoot and undershoot conditions are separate and their impact must be determined independently. The pulse magnitude, duration, and activity factor must all be used to determine if the overshoot/ undershoot pulse is within specifications. 56 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Front Side Bus Signal Quality Specifications 3.1.4 Overshoot/Undershoot Pulse Duration Pulse duration describes the total amount of time that an overshoot/undershoot event exceeds the overshoot/undershoot reference voltage. The total time could encompass several oscillations above the reference voltage. Multiple overshoot/undershoot pulses within a single overshoot/undershoot event may need to be measured to determine the total pulse duration. Note: 3.1.5 Oscillations below the reference voltage cannot be subtracted from the total overshoot/undershoot pulse duration. Activity Factor Activity Factor (AF) describes the frequency of overshoot (or undershoot) occurrence relative to a clock. Since the highest frequency of assertion of any common clock signal is every other clock, an AF = 1 indicates that the specific overshoot (or undershoot) waveform occurs every other clock cycle. Thus, an AF = 0.01 indicates that the specific overshoot (or undershoot) waveform occurs one time in every 200 clock cycles. For source synchronous signals (address, data, and associated strobes), the activity factor is in reference to the strobe edge. The highest frequency of assertion of any source synchronous signal is every active edge of its associated strobe. So, an AF = 1 indicates that the specific overshoot (or undershoot) waveform occurs every strobe cycle. The specifications provided in Table 3-3 through Table 3-6 show the maximum pulse duration allowed for a given overshoot/undershoot magnitude at a specific activity factor. Each table entry is independent of all others (meaning that the pulse duration reflects the existence of overshoot/ undershoot events of that magnitude ONLY). A platform with an overshoot/undershoot that just meets the pulse duration for a specific magnitude where the AF < 1 means that there can be no other overshoot/undershoot events, even of lesser magnitude (note that if AF = 1, then the event occurs at all times and no other events can occur). Note 1: Activity factor for common clock AGTL+ signals is referenced to BCLK[1:0] frequency. Note 2: Activity factor for source synchronous (2x) signals is referenced to ADSTB[1:0]#. Note 3: Activity factor for source synchronous (4x) signals is referenced to DSTBP[3:0]# and DSTBN[3:0]#. 3.1.6 Reading Overshoot/Undershoot Specification Tables The overshoot/undershoot specification for the processor is not a simple single value. Instead, many factors are needed to determine what the over/undershoot specification is. In addition to the magnitude of the overshoot, the following parameters must also be known: the width of the overshoot and the activity factor (AF). To determine the allowed overshoot for a particular overshoot event, the following must be done: 1. Determine the signal group that particular signal falls into. For AGTL+ signals operating in the 4x source synchronous domain, use Table 3-3. For AGTL+ signals operating in the 2x source synchronous domain, use Table 3-4. If the signal is an AGTL+ signal operating in the common clock domain, use Table 3-5. Finally, all other signals are referenced in Table 3-6. 2. Determine the magnitude of the overshoot or the undershoot (relative to VSS). 3. Determine the activity factor (how often does this overshoot occurs). 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 57 Front Side Bus Signal Quality Specifications 4. Next, from the appropriate specification table, determine the maximum pulse duration (in nanoseconds) allowed. 5. Compare the specified maximum pulse duration to the signal being measured. If the pulse duration measured is less than the pulse duration shown in the table, then the signal meets the specifications. Undershoot events must be analyzed separately from overshoot events as they are mutually exclusive. 3.1.7 Determining if a System Meets Over/Undershoot Specifications The overshoot/undershoot specifications listed in the following tables specify the allowable overshoot/undershoot for a single overshoot/undershoot event. Most systems will have multiple overshoot and/or undershoot events that each have their own set of parameters (duration, AF and magnitude). While each overshoot on its own may meet the overshoot specification, evaluate the cumulative overshoot of every cycle. A guideline to ensure a system passes the overshoot and undershoot specifications is shown below. Results from simulation may also be evaluated by utilizing the appropriate Processor Overshoot Checker Tool through the use of time-voltage data files. 1. If only one overshoot/undershoot event magnitude occurs, ensure it meets the over/undershoot specifications in the following tables OR 2. If multiple overshoots and/or multiple undershoots occur, measure the worst case pulse duration for each magnitude and compare the results against the AF = 1 specifications. If all of these worst case overshoot or undershoot events meet the specifications (measured time < specifications) in the table (where AF=1), then the system passes. Note: The following notes apply to Table 3-3 through Table 3-6. 1. Absolute Maximum Overshoot is measured referenced to VSS. Pulse Duration of overshoot is measured relative to VTT. 2. Absolute Maximum Undershoot and Pulse Duration of undershoot is measured relative to VSS. 3. Ringback below VTT cannot be subtracted from overshoots/undershoots. 4. Lesser undershoot does not allocate longer or larger overshoot. 5. OEM's are strongly encouraged to follow Intel layout recommendations. 6. All values specified by design characterization. 58 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Front Side Bus Signal Quality Specifications Table 3-3. Source Synchronous (667MHz) AGTL+ Signals Over/Undershoot Tolerance Absolute Maximum Overshoot (V) Absolute Maximum Undershoot (V) Pulse Duration (ns) AF = 1 Pulse Duration (ns) AF = 0.1 1.55 -0.35 0 0.2 1.50 -0.30 0 0.7 1.45 -0.25 0.1 1.5 1.40 -0.20 0.5 1.5 1.35 -0.15 1.3 1.5 1.30 -0.10 1.5 1.5 NOTES: 1. These specifications are measured at the processor pad. 2. Assumes a BCLK period of 6 ns. 3. AF is referenced to associated source synchronous strobes Table 3-4. Source Synchronous (333 MHz) AGTL+ Signals Over/Undershoot Tolerance Absolute Maximum Overshoot (V) Absolute Maximum Undershoot (V) Pulse Duration (ns) AF = 1 Pulse Duration (ns) AF = 0.1 1.55 -0.35 0 0.6 1.50 -0.30 0.1 1.9 1.45 -0.25 0.5 3.0 1.40 -0.20 1.3 3.0 1.35 -0.15 3.0 3.0 1.30 -0.10 3.0 3.0 NOTES: 1. These specifications are measured at the processor pad. 2. BCLK period is 6 ns. 3. AF is referenced to associated source synchronous strobe. Table 3-5. Common Clock (166 MHz) AGTL+ Signals Overshoot/Undershoot Tolerance Absolute Maximum Overshoot (V) Absolute Maximum Undershoot (V) Pulse Duration (ns) AF = 1 Pulse Duration (ns) AF = 0.1 1.55 -0.35 0.1 1.1 1.50 -0.30 0.3 3.2 1.45 -0.25 0.8 6.0 1.40 -0.20 2.3 6.0 1.35 -0.15 6.0 6.0 1.30 -0.10 6.0 6.0 NOTES: 1. These specifications are measured at the processor pad. 2. BCLK period is 6 ns. 3. AF is referenced to BCLK[1:0] 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 59 Front Side Bus Signal Quality Specifications Table 3-6. GTL+ Asynchronous, PWRGOOD, TAP Signals Over/Undershoot Tolerance Absolute Maximum Overshoot (V) Absolute Maximum Undershoot (V) Pulse Duration (ns) AF = 1 Pulse Duration (ns) AF = 0.1 1.55 -0.35 0.4 4.7 1.50 -0.30 1.3 12.8 1.45 -0.25 3.4 30.0 1.40 -0.20 9.3 30.0 1.35 -0.15 25.0 30.0 1.30 -0.10 30.0 30.0 NOTES: 1. These specifications are measured at the processor pad. 2. AF is referenced to a 33 MHz time domain. Figure 3-5. Maximum Acceptable Overshoot/Undershoot Waveform Time-dependent Overshoot VMAX VTT Maximum Absolute Overshoot GTLREF VOL VSS VMIN Maximum Absolute Undershoot § 60 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 4 Mechanical Specifications The 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache is packaged in a Flip-Chip Micro Pin Grid Array 4 (FC-mPGA4) package that interfaces with the motherboard via a mPGA604 socket. The package consists of a processor core mounted on a substrate pin-carrier. An integrated heat spreader (IHS) is attached to the package substrate and core and serves as the mating surface for processor component thermal solutions, such as a heatsink. Figure 4-1 shows a sketch of the processor package components and how they are assembled together. The package components shown in Figure 4-1 include the following: 1. Integrated Heat Spreader (IHS) 2. Processor die 3. FC-mPGA4 package 4. Pin-side capacitors 5. Package pin Figure 4-1. Processor Package Assembly Sketch 11 22 33 44 Note: 55 This drawing is not to scale and is for reference only. The mPGA604 socket is not shown. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 61 Mechanical Specifications 4.1 Package Mechanical Drawing The package mechanical drawings are shown in Figure 4-2 and Figure 4-3. The drawings include dimensions necessary to design a thermal solution for the processor. These dimensions include: 1. Package reference with tolerances (total height, length, width, etc.) 2. IHS parallelism and tilt 3. Pin dimensions 4. Top-side and back-side component keep-out dimensions 5. Reference datums All drawing dimensions are in mm[in]. 62 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Mechanical Specifications Figure 4-2. Processor Package Drawing (Sheet 1 of 2) 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 63 Mechanical Specifications Figure 4-3. 64 Processor Package Drawing (Sheet 2 of 2) 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Mechanical Specifications 4.2 Processor Component Keep-Out Zones The processor may contain components on the substrate that define component keep-out zone requirements. A thermal and mechanical solution design must not intrude into the required keepout zones. Decoupling capacitors are typically mounted to either the topside or pin-side of the package substrate. See Figure 4-2 and Figure 4-3 for keepout zones. 4.3 Package Loading Specifications Table 4-1 provides dynamic and static load specifications for the processor package. These mechanical load limits should not be exceeded during heatsink assembly, shipping conditions, or standard use condition. Also, any mechanical system or component testing should not exceed the maximum limits. The processor package substrate should not be used as a mechanical reference or load-bearing surface for thermal and mechanical solutions. The minimum loading specification must be maintained by any thermal and mechanical solution. Table 4-1. Processor Loading Specifications Parameter Static Compressive Load Dynamic Compressive Load Transient Minimum Maximum Unit Notes 44 10 222 50 N lbf 1, 2, 3, 4 44 10 288 65 N lbf 1, 2, 3, 5 222 N + 0.45 kg * 100 G 50 lbf (static) + 1 lbm * 100 G N lbf 1, 3, 4, 6, 7 288 N + 0.45 kg * 100 G 65 lbf (static) + 1 lbm * 100 G N lbf 1, 3, 5, 6, 7 445 100 N lbf 1, 3, 8 NOTES: 1. These specifications apply to uniform compressive loading in a direction perpendicular to the IHS top surface. 2. This is the minimum and maximum static force that can be applied by the heatsink and retention solution to maintain the heatsink and processor interface. 3. These parameters are based on limited testing for design characterization. Loading limits are for the package only and do not include the limits of the processor socket. 4. This specification applies for thermal retention solutions that allow baseboard deflection. 5. This specification applies either for thermal retention solutions that prevent baseboard deflection or for the Intel enabled reference solution. 6. Dynamic loading is defined as an 11 ms duration average load superimposed on the static load requirement. 7. Experimentally validated test condition used a heatsink mass of 1 lbm (~0.45 kg) with 100 G acceleration measured at heatsink mass. The dynamic portion of this specification in the product application can have flexibility in specific values, but the ultimate product of mass times acceleration should not exceed this validated dynamic load (1 lbm x 100 G = 100 lb). 8. Transient loading is defined as a 2 second duration peak load superimposed on the static load requirement, representative of loads experienced by the package during heatsink installation. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 65 Mechanical Specifications 4.4 Package Handling Guidelines Table 4-2 includes a list of guidelines on package handling in terms of recommended maximum loading on the processor IHS relative to a fixed substrate. These package handling loads may be experienced during heatsink removal. Table 4-2. Package Handling Guidelines Parameter Maximum Recommended Notes Shear 356 N [80 lbf] 1, 4, 5 Tensile 156 N [35 lbf] 2, 4, 5 Torque 8 N-m [70 lbf-in] 3, 4, 5 NOTES: 1. A shear load is defined as a load applied to the IHS in a direction parallel to the IHS top surface. 2. A tensile load is defined as a pulling load applied to the IHS in the direction normal to the IHS surface. 3. A torque load is defined as a twisting load applied to the IHS in an axis of rotation normal to the IHS top surface. 4. These guidelines are based on limited testing for design characterization and incidental applications (one time only). 5. Handling guidelines are for the package only and do not include the limits of the processor socket. 4.5 Package Insertion Specifications The processor can be inserted into and removed from a mPGA604 socket 15 times. The socket should meet the mPGA604 requirements detailed in the mPGA604 Socket Design Guidelines. 4.6 Processor Mass Specifications The typical mass of the processor is 34 g [1.20 oz]. This mass [weight] includes all the components that are included in the package. 4.7 Processor Materials Table 4-3 lists some of the package components and associated materials. Table 4-3. Processor Materials 66 Component Material Integrated Heat Spreader (IHS) Nickel Plated Copper Substrate Fiber-Reinforced Resin Substrate Pins Gold Plated Copper 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Mechanical Specifications 4.8 Processor Markings Figure 4-4 shows the topside markings and Figure 4-5 shows the bottom-side markings on the processor. These diagrams are to aid in the identification of the processor. Figure 4-4. Processor Topside Markings Process name i(m) ©’05 2D Matrix Includes ATPO and Serial Number (front end mark) Pin 1 Indicator NOTES: 1. All characters will be in upper case. 2. Drawing is not to scale. Figure 4-5. Processor Bottom-Side Markings Pin 1 Indicator Speed / Cache / Bus / Voltage Pin Field 3333MP / 8ML3 / 667 / 1.3875 V SL 6NY COSTA RICA C0096109-0021 Cavity with Components S-Spec Country of Assy FPO – Serial # (13 Characters) Text Line1 Text Line2 Text Line3 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 67 Mechanical Specifications 4.9 Processor Pin-Out Coordinates Figure 4-6 shows the top view of the processor pin coordinates. The coordinates are referred to throughout the document to identify processor pins. Figure 4-6. Processor Pin-Out Coordinates, Top View COMMON CLOCK 3 5 7 9 11 13 A B C D E F G H J K L M N P R T U V W Y AA AB AC AD AE 15 17 19 21 23 Async / JTAG 25 27 29 31 A B C D E F G H J K L M N P R T U V W Y AA AB AC Processor Top View Vcc/Vss Vcc/Vcache/Vss 1 COMMON CLOCK ADDRESS AD AE 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 DATA CLOCKS = Signal = VCC = Ground = VTT = Reserved/No Connect = VCache § 68 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 5 Pin Listing 5.1 Processor Pin Assignments Section 2.6 contains the front side bus signal groups for the processor (see Table 2-5). This section provides a sorted pin list in Table 5-1 and Table 5-2. Table 5-1 is a listing of all processor pins ordered alphabetically by pin name. Table 5-2 is a listing of all processor pins ordered by pin number. 5.1.1 Pin Listing by Pin Name Table 5-1. Pin Listing by Pin Name (cont’d) Table 5-1. Pin Listing by Pin Name Pin No. Signal Buffer Type Direction A28# E13 Source Sync Input/Output Input/Output A29# D12 Source Sync Input/Output Input/Output A30# C11 Source Sync Input/Output B7 Source Sync Input/Output Pin No. Signal Buffer Type Direction A3# A22 Source Sync Input/Output A4# A20 Source Sync A5# B18 Source Sync Pin Name Pin Name A6# C18 Source Sync Input/Output A31# A7# A19 Source Sync Input/Output A32# A6 Source Sync Input/Output A8# C17 Source Sync Input/Output A33# A7 Source Sync Input/Output A9# D17 Source Sync Input/Output A34# C9 Source Sync Input/Output A35# C8 Source Sync Source Sync A36# F16 Source Sync Source Sync A37# F22 Source Sync Source Sync A38# B6 Source Sync Source Sync A39# C16 Source Sync Source Sync Input/Output A20M# F27 Async GTL+ Input Source Sync Input/Output ADS# D19 Common Clk Input/Output A10 Source Sync Input/Output ADSTB0# F17 Source Sync Input/Output B10 Source Sync Input/Output ADSTB1# F14 Source Sync Input/Output E10 Common Clk Input/Output A10# A13 Source Sync Input/Output A11# B16 Source Sync Input/Output A12# B14 Source Sync Input/Output A13# B13 Source Sync Input/Output A14# A12 Source Sync Input/Output A15# C15 Source Sync Input/Output A16# C14 Source Sync Input/Output A17# D16 Source Sync Input/Output A18# D15 Source Sync A19# F15 A20# A21# A22# B11 Source Sync Input/Output AP0# A23# C12 Source Sync Input/Output AP1# D9 Common Clk Input/Output A24# E14 Source Sync Input/Output BCLK0 Y4 FSB Clk Input A25# D13 Source Sync Input/Output BCLK1 W5 FSB Clk Input F11 Common Clk Input/Output F20 Common Clk Input/Output A26# A9 Source Sync Input/Output BINIT# A27# B8 Source Sync Input/Output BNR# 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 69 Pin Listing Table 5-1. Pin Listing by Pin Name (cont’d) Pin No. Signal Buffer Type BOOT_SELECT G7 Power/Other Input BPM0# F6 Common Clk BPM1# F8 Common Clk BPM2# E7 BPM3# Table 5-1. Pin Listing by Pin Name (cont’d) Pin No. Signal Buffer Type Direction D18# AD25 Source Sync Input/Output Input/Output D19# AE25 Source Sync Input/Output Input/Output D20# AC24 Source Sync Input/Output Common Clk Input/Output D21# AD24 Source Sync Input/Output F5 Common Clk Input/Output D22# AE23 Source Sync Input/Output BPM4# E8 Common Clk Input/Output D23# AC23 Source Sync Input/Output BPM5# E4 Common Clk Input/Output D24# AA18 Source Sync Input/Output BPRI# D23 Common Clk Input D25# AC20 Source Sync Input/Output BR0# D20 Common Clk Input/Output D26# AC21 Source Sync Input/Output BR1# F12 Common Clk Input D27# AE22 Source Sync Input/Output BR2#1 E11 Common Clk Input D28# AE20 Source Sync Input/Output BR3#1 D10 Common Clk Input D29# AD21 Source Sync Input/Output BSEL0 AA3 Power/Other Output D30# AD19 Source Sync Input/Output BSEL1 AB3 Power/Other Output D31# AB17 Source Sync Input/Output COMP0 AD16 Power/Other Input D32# AB16 Source Sync Input/Output CVID0 E2 Power/Other Output D33# AA16 Source Sync Input/Output CVID1 D1 Power/Other Output D34# AC17 Source Sync Input/Output CVID2 C2 Power/Other Output D35# AE13 Source Sync Input/Output CVID3 A2 Power/Other Output D36# AD18 Source Sync Input/Output D0# Y26 Source Sync Input/Output D37# AB15 Source Sync Input/Output D1# AA27 Source Sync Input/Output D38# AD13 Source Sync Input/Output D2# Y24 Source Sync Input/Output D39# AD14 Source Sync Input/Output D3# AA25 Source Sync Input/Output D40# AD11 Source Sync Input/Output D4# AD27 Source Sync Input/Output D41# AC12 Source Sync Input/Output D5# Y23 Source Sync Input/Output D42# AE10 Source Sync Input/Output D6# AA24 Source Sync Input/Output D43# AC11 Source Sync Input/Output D7# AB26 Source Sync Input/Output D44# AE9 Source Sync Input/Output D8# AB25 Source Sync Input/Output D45# AD10 Source Sync Input/Output D9# AB23 Source Sync Input/Output D46# AD8 Source Sync Input/Output Pin Name Direction Pin Name D10# AA22 Source Sync Input/Output D47# AC9 Source Sync Input/Output D11# AA21 Source Sync Input/Output D48# AA13 Source Sync Input/Output D12# AB20 Source Sync Input/Output D49# AA14 Source Sync Input/Output D13# AB22 Source Sync Input/Output D50# AC14 Source Sync Input/Output D14# AB19 Source Sync Input/Output D51# AB12 Source Sync Input/Output D15# AA19 Source Sync Input/Output D52# AB13 Source Sync Input/Output D16# AE26 Source Sync Input/Output D53# AA11 Source Sync Input/Output D17# AC26 Source Sync Input/Output D54# AA10 Source Sync Input/Output 70 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Pin Listing Table 5-1. Pin Listing by Pin Name (cont’d) Table 5-1. Pin Listing by Pin Name (cont’d) Pin No. Signal Buffer Type Direction D55# AB10 Source Sync Input/Output FORCEPR# D56# AC8 Source Sync Input/Output GTLREF0 W23 Power/Other Input D57# AD7 Source Sync Input/Output GTLREF1 W9 Power/Other Input D58# AE7 Source Sync Input/Output GTLREF2 F23 Power/Other Input D59# AC6 Source Sync Input/Output GTLREF3 F9 Power/Other Input D60# AC5 Source Sync Input/Output HIT# E22 Common Clk Input/Output D61# AA8 Source Sync Input/Output HITM# A23 Common Clk Input/Output D62# Y9 Source Sync Input/Output ID0# A26 Common Clk Input D63# AB6 Source Sync Input/Output ID1# B26 Common Clk Input DBI0# AC27 Source Sync Input/Output ID2# D25 Common Clk Input DBI1# AD22 Source Sync Input/Output ID3# D27 Common Clk Input DBI2# AE12 Source Sync Input/Output ID4# C28 Common Clk Input DBI3# AB9 Source Sync Input/Output ID5# B29 Common Clk Input DBSY# F18 Common Clk Input/Output ID6# B30 Common Clk Input DEFER# C23 Common Clk Input ID7# A30 Common Clk Input DEP0# AD31 Source Sync Input/Output IDS# A28 Common Clk Input DEP1# AD30 Source Sync Input/Output IERR# E5 Async GTL+ Output DEP2# AE16 Source Sync Input/Output IGNNE# C26 Async GTL+ Input DEP3# AE15 Source Sync Input/Output INIT# D6 Async GTL+ Input DEP4# AE8 Source Sync Input/Output LINT0/INTR B24 Async GTL+ Input DEP5# AD6 Source Sync Input/Output LINT1/NMI G23 Async GTL+ Input DEP6# AC4 Source Sync Input/Output LOCK# A17 Common Clk Input/Output DEP7# AA4 Source Sync Input/Output MCERR# D7 Common Clk Input/Output DP0# AC18 Common Clk Input/Output ODTEN B5 Power/Other Input DP1# AE19 Common Clk Input/Output OOD# D29 Common Clk Input DP2# AC15 Common Clk Input/Output PROCHOT# B25 Async GTL+ Output DP3# AE17 Common Clk Input/Output PWRGOOD AB7 Async GTL+ Input DRDY# E18 Common Clk Input/Output REQ0# B19 Source Sync Input/Output DSTBN0# Y21 Source Sync Input/Output REQ1# B21 Source Sync Input/Output DSTBN1# Y18 Source Sync Input/Output REQ2# C21 Source Sync Input/Output DSTBN2# Y15 Source Sync Input/Output REQ3# C20 Source Sync Input/Output DSTBN3# Y12 Source Sync Input/Output REQ4# B22 Source Sync Input/Output DSTBP0# Y20 Source Sync Input/Output Reserved A31 DSTBP1# Y17 Source Sync Input/Output Reserved C1 DSTBP2# Y14 Source Sync Input/Output Reserved E16 DSTBP3# Y11 Source Sync Input/Output Reserved W3 FERR#/PBE# E27 Async GTL+ Output Reserved Y27 Pin Name Pin Name 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Pin No. Signal Buffer Type A15 Power/Other Direction Input 71 Pin Listing Table 5-1. Pin Listing by Pin Name (cont’d) Pin Name Pin No. Signal Buffer Type Table 5-1. Pin Listing by Pin Name (cont’d) Direction Pin Name Pin No. Signal Buffer Type E19 Common Clk Input Input Reserved Y28 TRDY# Reserved AC1 TRST# F24 TAP Reserved AE4 VCACHE B4 Power/Other Reserved AE30 VCACHE H1 Power/Other RESET# Y8 Common Clk Input VCACHE H3 Power/Other RS0# E21 Common Clk Input VCACHE H5 Power/Other RS1# D22 Common Clk Input VCACHE H7 Power/Other RS2# F21 Common Clk Input VCACHE H9 Power/Other RSP# C6 Common Clk Input VCACHE K1 Power/Other SKTOCC# A3 Power/Other Output VCACHE K3 Power/Other SLEW_CTRL AC30 Power/Other Input VCACHE K5 Power/Other SM_ALERT# AD28 SMBus Output VCACHE K7 Power/Other SM_CLK AC28 SMBus Input VCACHE K9 Power/Other SM_DAT AC29 SMBus Input/Output VCACHE M1 Power/Other SM_EP_A0 AA29 SMBus Input VCACHE M3 Power/Other SM_EP_A1 AB29 SMBus Input VCACHE M5 Power/Other SM_EP_A2 AB28 SMBus Input VCACHE M7 Power/Other SM_TS1_A0 AA28 SMBus Input VCACHE M9 Power/Other SM_TS1_A1 Y29 SMBus Input VCACHE N1 Power/Other SM_VCC AE28 Power/Other VCACHE N3 Power/Other SM_VCC AE29 Power/Other VCACHE N5 Power/Other SM_WP AD29 SMBus Input VCACHE N7 Power/Other SMI# C27 Async GTL+ Input VCACHE N9 Power/Other STPCLK# D4 Async GTL+ Input VCACHE R1 Power/Other TCK E24 TAP Input VCACHE R3 Power/Other TDI C24 TAP Input VCACHE R5 Power/Other TDO E25 TAP Output VCACHE R7 Power/Other TEST_BUS A16 Power/Other Input VCACHE R9 Power/Other TESTHI0 W6 Power/Other Input VCACHE U1 Power/Other TESTHI1 W7 Power/Other Input VCACHE U3 Power/Other TESTHI2 W8 Power/Other Input VCACHE U5 Power/Other TESTHI3 Y6 Power/Other Input VCACHE U7 Power/Other TESTHI4 AA7 Power/Other Input VCACHE U9 Power/Other TESTHI5 AD5 Power/Other Input VCC A8 Power/Other TESTHI6 AE5 Power/Other Input VCC A14 Power/Other THERMTRIP# F26 Async GTL+ Output VCC A18 Power/Other TMS A25 TAP Input VCC A24 Power/Other 72 Direction 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Pin Listing Table 5-1. Pin Listing by Pin Name (cont’d) Pin No. Signal Buffer Type VCC B20 Power/Other VCC C4 VCC C22 VCC C30 VCC Table 5-1. Pin Listing by Pin Name (cont’d) Pin No. Signal Buffer Type VCC J28 Power/Other Power/Other VCC J30 Power/Other Power/Other VCC K23 Power/Other Power/Other VCC K25 Power/Other D8 Power/Other VCC K27 Power/Other VCC D14 Power/Other VCC K29 Power/Other VCC D18 Power/Other VCC K31 Power/Other VCC D24 Power/Other VCC L2 Power/Other VCC D31 Power/Other VCC L4 Power/Other VCC E6 Power/Other VCC L6 Power/Other VCC E20 Power/Other VCC L8 Power/Other VCC E26 Power/Other VCC L24 Power/Other VCC E28 Power/Other VCC L26 Power/Other VCC E30 Power/Other VCC L28 Power/Other VCC F1 Power/Other VCC L30 Power/Other VCC F4 Power/Other VCC M23 Power/Other VCC F29 Power/Other VCC M25 Power/Other VCC F31 Power/Other VCC M27 Power/Other VCC G2 Power/Other VCC M29 Power/Other VCC G4 Power/Other VCC M31 Power/Other VCC G6 Power/Other VCC N23 Power/Other VCC G8 Power/Other VCC N25 Power/Other VCC G24 Power/Other VCC N27 Power/Other VCC G26 Power/Other VCC N29 Power/Other VCC G28 Power/Other VCC N31 Power/Other VCC G30 Power/Other VCC P2 Power/Other VCC H23 Power/Other VCC P4 Power/Other VCC H25 Power/Other VCC P6 Power/Other VCC H27 Power/Other VCC P8 Power/Other VCC H29 Power/Other VCC P24 Power/Other VCC H31 Power/Other VCC P26 Power/Other VCC J2 Power/Other VCC P28 Power/Other VCC J4 Power/Other VCC P30 Power/Other VCC J6 Power/Other VCC R23 Power/Other VCC J8 Power/Other VCC R25 Power/Other VCC J24 Power/Other VCC R27 Power/Other VCC J26 Power/Other VCC R29 Power/Other Pin Name Direction Pin Name 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Direction 73 Pin Listing Table 5-1. Pin Listing by Pin Name (cont’d) Pin No. Signal Buffer Type VCC R31 Power/Other VCC T2 VCC T4 VCC Table 5-1. Pin Listing by Pin Name (cont’d) Pin No. Signal Buffer Type VCC AB8 Power/Other Power/Other VCC AB14 Power/Other Power/Other VCC AB18 Power/Other T6 Power/Other VCC AB24 Power/Other VCC T8 Power/Other VCC AB30 Power/Other VCC T24 Power/Other VCC AC3 Power/Other VCC T26 Power/Other VCC AC16 Power/Other VCC T28 Power/Other VCC AC22 Power/Other VCC T30 Power/Other VCC AC31 Power/Other VCC U23 Power/Other VCC AD2 Power/Other VCC U25 Power/Other VCC AD20 Power/Other VCC U27 Power/Other VCC AD26 Power/Other VCC U29 Power/Other VCC AE14 Power/Other VCC U31 Power/Other VCC AE18 Power/Other VCC V2 Power/Other VCC AE24 Power/Other VCC V4 Power/Other VCCA AB4 Power/Other Input VCC V6 Power/Other VCCA_CACHE AE3 Power/Other Input VCC V8 Power/Other VCC_CACHE_SENSE B31 Power/Other Output VCC V24 Power/Other VCCIOPLL AD4 Power/Other Input VCC V26 Power/Other VCCPLL AD1 Power/Other Input VCC V28 Power/Other VCCSENSE B27 Power/Other Output VCC V30 Power/Other VID0 F3 Power/Other Output VCC W1 Power/Other VID1 E3 Power/Other Output VCC W25 Power/Other VID2 D3 Power/Other Output VCC W27 Power/Other VID3 C3 Power/Other Output VCC W29 Power/Other VID4 B3 Power/Other Output VCC W31 Power/Other VID5 A1 Power/Other Output VCC Y2 Power/Other VIDPWRGD B1 Power/Other Input VCC Y16 Power/Other VSS A5 Power/Other VCC Y22 Power/Other VSS A11 Power/Other VCC Y30 Power/Other VSS A21 Power/Other VCC AA1 Power/Other VSS A27 Power/Other VCC AA6 Power/Other VSS A29 Power/Other VCC AA20 Power/Other VSS B2 Power/Other VCC AA26 Power/Other VSS B9 Power/Other VCC AA31 Power/Other VSS B15 Power/Other VCC AB2 Power/Other VSS B17 Power/Other Pin Name 74 Direction Pin Name Direction 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Pin Listing Table 5-1. Pin Listing by Pin Name (cont’d) Pin No. Signal Buffer Type VSS B23 Power/Other VSS B28 VSS C7 VSS Table 5-1. Pin Listing by Pin Name (cont’d) Pin No. Signal Buffer Type VSS H8 Power/Other Power/Other VSS H24 Power/Other Power/Other VSS H26 Power/Other C13 Power/Other VSS H28 Power/Other VSS C19 Power/Other VSS H30 Power/Other VSS C25 Power/Other VSS J1 Power/Other VSS C29 Power/Other VSS J3 Power/Other VSS D2 Power/Other VSS J5 Power/Other VSS D5 Power/Other VSS J7 Power/Other VSS D11 Power/Other VSS J9 Power/Other VSS D21 Power/Other VSS J23 Power/Other VSS D28 Power/Other VSS J25 Power/Other VSS D30 Power/Other VSS J27 Power/Other VSS E9 Power/Other VSS J29 Power/Other VSS E15 Power/Other VSS J31 Power/Other VSS E17 Power/Other VSS K2 Power/Other VSS E23 Power/Other VSS K4 Power/Other VSS E29 Power/Other VSS K6 Power/Other VSS E31 Power/Other VSS K8 Power/Other VSS F2 Power/Other VSS K24 Power/Other VSS F7 Power/Other VSS K26 Power/Other VSS F13 Power/Other VSS K28 Power/Other VSS F19 Power/Other VSS K30 Power/Other VSS F25 Power/Other VSS L1 Power/Other VSS F28 Power/Other VSS L3 Power/Other VSS F30 Power/Other VSS L5 Power/Other VSS G1 Power/Other VSS L7 Power/Other VSS G3 Power/Other VSS L9 Power/Other VSS G5 Power/Other VSS L23 Power/Other VSS G9 Power/Other VSS L25 Power/Other VSS G25 Power/Other VSS L27 Power/Other VSS G27 Power/Other VSS L29 Power/Other VSS G29 Power/Other VSS L31 Power/Other VSS G31 Power/Other VSS M2 Power/Other VSS H2 Power/Other VSS M4 Power/Other VSS H4 Power/Other VSS M6 Power/Other VSS H6 Power/Other VSS M8 Power/Other Pin Name Direction Pin Name 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Direction 75 Pin Listing Table 5-1. Pin Listing by Pin Name (cont’d) Pin No. Signal Buffer Type VSS M24 Power/Other VSS M26 VSS M28 VSS Table 5-1. Pin Listing by Pin Name (cont’d) Pin No. Signal Buffer Type VSS T27 Power/Other Power/Other VSS T29 Power/Other Power/Other VSS T31 Power/Other M30 Power/Other VSS U2 Power/Other VSS N2 Power/Other VSS U4 Power/Other VSS N4 Power/Other VSS U6 Power/Other VSS N6 Power/Other VSS U8 Power/Other VSS N8 Power/Other VSS U24 Power/Other VSS N24 Power/Other VSS U26 Power/Other VSS N26 Power/Other VSS U28 Power/Other VSS N28 Power/Other VSS U30 Power/Other VSS N30 Power/Other VSS V1 Power/Other VSS P1 Power/Other VSS V3 Power/Other VSS P3 Power/Other VSS V5 Power/Other VSS P5 Power/Other VSS V7 Power/Other VSS P7 Power/Other VSS V9 Power/Other VSS P9 Power/Other VSS V23 Power/Other VSS P23 Power/Other VSS V25 Power/Other VSS P25 Power/Other VSS V27 Power/Other VSS P27 Power/Other VSS V29 Power/Other VSS P29 Power/Other VSS V31 Power/Other VSS P31 Power/Other VSS W2 Power/Other VSS R2 Power/Other VSS W4 Power/Other VSS R4 Power/Other VSS W24 Power/Other VSS R6 Power/Other VSS W26 Power/Other VSS R8 Power/Other VSS W28 Power/Other VSS R24 Power/Other VSS W30 Power/Other VSS R26 Power/Other VSS Y1 Power/Other VSS R28 Power/Other VSS Y3 Power/Other VSS R30 Power/Other VSS Y5 Power/Other VSS T1 Power/Other VSS Y7 Power/Other VSS T3 Power/Other VSS Y13 Power/Other VSS T5 Power/Other VSS Y19 Power/Other VSS T7 Power/Other VSS Y25 Power/Other VSS T9 Power/Other VSS Y31 Power/Other VSS T23 Power/Other VSS AA2 Power/Other VSS T25 Power/Other VSS AA9 Power/Other Pin Name 76 Direction Pin Name Direction 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Pin Listing Table 5-1. Pin Listing by Pin Name (cont’d) Pin No. Signal Buffer Type VSS AA15 Power/Other VSS AA17 VSS AA23 VSS Table 5-1. Pin Listing by Pin Name (cont’d) Pin No. Signal Buffer Type VSSA AA5 Power/Other Power/Other VSSA_CACHE AE2 Power/Other Input Power/Other VSS_CACHE_SENSE C31 Power/Other Output AA30 Power/Other VSSSENSE D26 Power/Other Output VSS AB1 Power/Other VTT A4 Power/Other VSS AB5 Power/Other VTT B12 Power/Other VSS AB11 Power/Other VTT C5 Power/Other VSS AB21 Power/Other VTT C10 Power/Other VSS AB27 Power/Other VTT E12 Power/Other VSS AB31 Power/Other VTT F10 Power/Other VSS AC2 Power/Other VTT Y10 Power/Other VSS AC7 Power/Other VTT AA12 Power/Other VSS AC13 Power/Other VTT AC10 Power/Other VSS AC19 Power/Other VTT AD12 Power/Other VSS AC25 Power/Other VTTEN E1 Power/Other VSS AD3 Power/Other VSS AD9 Power/Other VSS AD15 Power/Other VSS AD17 Power/Other VSS AD23 Power/Other VSS AE6 Power/Other VSS AE11 Power/Other VSS AE21 Power/Other VSS AE27 Power/Other Pin Name Direction Pin Name Direction Input Output NOTES: 1. In systems utilizing the 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache, the system designer must pull-up these signals to the processor VTT. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 77 Pin Listing 5.1.2 Pin Listing by Pin Number Table 5-2. Pin Listing by Pin Number (cont’d) Table 5-2. Pin Listing by Pin Number Pin No. Pin Name Signal Buffer Type Direction Pin No. Pin Name Signal Buffer Type Direction A1 VID5 Power/Other Output B2 VSS Power/Other A2 CVID3 Power/Other Output B3 VID4 Power/Other A3 SKTOCC# Power/Other Output B4 VCACHE Power/Other A4 VTT Power/Other B5 ODTEN Power/Other Input B6 A38# Source Sync Input/Output Output A5 VSS Power/Other A6 A32# Source Sync Input/Output B7 A31# Source Sync Input/Output A7 A33# Source Sync Input/Output B8 A27# Source Sync Input/Output A8 VCC Power/Other B9 VSS Power/Other A21# Source Sync Input/Output Input/Output A9 A26# Source Sync Input/Output B10 A10 A20# Source Sync Input/Output B11 A22# Source Sync A11 VSS Power/Other B12 VTT Power/Other A12 A14# Source Sync Input/Output B13 A13# Source Sync Input/Output Input/Output B14 A12# Source Sync Input/Output B15 VSS Power/Other A11# Source Sync A13 A10# Source Sync A14 VCC Power/Other A15 FORCEPR# Power/Other Input B16 A16 TEST_BUS Power/Other Input B17 VSS Power/Other B18 A5# Source Sync Input/Output B19 REQ0# Common Clk Input/Output Input/Output A17 LOCK# Common Clk A18 VCC Power/Other A19 A7# Source Sync Input/Output B20 VCC Power/Other A20 A4# Source Sync Input/Output B21 REQ1# Common Clk Input/Output A21 VSS Power/Other A22 A3# Source Sync Input/Output B22 REQ4# Common Clk Input/Output A23 HITM# Common Clk Input/Output B23 VSS Power/Other B24 LINT0/INTR Async GTL+ Input A24 VCC Power/Other B25 PROCHOT# Power/Other Output A25 TMS TAP Input ID1# Input Common Clk Input ID0# Common Clk B26 A26 B27 Output VSS Power/Other VCCSENSE Power/Other A27 B28 Power/Other IDS# Common Clk VSS A28 B29 ID5# Common Clk Input A29 VSS Power/Other ID6# Common Clk Input ID7# Common Clk B30 A30 B31 Reserved VCC_CACHE_SENSE Power/Other A31 C1 Reserved B1 VIDPWRGD C2 CVID2 78 Power/Other Input/Output Input Input Input Power/Other Output 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Pin Listing Table 5-2. Pin Listing by Pin Number (cont’d) Pin No. Pin Name Signal Buffer Type Direction C3 VID3 Power/Other Output C4 VCC Power/Other C5 VTT Power/Other C6 RSP# Common Clk C7 VSS Power/Other C8 A35# Source Sync Input/Output C9 A34# Source Sync Input/Output Input Table 5-2. Pin Listing by Pin Number (cont’d) Pin No. Pin Name Signal Buffer Type Direction D7 MCERR# Common Clk D8 VCC Power/Other D9 AP1# Common Clk Input/Output D10 BR3#1 Common Clk Input D11 VSS Power/Other D12 A29# Source Sync Input/Output Input/Output Input/Output C10 VTT Power/Other C11 A30# Source Sync Input/Output D13 A25# Source Sync C12 A23# Source Sync Input/Output D14 VCC Power/Other C13 VSS Power/Other D15 A18# Source Sync Input/Output A17# Source Sync Input/Output Input/Output C14 A16# Source Sync Input/Output D16 C15 A15# Source Sync Input/Output D17 A9# Source Sync C16 A39# Source Sync Input/Output D18 VCC Power/Other C17 A8# Source Sync Input/Output D19 ADS# Common Clk Input/Output C18 A6# Source Sync Input/Output C19 VSS Power/Other D20 BR0# Common Clk Input/Output C20 REQ3# Common Clk D21 VSS Power/Other D22 RS1# REQ2# Common Clk Input C21 Common Clk C22 VCC Power/Other D23 BPRI# Common Clk Input C23 DEFER# Common Clk Input D24 VCC Power/Other C24 TDI TAP Input D25 ID2# Common Clk Input C25 VSS Power/Other Input D26 VSSSENSE Power/Other Output C26 IGNNE# Async GTL+ Input D27 ID3# Input C27 SMI# Async GTL+ Input Common Clk VSS Power/Other ID4# Common Clk D28 C28 Input D29 OOD# C29 VSS Power/Other Common Clk C30 VCC Power/Other D30 VSS Power/Other C31 VSS_CACHE_SENSE Power/Other D31 VCC Power/Other D1 CVID1 Power/Other E1 VTTEN Power/Other Output D2 VSS Power/Other E2 CVID0 Power/Other Output D3 VID2 Power/Other E3 VID1 Power/Other Output E4 BPM5# Common Clk Input/Output E5 IERR# Common Clk Output D4 STPCLK# Async GTL+ D5 VSS Power/Other D6 INIT# Async GTL+ Input/Output Input/Output Output Output Input Input 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Input 79 Pin Listing Table 5-2. Pin Listing by Pin Number (cont’d) Pin No. E6 Pin Name Signal Buffer Type Direction Table 5-2. Pin Listing by Pin Number (cont’d) Pin No. Pin Name Signal Buffer Type Direction VCC Power/Other E7 BPM2# Common Clk Input/Output E8 BPM4# Common Clk Input/Output E9 VSS Power/Other F9 E10 AP0# Common Clk Input/Output F10 BINIT# Input Common Clk Input/Output BR2#1 Common Clk F11 E11 E12 VTT Power/Other F12 BR1# Common Clk Input E13 A28# Source Sync Input/Output F13 VSS Power/Other E14 A24# Source Sync Input/Output F14 ADSTB1# Source Sync Input/Output E15 VSS Power/Other F15 A19# Source Sync Input/Output E16 Reserved F16 A36# Source Sync Input/Output E17 VSS Power/Other F17 ADSTB0# Source Sync Input/Output E18 DRDY# Common Clk Input/Output F18 DBSY# Common Clk Input/Output E19 TRDY# Common Clk Input F19 VSS Power/Other E20 VCC Power/Other F20 BNR# Common Clk Input/Output E21 RS0# Common Clk Input F21 RS2# Common Clk Input E22 HIT# Common Clk Input/Output F22 A37# Source Sync Input/Output F23 GTLREF2 Power/Other Input F24 TRST# TAP Input F25 VSS Power/Other F26 THERMTRIP# Async GTL+ Output F27 A20M# Async GTL+ Input F28 VSS Power/Other F29 VCC Power/Other F30 VSS Power/Other F31 VCC Power/Other G1 VSS Power/Other G2 VCC Power/Other G3 VSS Power/Other G4 VCC Power/Other G5 VSS Power/Other G6 VCC Power/Other G7 BOOT_SELECT Power/Other E23 VSS Power/Other E24 TCK TAP Input E25 TDO TAP Output E26 VCC Power/Other E27 FERR#/PBE# Async GTL+ E28 VCC Power/Other E29 VSS Power/Other E30 VCC Power/Other E31 VSS Power/Other F1 VCC Power/Other F2 VSS Power/Other F3 VID0 Power/Other F4 VCC Power/Other F5 BPM3# Common Clk 80 Output Output Input/Output F6 BPM0# Common Clk F7 VSS Power/Other F8 BPM1# Common Clk Input/Output GTLREF3 Power/Other Input VTT Power/Other Input/Output Input 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Pin Listing Table 5-2. Pin Listing by Pin Number (cont’d) Pin No. Pin Name Signal Buffer Type G8 VCC Power/Other G9 VSS Power/Other G23 LINT1/NMI Async GTL+ G24 VCC G25 Direction Table 5-2. Pin Listing by Pin Number (cont’d) Pin No. Pin Name Signal Buffer Type J9 VSS Power/Other J23 VSS Power/Other J24 VCC Power/Other Power/Other J25 VSS Power/Other VSS Power/Other J26 VCC Power/Other G26 VCC Power/Other J27 VSS Power/Other G27 VSS Power/Other J28 VCC Power/Other G28 VCC Power/Other J29 VSS Power/Other G29 VSS Power/Other J30 VCC Power/Other G30 VCC Power/Other J31 VSS Power/Other G31 VSS Power/Other K1 VCACHE Power/Other H1 VCACHE Power/Other K2 VSS Power/Other H2 VSS Power/Other K3 VCACHE Power/Other H3 VCACHE Power/Other K4 VSS Power/Other H4 VSS Power/Other K5 VCACHE Power/Other H5 VCACHE Power/Other K6 VSS Power/Other H6 VSS Power/Other K7 VCACHE Power/Other H7 VCACHE Power/Other K8 VSS Power/Other H8 VSS Power/Other K9 VCACHE Power/Other H9 VCACHE Power/Other K23 VCC Power/Other H23 VCC Power/Other K24 VSS Power/Other H24 VSS Power/Other K25 VCC Power/Other H25 VCC Power/Other K26 VSS Power/Other H26 VSS Power/Other K27 VCC Power/Other H27 VCC Power/Other K28 VSS Power/Other H28 VSS Power/Other K29 VCC Power/Other H29 VCC Power/Other K30 VSS Power/Other H30 VSS Power/Other K31 VCC Power/Other H31 VCC Power/Other L1 VSS Power/Other J1 VSS Power/Other L2 VCC Power/Other J2 VCC Power/Other L3 VSS Power/Other J3 VSS Power/Other L4 VCC Power/Other J4 VCC Power/Other L5 VSS Power/Other J5 VSS Power/Other L6 VCC Power/Other J6 VCC Power/Other L7 VSS Power/Other J7 VSS Power/Other L8 VCC Power/Other J8 VCC Power/Other L9 VSS Power/Other Input 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Direction 81 Pin Listing Table 5-2. Pin Listing by Pin Number (cont’d) Pin No. Pin Name Signal Buffer Type Direction Table 5-2. Pin Listing by Pin Number (cont’d) Pin No. Pin Name Signal Buffer Type L23 VSS Power/Other N24 VSS Power/Other L24 VCC Power/Other N25 VCC Power/Other L25 VSS Power/Other N26 VSS Power/Other L26 VCC Power/Other N27 VCC Power/Other L27 VSS Power/Other N28 VSS Power/Other L28 VCC Power/Other N29 VCC Power/Other L29 VSS Power/Other N30 VSS Power/Other L30 VCC Power/Other N31 VCC Power/Other L31 VSS Power/Other P1 VSS Power/Other M1 VCACHE Power/Other P2 VCC Power/Other M2 VSS Power/Other P3 VSS Power/Other M3 VCACHE Power/Other P4 VCC Power/Other M4 VSS Power/Other P5 VSS Power/Other M5 VCACHE Power/Other P6 VCC Power/Other M6 VSS Power/Other P7 VSS Power/Other M7 VCACHE Power/Other P8 VCC Power/Other M8 VSS Power/Other P9 VSS Power/Other M9 VCACHE Power/Other P23 VSS Power/Other M23 VCC Power/Other P24 VCC Power/Other M24 VSS Power/Other P25 VSS Power/Other M25 VCC Power/Other P26 VCC Power/Other M26 VSS Power/Other P27 VSS Power/Other M27 VCC Power/Other P28 VCC Power/Other M28 VSS Power/Other P29 VSS Power/Other M29 VCC Power/Other P30 VCC Power/Other M30 VSS Power/Other P31 VSS Power/Other M31 VCC Power/Other R1 VCACHE Power/Other N1 VCACHE Power/Other R2 VSS Power/Other N2 VSS Power/Other R3 VCACHE Power/Other N3 VCACHE Power/Other R4 VSS Power/Other N4 VSS Power/Other R5 VCACHE Power/Other N5 VCACHE Power/Other R6 VSS Power/Other N6 VSS Power/Other R7 VCACHE Power/Other N7 VCACHE Power/Other R8 VSS Power/Other N8 VSS Power/Other R9 VCACHE Power/Other N9 VCACHE Power/Other R23 VCC Power/Other N23 VCC Power/Other R24 VSS Power/Other 82 Direction 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Pin Listing Table 5-2. Pin Listing by Pin Number (cont’d) Pin No. Pin Name Signal Buffer Type Direction Table 5-2. Pin Listing by Pin Number (cont’d) Pin No. Pin Name Signal Buffer Type Direction R25 VCC Power/Other U26 VSS Power/Other R26 VSS Power/Other U27 VCC Power/Other R27 VCC Power/Other U28 VSS Power/Other R28 VSS Power/Other U29 VCC Power/Other R29 VCC Power/Other U30 VSS Power/Other R30 VSS Power/Other U31 VCC Power/Other R31 VCC Power/Other V1 VSS Power/Other T1 VSS Power/Other V2 VCC Power/Other T2 VCC Power/Other V3 VSS Power/Other T3 VSS Power/Other V4 VCC Power/Other T4 VCC Power/Other V5 VSS Power/Other T5 VSS Power/Other V6 VCC Power/Other T6 VCC Power/Other V7 VSS Power/Other T7 VSS Power/Other V8 VCC Power/Other T8 VCC Power/Other V9 VSS Power/Other T9 VSS Power/Other V23 VSS Power/Other T23 VSS Power/Other V24 VCC Power/Other T24 VCC Power/Other V25 VSS Power/Other T25 VSS Power/Other V26 VCC Power/Other T26 VCC Power/Other V27 VSS Power/Other T27 VSS Power/Other V28 VCC Power/Other T28 VCC Power/Other V29 VSS Power/Other T29 VSS Power/Other V30 VCC Power/Other T30 VCC Power/Other V31 VSS Power/Other T31 VSS Power/Other W1 VCC Power/Other U1 VCACHE Power/Other W2 VSS Power/Other U2 VSS Power/Other W3 Reserved U3 VCACHE Power/Other W4 VSS Power/Other U4 VSS Power/Other W5 BCLK1 FSB Clk Input U5 VCACHE Power/Other W6 TESTHI0 Power/Other Input U6 VSS Power/Other W7 TESTHI1 Power/Other Input U7 VCACHE Power/Other W8 TESTHI2 Power/Other Input U8 VSS Power/Other W9 GTLREF1 Power/Other Input U9 VCACHE Power/Other W23 GTLREF0 Power/Other Input U23 VCC Power/Other W24 VSS Power/Other U24 VSS Power/Other W25 VCC Power/Other U25 VCC Power/Other W26 VSS Power/Other 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 83 Pin Listing Table 5-2. Pin Listing by Pin Number (cont’d) Pin No. Pin Name Signal Buffer Type Direction Table 5-2. Pin Listing by Pin Number (cont’d) Pin No. Pin Name Direction W27 VCC Power/Other AA1 W28 VSS Power/Other AA2 VSS Power/Other W29 VCC Power/Other AA3 BSEL0 Power/Other Output W30 VSS Power/Other AA4 DEP7# Source Sync Input/Output W31 VCC Power/Other AA5 VSSA Power/Other Input Y1 VSS Power/Other AA6 VCC Power/Other Y2 VCC Power/Other AA7 TESTHI4 Power/Other Input Y3 VSS Power/Other AA8 D61# Source Sync Input/Output Y4 BCLK0 FSB Clk AA9 VSS Power/Other Y5 VSS Power/Other Y6 TESTHI3 Power/Other Y7 VSS Power/Other Y8 RESET# Common Clk Input Y9 D62# Source Sync Input/Output Y10 VTT Power/Other Y11 DSTBP3# Source Sync Input/Output Y12 DSTBN3# Source Sync Input/Output Y13 VSS Power/Other Y14 DSTBP2# Source Sync Input/Output Y15 DSTBN2# Source Sync Input/Output Y16 VCC Power/Other Y17 DSTBP1# Source Sync Input/Output Input/Output Input Input Y18 DSTBN1# Source Sync Y19 VSS Power/Other Y20 DSTBP0# Source Sync Input/Output Y21 DSTBN0# Source Sync Input/Output Y22 VCC Power/Other Y23 D5# Source Sync Input/Output Y24 D2# Source Sync Input/Output Y25 VSS Power/Other Y26 D0# Source Sync Y27 Reserved Y28 Reserved Y29 SM_TS1_A1 SMBus Y30 VCC Power/Other Y31 VSS Power/Other 84 Input/Output Input VCC Signal Buffer Type Power/Other AA10 D54# Source Sync Input/Output AA11 D53# Source Sync Input/Output AA12 VTT Power/Other AA13 D48# Source Sync Input/Output AA14 D49# Source Sync Input/Output AA15 VSS Power/Other AA16 D33# Source Sync AA17 VSS Power/Other AA18 D24# Source Sync Input/Output AA19 D15# Source Sync Input/Output AA20 VCC Power/Other AA21 D11# Source Sync Input/Output AA22 D10# Source Sync Input/Output AA23 VSS Power/Other AA24 D6# Source Sync Input/Output AA25 D3# Source Sync Input/Output Input/Output AA26 VCC Power/Other AA27 D1# Source Sync Input/Output AA28 SM_TS1_A0 SMBus Input AA29 SM_EP_A0 SMBus Input AA30 VSS Power/Other AA31 VCC Power/Other AB1 VSS Power/Other AB2 VCC Power/Other AB3 BSEL1 Power/Other Output AB4 VCCA Power/Other Input AB5 VSS Power/Other AB6 D63# Source Sync Input/Output 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Pin Listing Table 5-2. Pin Listing by Pin Number (cont’d) Pin No. AB7 Pin Name PWRGOOD Signal Buffer Type Async GTL+ Direction Input Table 5-2. Pin Listing by Pin Number (cont’d) Pin No. Pin Name Signal Buffer Type Direction AC13 VSS Power/Other AC14 D50# Source Sync Input/Output AC15 DP2# Common Clk Input/Output AC16 VCC Power/Other AC17 D34# Source Sync Input/Output AC18 DP0# Common Clk Input/Output AB8 VCC Power/Other AB9 DBI3# Source Sync Input/Output AB10 D55# Source Sync Input/Output AB11 VSS Power/Other AB12 D51# Source Sync Input/Output AB13 D52# Source Sync Input/Output AB14 VCC Power/Other AC19 VSS Power/Other AB15 D37# Source Sync Input/Output AC20 D25# Source Sync Input/Output AB16 D32# Source Sync Input/Output AC21 D26# Source Sync Input/Output AB17 D31# Source Sync Input/Output AC22 VCC Power/Other AB18 VCC Power/Other AC23 D23# Source Sync Input/Output AB19 D14# Source Sync Input/Output AC24 D20# Source Sync Input/Output AB20 D12# Source Sync Input/Output AC25 VSS Power/Other AB21 VSS Power/Other AC26 D17# Source Sync Input/Output AB22 D13# Source Sync Input/Output AC27 DBI0# Source Sync Input/Output AB23 D9# Source Sync Input/Output AC28 SM_CLK SMBus Input AB24 VCC Power/Other AC29 SM_DAT SMBus Output AB25 D8# Source Sync Input/Output AC30 SLEW_CTRL Power/Other Input AB26 D7# Source Sync Input/Output AC31 VCC Power/Other AB27 VSS Power/Other AD1 VCCPLL Power/Other AB28 SM_EP_A2 SMBus Input AD2 VCC Power/Other AB29 SM_EP_A1 SMBus Input AD3 VSS Power/Other AB30 VCC Power/Other AD4 VCCIOPLL Power/Other Input AB31 VSS Power/Other AD5 TESTHI5 Power/Other Input AC1 Reserved AD6 DEP5# Source Sync Input/Output AC2 VSS Power/Other AD7 D57# Source Sync Input/Output AC3 VCC Power/Other AD8 D46# Source Sync Input/Output AC4 DEP6# Source Sync Input/Output AD9 VSS Power/Other AC5 D60# Source Sync Input/Output AD10 D45# Source Sync Input/Output AC6 D59# Source Sync Input/Output AD11 D40# Source Sync Input/Output AC7 VSS Power/Other AD12 VTT Power/Other AC8 D56# Source Sync Input/Output AD13 D38# Source Sync Input/Output AC9 D47# Source Sync Input/Output AD14 D39# Source Sync Input/Output AC10 VTT Power/Other AD15 VSS Power/Other AC11 D43# Source Sync Input/Output AD16 COMP0 Power/Other AC12 D41# Source Sync Input/Output AD17 VSS Power/Other 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Input Input 85 Pin Listing Table 5-2. Pin Listing by Pin Number (cont’d) Pin No. AD18 Pin Name D36# Table 5-2. Pin Listing by Pin Number (cont’d) Signal Buffer Type Direction Signal Buffer Type Direction Source Sync Input/Output AE12 DBI2# Source Sync Input/Output Input/Output AE13 D35# Source Sync Input/Output AE14 VCC Power/Other Pin No. Pin Name AD19 D30# Source Sync AD20 VCC Power/Other AD21 D29# Source Sync Input/Output AE15 DEP3# Source Sync Input/Output AD22 DBI1# Source Sync Input/Output AE16 DEP2# Source Sync Input/Output AD23 VSS Power/Other AE17 DP3# D21# Source Sync Input/Output Common Clk Input/Output AD24 AD25 D18# Source Sync Input/Output AE18 VCC Power/Other AD26 VCC Power/Other AE19 DP1# Common Clk Input/Output Input/Output AD27 D4# Source Sync Input/Output AE20 D28# Source Sync AD28 SM_ALERT# SMBus Output AE21 VSS Power/Other AD29 SM_WP SMBus Input AE22 D27# Source Sync Input/Output AD30 DEP1# Source Sync Input/Output AE23 D22# Source Sync Input/Output AD31 DEP0# Source Sync Input/Output AE24 VCC Power/Other AE2 VSSA_CACHE Power/Other Input AE25 D19# Source Sync Input/Output AE3 VCCA_CACHE Power/Other Input AE26 D16# Source Sync Input/Output AE4 Reserved AE27 VSS Power/Other AE5 TESTHI6 Power/Other Input AE28 SM_VCC Power/Other AE6 VSS Power/Other AE29 SM_VCC Power/Other AE7 D58# Source Sync Input/Output AE30 Reserved AE8 DEP4# Source Sync Input/Output AE9 D44# Source Sync Input/Output AE10 D42# Source Sync Input/Output AE11 VSS Power/Other NOTES: 1. In systems utilizing the 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache, the system designer must pull-up these signals to the processor VTT. § 86 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 6 Signal Definitions 6.1 Signal Definitions Table 6-1. Signal Definitions (Sheet 1 of 9) Name A[39:3]# A20M# Type Description I/O A[39:3]# (Address) define a 240-byte physical memory address space. In subphase 1 of the address phase, these pins transmit the address of a transaction. In sub-phase 2, these pins transmit transaction type information. These signals must connect the appropriate pins of all agents on the processor front side bus. A[39:3]# are protected by parity signals AP[1:0]#. A[39:3]# are source synchronous signals and are latched into the receiving buffers by ADSTB[1:0]#. On the active-to-inactive transition of RESET#, the processors sample a subset of the A[39:3]# pins to determine their power-on configuration. See Section 8.1. I If A20M# (Address-20 Mask) is asserted, the processor masks physical address bit 20 (A20#) before looking up a line in any internal cache and before driving a read/write transaction on the bus. Asserting A20M# emulates the 8086 processor's address wrap-around at the 1-Mbyte boundary. Assertion of A20M# is only supported in real mode. A20M# is an asynchronous signal. However, to ensure recognition of this signal following an I/O write instruction, it must be valid 6 clks before the I/O write’s response. ADS# I/O ADS# (Address Strobe) is asserted to indicate the validity of the transaction address on the A[39:3]# and transaction request type on REQ[4:0]# pins. All bus agents observe the ADS# activation to begin parity checking, protocol checking, address decode, internal snoop, or deferred reply ID match operations associated with the new transaction. This signal must connect the appropriate pins on all processor front side bus agents. ADSTB[1:0]# I/O Address strobes are used to latch A[39:3]# and REQ[4:0]# on their rising and falling edge. AP[1:0]# (Address Parity) are driven by the requestor one common clock after ADS#, A[39:3]#, REQ[4:0]# are driven. A correct parity signal is electrically high if an even number of covered signals are electrically low and electrically low if an odd number of covered signals are electrically low. This allows parity to be electrically high when all the covered signals are electrically high. AP[1:0]# should connect the appropriate pins of all processor front side bus agents. The following table defines the coverage for these signals. AP[1:0]# BCLK[1:0] I/O I Request Signals Subphase 1 Subphase 2 A[39:24]# AP0# AP1# A[23:3]# AP1# AP0# REQ[4:0]# AP1# AP0# The differential bus clock pair BCLK[1:0] determines the bus frequency. All processor front side bus agents must receive these signals to drive their outputs and latch their inputs. All external timing parameters are specified with respect to the rising edge of BCLK0 crossing the falling edge of BCLK1. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 87 Signal Definitions Table 6-1. Signal Definitions (Sheet 2 of 9) Name BINIT# BNR# BOOT_ SELECT BPM[5:0]# BPRI# 88 Type Description I/O BINIT# (Bus Initialization) may be observed and driven by all processor front side bus agents. If used, BINIT# must connect the appropriate pins of all such agents. If the BINIT# driver is enabled, BINIT# is asserted to signal any bus condition that prevents reliable future operation. If BINIT# observation is enabled during power-on configuration (see Section 8.1) and BINIT# is sampled asserted, symmetric agents reset their bus LOCK# activity and bus request arbitration state machines. The bus agents do not reset their I/O Queue (IOQ) and transaction tracking state machines upon observation of BINIT# assertion. Once the BINIT# assertion has been observed, the bus agents will re-arbitrate for the front side bus and attempt completion of their bus queue and IOQ entries. If BINIT# observation is enabled during power on configuration, a central agent may handle an assertion of BINIT# as appropriate to the error handling architecture of the system. I/O BNR# (Block Next Request) is used to assert a bus stall by any bus agent who is unable to accept new bus transactions. During a bus stall, the current bus owner cannot issue any new transactions. Since multiple agents might need to request a bus stall at the same time, BNR# is a wire-OR signal which must connect the appropriate pins of all processor system bus agents. In order to avoid wire-OR glitches associated with simultaneous edge transitions driven by multiple drivers, BNR# is activated on specific clock edges and sampled on specific clock edges. I The BOOT_SELECT input informs the processor whether the platform supports the processor. Incompatible platform designs will have this input connected to VSS. Thus, this pin is essentially an electrical key to prevent the processor from running in a system that is not designed for it. For platforms that are designed to support the 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache, this pin should be a no-connect. I/O BPM[5:0]# (Breakpoint Monitor) are breakpoint and performance monitor signals. They are outputs from the processor which indicate the status of breakpoints and programmable counters used for monitoring processor performance. BPM[5:0]# should connect the appropriate pins of all processor front side bus agents. BPM4# provides PRDY# (Probe Ready) functionality for the TAP port. PRDY# is a processor output used by debug tools to determine processor debug readiness. BPM5# provides PREQ# (Probe Request) functionality for the TAP port. PREQ# is a processor input and is used by debug tools to request debug operation of the processors. BPM[5:4]# must be bussed to all bus agents. I BPRI# (Bus Priority Request) is used to arbitrate for ownership of the processor front side bus. It must connect the appropriate pins of all processor front side bus agents. Observing BPRI# active (as asserted by the priority agent) causes all other agents to stop issuing new requests, unless such requests are part of an ongoing locked operation. The priority agent keeps BPRI# asserted until its requests are issued, then releases the bus by deasserting BPRI#. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Signal Definitions Table 6-1. Signal Definitions (Sheet 3 of 9) Name Type Description BR[3:0]# (Bus Request) drive the BREQ[3:0]# signals in the system. The BREQ[3:0]# signals are interconnected in a rotating manner to individual processor pins. The tables below give the rotating interconnect between the processor and bus signals for 3-load configurations. BR[1:0]# Signals Rotating Interconnect, 3-Load Configuration BR0# BR[3:1]# I/O I Bus Signal Agent 0 Pins Agent 1 Pins BREQ0# BR0# BR1# BREQ1# BR1# BR0# BR2# and BR3# must not be utilized in 3-load configurations. However, they must still be terminated. During power-on configuration, the central agent must assert the BR0# bus signal. All symmetric agents sample their BR[3:0]# pins on the active-toinactive transition of RESET#. The pin which the agent samples asserted determines its agent ID. BSEL[1:0] O These output signals are used to select the front side bus frequency. The frequency is determined by the processor(s), chipset, and frequency synthesizer capabilities. All front side bus agents must operate at the same frequency. Individual processors will only operate at their specified front side bus frequency. See Table 2-2 for output values. COMP0 I COMP0 must be terminated to VSS on the baseboard using precision resistors. This input configures the AGTL+ drivers of the processor. Refer to Table 2-19. O CVID[3:0] (Cache Voltage ID) pins are used to support automatic selection of VCACHE. These are open drain signals that are driven by the processor and must be pulled to no more than 3.3 V (+5% tolerance) with a resistor. Conversely, the VCACHE VR output must be disabled prior to the voltage supply for these pins becoming invalid. The CVID pins are needed to support processor voltage specification variations. See Table 2-4 for definitions of these pins. The VCACHE VR must supply the voltage that is requested by these pins, or disable itself. I/O D[63:0]# (Data) are the data signals. These signals provide a 64-bit data path between the processor front side bus agents, and must connect the appropriate pins on all such agents. The data driver asserts DRDY# to indicate a valid data transfer. D[63:0]# are quad-pumped signals, and will thus be driven four times in a common clock period. D[63:0]# are latched off the falling edge of both DSTBP[3:0]# and DSTBN[3:0]#. Each group of 16 data signals correspond to a pair of one DSTBP# and one DSTBN#. The following table shows the grouping of data signals to strobes and DBI#. Furthermore, the DBI# pins determine the polarity of the data signals. Each group of 16 data signals corresponds to one DBI# signal. When the DBI# signal is active, the corresponding data group is inverted and therefore sampled active high. I/O DBI[3:0]# are source synchronous and indicate the polarity of the D[63:0]# and DEP[7:0]# signals. The DBI[3:0]# signals are activated when the data on the data bus is inverted. If more than half the data bits, within an 18-bit group (including ECC bits), would have been asserted electrically low, the bus agent may invert the data bus and corresponding ECC signals for that particular subphase for that 18-bit group. CVID[3:0] D[63:0]# DBI[3:0]# 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 89 Signal Definitions Table 6-1. Signal Definitions (Sheet 4 of 9) Name Type Description I/O DBSY# (Data Bus Busy) is asserted by the agent responsible for driving data on the processor front side bus to indicate that the data bus is in use. The data bus is released after DBSY# is deasserted. This signal must connect the appropriate pins on all processor front side bus agents. I DEFER# is asserted by an agent to indicate that a transaction cannot be guaranteed in-order completion. Assertion of DEFER# is normally the responsibility of the addressed memory or I/O agent. This signal must connect the appropriate pins of all processor front side bus agents. DEP[7:0]# I/O The DEP[7:0]# (data bus ECC protection) signals provide optional ECC protection for the data bus. They are driven by the agent responsible for driving D[63:0]#, and, if ECC is implemented, must connect the appropriate pins of all bus agents which use them. Furthermore, the DBI# pins determine the polarity of the ECC signals. Each pair of 2 ECC signals corresponds to one DBI# signal. When the DBI# signal is active, the corresponding ECC pair is inverted and therefore sampled active high. DP[3:0]# I/O DP[3:0]# (Data Parity) provide optional parity protection for the data bus. They are driven by the agent responsible for driving D[63:0]#, and, if parity is implemented, must connect the appropriate pins of all bus agents which use them. DRDY# I/O DRDY# (Data Ready) is asserted by the data driver on each data transfer, indicating valid data on the data bus. In a multi-common clock data transfer, DRDY# may be deasserted to insert idle clocks. This signal must connect the appropriate pins of all processor front side bus agents. DSTBN[3:0]# I/O Data strobe used to latch in D[63:0]#, DEP[7:0]# and DBI[3:0]#. DSTBP[3:0]# I/O Data strobe used to latch in D[63:0]#, DEP[7:0]# and DBI[3:0]#. FERR#/PBE# O FERR#/PBE# (floating-point error/pending break event) is a multiplexed signal and its meaning is qualified by STPCLK#. When STPCLK# is not asserted, FERR#/PBE# indicates a floating-point error and will be asserted when the processor detects an unmasked floating-point error. When STPCLK# is not asserted, FERR#/PBE# is similar to the ERROR# signal on the Intel 387 coprocessor, and is included for compatibility with systems using MS-DOS*type floating-point error reporting. When STPCLK# is asserted, an assertion of FERR#/PBE# indicates that the processor has a pending break event waiting for service. The assertion of FERR#/PBE# indicates that the processor should be returned to the Normal state. For additional information on the pending break event functionality, including the identification of support of the feature and enable/disable information, refer to Volume 3 of the IA-32 Intel® Architecture Software Developer’s Manual and the AP-485 Intel® Processor Identification and the CPUID Instruction application note. FORCEPR# I This input can be used to force activation of the Thermal Control Circuit. GTLREF[3:0] I GTLREF determines the signal reference level for AGTL+ input pins. GTLREF is used by the AGTL+ receivers to determine if a signal is an electrical 0 or an electrical 1. Please refer to Table 2-19 for further details. I/O I/O HIT# (Snoop Hit) and HITM# (Hit Modified) convey transaction snoop operation results. Any front side bus agent may assert both HIT# and HITM# together to indicate that it requires a snoop stall, which can be continued by reasserting HIT# and HITM# together, every other common clock. Since multiple agents may deliver snoop results at the same time, HIT# and HITM# are wire-OR signals which must connect the appropriate pins of all processor front side bus agents. In order to avoid wire-OR glitches associated with simultaneous edge transitions driven by multiple drivers, HIT# and HITM# are activated on specific clock edges and sampled on specific clock edges. DBSY# DEFER# HIT# HITM# 90 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Signal Definitions Table 6-1. Signal Definitions (Sheet 5 of 9) Name Type Description ID[7:0]# I ID[7:0]# are the Transaction ID signals. They are driven during the Deferred Phase by the deferring agent. IDS# I IDS# is the ID Strobe signal. It is asserted to begin the Deferred Phase. O IERR# (Internal Error) is asserted by a processor as the result of an internal error. Assertion of IERR# is usually accompanied by a SHUTDOWN transaction on the processor front side bus. This transaction may optionally be converted to an external error signal (e.g., NMI) by system core logic. The processor will keep IERR# asserted until the assertion of RESET#. I IGNNE# (Ignore Numeric Error) is asserted to force the processor to ignore a numeric error and continue to execute noncontrol floating-point instructions. If IGNNE# is deasserted, the processor generates an exception on a noncontrol floating-point instruction if a previous floating-point instruction caused an error. IGNNE# has no effect when the NE bit in control register 0 (CR0) is set. IGNNE# is an asynchronous signal. However, to ensure recognition of this signal following an I/O write instruction, it must be valid a 6 clks before the I/O write’s response. I INIT# (Initialization), when asserted, resets integer registers inside all processors without affecting their internal caches or floating-point registers. Each processor then begins execution at the power-on Reset vector configured during power-on configuration. The processor continues to handle snoop requests during INIT# assertion. INIT# is an asynchronous signal and must connect the appropriate pins of all processor front side bus agents. If INIT# is sampled active on the active to inactive transition of RESET#, then the processor executes its Built-in Self-Test (BIST). I LINT[1:0] (Local APIC Interrupt) must connect the appropriate pins of all front side bus agents. When the APIC functionality is disabled, the LINT0 signal becomes INTR, a maskable interrupt request signal, and LINT1 becomes NMI, a nonmaskable interrupt. INTR and NMI are backward compatible with the signals of those names on the Pentium® processor. Both signals are asynchronous. These signals must be software configured via BIOS programming of the APIC register space to be used either as NMI/INTR or LINT[1:0]. Because the APIC is enabled by default after Reset, operation of these pins as LINT[1:0] is the default configuration. I/O LOCK# indicates to the system that a set of transactions must occur atomically. This signal must connect the appropriate pins of all processor front side bus agents. For a locked sequence of transactions, LOCK# is asserted from the beginning of the first transaction to the end of the last transaction. When the priority agent asserts BPRI# to arbitrate for ownership of the processor front side bus, it will wait until it observes LOCK# deasserted. This enables symmetric agents to retain ownership of the processor front side bus throughout the bus locked operation and ensure the atomicity of lock. IERR# IGNNE# INIT# LINT0/INTR LINT1/NMI LOCK# 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 91 Signal Definitions Table 6-1. Signal Definitions (Sheet 6 of 9) Name MCERR# Description I/O MCERR# (Machine Check Error) is asserted to indicate an unrecoverable error or a bus protocol violation. It may be driven by all processor front side bus agents. MCERR# assertion conditions are configurable at a system level. Assertion options are defined as follows: • Enabled or disabled. • Asserted, if configured, for internal errors along with IERR#. • Asserted, if configured, by the request initiator of a bus transaction after it observes an error. • Asserted by any bus agent when it observes an error in a bus transaction. For more details regarding machine check architecture, refer to the IA-32 Intel® Software Developer’s Manual, Volume 3: System Programming Guide. Since multiple agents may drive this signal at the same time, MCERR# is a wired-OR signal which must connect the appropriate pins of all processor front side bus agents. In order to avoid wire-OR glitches associated with simultaneous edge transitions driven by multiple drivers, MCERR# is activated on specific clock edges and sampled on specific clock edges. ODTEN I ODTEN (On-die termination enable) should be connected to VTT through a resistor to enable on-die termination for end bus agents. For middle bus agents, pull this signal down via a resistor to ground to disable on-die termination. Whenever ODTEN is high, on-die termination will be active, regardless of other states of the bus. OOD# I OOD# allows data delivery to occur subsequent to IDS# assertion during the Deferred Phase. PROCHOT# O The assertion of PROCHOT# (processor hot) indicates that the processor die temperature has reached its thermal limit. See Section 7.2.4 for more details. I PWRGOOD (Power Good) is an input. The processor requires this signal to be a clean indication that all processor clocks and power supplies are stable and within their specifications. “Clean” implies that the signal will remain low (capable of sinking leakage current), without glitches, from the time that the power supplies are turned on until they come within specification. The signal must then transition monotonically to a high state. Figure 2-18 illustrates the relationship of PWRGOOD to the RESET# signal. PWRGOOD can be driven inactive at any time, but clocks and power must again be stable before a subsequent rising edge of PWRGOOD. It must also meet the minimum pulse width specification in Table 2-21, and be followed by a 1 ms active RESET# pulse. The PWRGOOD signal must be supplied to the processor. This signal is used to protect internal circuits against voltage sequencing issues. It should be driven high throughout boundary scan operation. I/O REQ[4:0]# (Request Command) must connect the appropriate pins of all processor front side bus agents. They are asserted by the current bus owner to define the currently active transaction type. These signals are source synchronous to ADSTB[1:0]#. Refer to the AP[1:0]# signal description for details on parity checking of these signals. I Asserting the RESET# signal resets all processors to known states and invalidates their internal caches without writing back any of their contents. For a power-on Reset, RESET# must stay active for at least 1 ms after VCC and BCLK have reached their specified levels. On observing active RESET#, all front side bus agents will deassert their outputs within two clocks. RESET# must not be kept asserted for more than 10 ms. A number of bus signals are sampled at the active-to-inactive transition of RESET# for power-on configuration. These configuration options are described in Section 8.1. PWRGOOD REQ[4:0]# RESET# 92 Type 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Signal Definitions Table 6-1. Signal Definitions (Sheet 7 of 9) Name Type Description I RS[2:0]# (Response Status) are driven by the response agent (the agent responsible for completion of the current transaction), and must connect to the appropriate pins of all processor front side bus agents. RSP# I RSP# (Response Parity) is driven by the response agent (the agent responsible for completion of the current transaction) during assertion of RS[2:0]#, the signals for which RSP# provides parity protection. It must connect to the appropriate pins of all processor front side bus agents. A correct parity signal is electrically high if an even number of covered signals are electrically low and electrically low if an odd number of covered signals are electrically low. If RS[2:0]# are all electrically high, RSP# is also electrically high, since this indicates it is not being driven by any agent guaranteeing correct parity. SKTOCC# O SKTOCC# (Socket occupied) will be pulled to ground by the processor to indicate that the processor is present. There is no connection to the processor silicon for this signal. SLEW_CTRL I SLEW_CTRL must be terminated to VSS on the baseboard using precision resistors. This input configures the slew rate of the AGTL+ drivers. Refer to Table 2-19 for implementation details. SM_ALERT# O SM_ALERT# (SMBus Alert) is an asynchronous interrupt line associated with the SMBus Thermal Sensor device. It is an open-drain output and the processor includes a 10kΩ pull-up resistor to SM_VCC for this signal. For more information on the usage of the SM_ALERT# pin, see Section 8.4.7. SM_CLK I/O The SM_CLK (SMBus Clock) signal is an input clock to the system management logic which is required for operation of the system management features of the processor. This clock is driven by the SMBus controller and is asynchronous to other clocks in the processor.The processor includes a 10 kΩ pull-up resistor to SM_VCC for this signal. SM_DAT I/O The SM_DAT (SMBus Data) signal is the data signal for the SMBus. This signal provides the single-bit mechanism for transferring data between SMBus devices. The processor includes a 10k Ω pull-up resistor to SM_VCC for this signal. I The SM_EP_A (EEPROM Select Address) pins are decoded on the SMBus in conjunction with the upper address bits in order to maintain unique addresses on the SMBus in a system with multiple processors. To set an SM_EP_A line high, a pull-up resistor should be used that is no larger than 1 kΩ. The processor includes a 10 kΩ pull-down resistor to VSS for each of these signals. For more information on the usage of these pins, see Section 8.4.8. SM_TS_A[1:0] I The SM_TS_A (Thermal Sensor Select Address) pins are decoded on the SMBus in conjunction with the upper address bits in order to maintain unique addresses on the SMBus in a system with multiple processors. The device’s addressing, as implemented, includes a Hi-Z state for both address pins. The use of the Hi-Z state is achieved by leaving the input floating (unconnected). For more information on the usage of these pins, see Section 8.4.8. SM_VCC I SM_VCC provides power to the SMBus components on the processor package. SM_WP I WP (Write Protect) can be used to write protect the Scratch EEPROM. The Scratch EEPROM is write-protected when this input is pulled high to SM_VCC. The processor includes a 10 kΩ pull-down resistor to VSS for this signal. RS[2:0]# SM_EP_A[2:0] 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 93 Signal Definitions Table 6-1. Signal Definitions (Sheet 8 of 9) Name Description I SMI# (System Management Interrupt) is asserted asynchronously by system logic. On accepting a System Management Interrupt, processors save the current state and enter System Management Mode (SMM). An SMI Acknowledge transaction is issued, and the processor begins program execution from the SMM handler. It is required that SMI# assertion be observed 8 BCLKs before the Response Status (RS[2:0]#) is observed by the processor. If SMI# is asserted during the deassertion of RESET#, the processor will tristate its outputs. STPCLK# I STPCLK# (Stop Clock), when asserted, causes processors to enter a low power Stop-Grant state. The processor issues a Stop-Grant Acknowledge transaction, and stops providing internal clock signals to all processor core units except the front side bus and APIC units. The processor continues to snoop bus transactions and service interrupts while in Stop-Grant state. When STPCLK# is deasserted, the processor restarts its internal clock to all units and resumes execution. The assertion of STPCLK# has no effect on the bus clock; STPCLK# is an asynchronous input. TCK I TCK (Test Clock) provides the clock input for the processor Test Access Port. TDI I TDI (Test Data In) transfers serial test data into the processor. TDI provides the serial input needed for JTAG specification support. TDO O TDO (Test Data Out) transfers serial test data out of the processor. TDO provides the serial output needed for JTAG specification support. TEST_BUS I Must be connected to all other processor TEST_BUS signals in the system. TESTHI[6:0] I TESTHI[6:0] must be connected to a VTT power source through a resistor for proper processor operation. See Section 2.4 for more details. THERMTRIP# O Assertion of THERMTRIP# (Thermal Trip) indicates the processor junction temperature has reached a temperature beyond which permanent silicon damage may occur. THERMTRIP# (Thermal Trip) will activate at a temperature that is approximately 20°C above the maximum case temperature (TC). Measurement of the temperature is accomplished through an internal thermal sensor. Upon assertion of THERMTRIP#, the processor will shut off its internal clocks (thus halting program execution) in an attempt to reduce the processor junction temperature. To protect the processor its core voltage (VCC) must be removed following the assertion of THERMTRIP#. See Figure 2-15 and Table 2-23 for the appropriate power down sequence and timing requirements. Driving of the THERMTRIP# signals is enabled within 10 µs of the assertion of PWRGOOD and is disabled on de-assertion of PWRGOOD. Once activated, THERMTRIP# remains latched until PWRGOOD is de-asserted. While the deassertion of the PWRGOOD signal will de-assert THERMTRIP#, if the processor’s junction temperature remains at or above the trip level, THERMTRIP# will again be asserted within 10 µs of the assertion of PWRGOOD. Thermtrip should not be sampled until 10 µs after PWRGOOD assertion at the processor. TMS I TMS (Test Mode Select) is a JTAG specification support signal used by debug tools. TRDY# I TRDY# (Target Ready) is asserted by the target (chipset) to indicate that it is ready to receive a write or implicit writeback data transfer. TRDY# must connect the appropriate pins of all front side bus agents. TRST# I TRST# (Test Reset) resets the Test Access Port (TAP) logic. TRST# must be driven electrically low during power on Reset. VCACHE I VCACHE provides power to the L3 cache on the processor. VCC I VCC provides power to the core logic of the processor. SMI# 94 Type 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Signal Definitions Table 6-1. Signal Definitions (Sheet 9 of 9) Name Type Description VCCA I VCCA provides isolated power for the analog portion of the internal PLL’s. Use a discrete RLC filter to provide clean power. VCCA_CACHE I VCCA_CACHE provides isolated power for the L3 cache PLL. Use a discrete RLC filter to provide clean power. VCC_CACHE_SENSE VSS_CACHE_SENSE O VCC_CACHE_SENSE and VSS_CACHE_SENSE provide isolated, low impedance connections to the processor cache voltage (VCACHE) and ground (VSS). They can be used to sense or measure voltage or ground near the silicon with little noise. VCCIOPLL I VCCIOPLL provides isolated power for digital portion of the internal PLL’s. VCCPLL I The on-die PLL filter solution will not be implemented on this platform. The VCCPLL input should be left unconnected. VCCSENSE VSSSENSE O VCCSENSE and VSSSENSE provide isolated, low impedance connections to the processor core voltage (VCC) and ground (VSS). These signals must be connected to the voltage regulator feedback signals, which ensure the output voltage (i.e. processor voltage) remains within specification. VID[5:0] O VID[5:0] (Voltage ID) pins are used to support automatic selection of VCC. These are open drain signals that are driven by the processor and must be pulled to no more than 3.3 V (+5% tolerance) with a resistor. Conversely, the VCC VR output must be disabled prior to the voltage supply for these pins becoming invalid. The VID pins are needed to support processor voltage specification variations. See Table 2-3 for definitions of these pins. The VCC VR must supply the voltage that is requested by these pins, or disable itself. VIDPWRGD I The processor requires this input to determine that the supply voltage for BSEL[1:0], VID[5:0], and CVID[3:0] is stable and within specification. VSS I VSS is the ground plane for the processor. VSSA I VSSA provides an isolated, internal ground for internal PLL’s. Do not connect directly to ground. This pin is to be connected to VCCA and VCCIOPLL through a discrete filter circuit. VSSA_CACHE I VSSA_CACHE provides an isolated, internal ground for the L3 cache PLL. Do not connect directly to ground. VTT I VTT is the front side bus termination voltage. O VTTEN can be used as an output enable for the VTT regulator. VTTEN is used as an electrical key to prevent processors with mechanically-equivalent pinouts from accidentally booting in a 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache platform. Since VTTEN is an open circuit on the processor package, VTTEN must be pulled up on the motherboard. VTTEN § 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 95 Signal Definitions 96 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 7 Thermal Specifications 7.1 Package Thermal Specifications The 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache requires a thermal solution to maintain temperatures within operating limits. Any attempt to operate the processor outside these operating limits may result in permanent damage to the processor and potentially other components within the system. As processor technology changes, thermal management becomes increasingly crucial when building computer systems. Maintaining the proper thermal environment is key to reliable, long-term system operation. A complete solution includes both component and system level thermal management features. Component level thermal solutions can include active or passive heatsinks attached to the processor Integrated Heat Spreader (IHS). Typical system level thermal solutions may consist of system fans combined with ducting and venting. For more information on designing a component level thermal solution, refer to the Processor Thermal/Mechanical Design Guidelines. Note: 7.1.1 The boxed processor will ship with a component thermal solution. Refer to Section 9 for details on the boxed processor. Thermal Specifications To allow the optimal operation and long-term reliability of Intel processor-based systems, the processor must remain within the minimum and maximum case temperature (TCASE) specifications as defined by the applicable thermal profile (see Table 7-1 and Figure 7-1). Thermal solutions not designed to provide this level of thermal capability may affect the long-term reliability of the processor and system. The 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache introduces a new methodology for managing processor temperatures which is intended to support acoustic noise reduction through fan speed control and assure processor reliability. Selection of the appropriate fan speed will be based on the temperature reported by the processor’s Thermal Diode. If the diode temperature is greater than or equal to Tcontrol (see Section 7.2.7), then the processor case temperature must remain at or below the temperature as specified by the thermal profile (see Figure 7-1). If the diode temperature is less than Tcontrol, then the case temperature is permitted to exceed the thermal profile, but the diode temperature must remain at or below Tcontrol. Systems that implement fan speed control must be designed to take these conditions into account. Systems that do not alter the fan speed only need to guarantee the case temperature meets the thermal profile specifications. The processor thermal profile ensures adherence to Intel reliability requirements. The thermal profile is representative of a volumetrically unconstrained thermal solution (i.e. industry enabled 2U+ heat sink). In this scenario, it is expected that the Thermal Control Circuit (TCC) would only be activated for very brief periods of time when running the most power intensive applications. The upper point of the thermal profile consists of the Thermal Design Power (TDP) defined in Table 7-1 and the associated TCASE value. The lower point of the thermal profile consists of x = PCONTROL_BASE and y = TCASE_MAX @ PCONTROL_BASE. Pcontrol is defined as the processor 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 97 Thermal Specifications power at which TCASE, calculated from the thermal profile, corresponds to the lowest possible value of Tcontrol. This point is associated with the Tcontrol value (see Section 7.2.7). However, because Tcontrol represents a diode temperature, it is necessary to define the associated case temperature. This is TCASE_MAX @ PCONTROL_BASE. Please see Section 7.2.7 and the Processor Thermal/Mechanical Design Guidelines for proper usage of the Tcontrol specification. The case temperature is defined at the geometric top center of the processor IHS. Analysis indicates that real applications are unlikely to cause the processor to consume maximum power dissipation for sustained time periods. Intel recommends that complete thermal solution designs target the TDP indicated in Table 7-1, instead of the maximum processor power consumption. The Thermal Monitor feature is intended to help protect the processor in the event that an application exceeds the TDP recommendation for a sustained time period. For more details on this feature, refer to Section 7.2. To ensure maximum flexibility for future requirements, systems should be designed to the Flexible Motherboard (FMB) guidelines, even if a processor with a lower thermal dissipation is currently planned. Thermal Monitor or Thermal Monitor 2 feature must be enabled for the processor to remain within specification. Table 7-1. Processor Thermal Specifications Core Frequency Maximum Power3 (W) Thermal Design Power (W) Minimum TCASE (°C) Maximum TCASE (°C) Notes Launch - FMB1 136 129 5 See Figure 7-1 and Table 7-2 1,2 NOTE: 1. Thermal Design Power (TDP) should be used for processor thermal solution design targets. The TDP is not the maximum power that the processor can dissipate. TDP is measured at maximum TCASE. 2. FMB, or Flexible Motherboard, guidelines provide a design target for meeting future thermal requirements. See Section 2.10.1 for further information on FMB. 3. Maximum Power is the maximum thermal power that can be dissipated by the processor through the integrated heat spreader. Maximum Power is measured at maximum TCASE. 98 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Thermal Specifications Figure 7-1. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Thermal Profile 75 TCASE_MAX @ TDP 70 Temperature [C] 65 60 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache Thermal Profile y = 0.217 * x + 45 55 TCASE_MAX @ PCONTROL_BASE 50 TDP PCONTROL_BASE 45 0 10 20 30 40 50 60 70 80 90 100 110 120 130 Power [W] NOTE: Refer to the Processor Thermal/Mechanical Design Guidelines for system and environmental implementation details. Table 7-2. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Thermal Profile Power [W] TCASE_MAX [°C] Power [W] TCASE_MAX [°C] PCONTROL_BASE = 23 50 80 62 30 52 85 63 35 53 90 65 40 54 95 66 45 55 100 67 50 56 105 68 55 57 110 69 60 58 115 70 65 59 120 71 70 60 125 72 75 61 129 73 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 99 Thermal Specifications 7.1.2 Thermal Metrology The maximum and minimum case temperatures (TCASE) specified in Table 7-1 are measured at the geometric top center of the processor integrated heat spreader (IHS). Figure 7-2 illustrates the location where TCASE temperature measurements should be made. For detailed guidelines on temperature measurement methodology, refer to the Processor Thermal/Mechanical Design Guidelines. Figure 7-2. Case Temperature (TCASE) Measurement Location Measure from edge of IHS 19.2 mm [0.756 in] Measure T CASE at this point (geometric center of IHS) 19.2 mm [0.756 in] 53.34 mm FC-mPGA4 Package Thermal grease should cover entire area of IHS 7.2 Processor Thermal Features 7.2.1 Thermal Monitor The Thermal Monitor feature helps control the processor temperature by activating the Thermal Control Circuit (TCC) when the processor silicon reaches its maximum operating temperature. The TCC reduces processor power consumption as needed by modulating (starting and stopping) the internal processor core clocks. The Thermal Monitor (or Thermal Monitor 2) must be enabled for the processor to be operating within specifications. The temperature at which Thermal Monitor activates the thermal control circuit is not user configurable and is not software visible. Bus traffic is snooped in the normal manner, and interrupt requests are latched (and serviced during the time that the clocks are on) while the TCC is active. When the Thermal Monitor is enabled and a high temperature situation exists (i.e. TCC is active), the clocks will be modulated by alternately turning the clocks off and on at a duty cycle specific to the processor (typically 30-50%). Clocks will not be off for more than 3 microseconds when the TCC is active. Cycle times are processor speed dependent and will decrease as processor core frequencies increase. A small amount of hysteresis has been included to prevent rapid active/ 100 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Thermal Specifications inactive transitions of the TCC when the processor temperature is near its maximum operating temperature. Once the temperature has dropped below the maximum operating temperature and the hysteresis timer has expired, the TCC goes inactive and clock modulation ceases. With a thermal solution designed to meet the thermal profile, it is anticipated that the TCC would only be activated for very short periods of time when running the most power intensive applications. The processor performance impact due to these brief periods of TCC activation is expected to be so minor that it would be immeasurable. A thermal solution that is significantly under-designed may not be capable of cooling the processor even when the TCC is active continuously. The duty cycle for the TCC, when activated by the Thermal Monitor, is factory configured and cannot be modified. The Thermal Monitor does not require any additional hardware, software drivers, or interrupt handling routines. 7.2.2 Thermal Monitor 2 The processor also supports an additional power reduction capability known as Thermal Monitor 2 (TM2). This mechanism provides an efficient means for limiting the processor temperature by reducing the power consumption within the processor. The Thermal Monitor (or Thermal Monitor 2) feature must be enabled for the processor to be operating within specifications. When Thermal Monitor 2 is enabled and a high temperature situation is detected, the Thermal Control Circuit (TCC) will be activated. The TCC causes the processor to adjust its operating frequency (via the bus multiplier) and input voltage (via the VID signals). This combination of reduced frequency and VID results in a decrease to the processor power consumption. A processor enabled for Thermal Monitor 2 includes two operating points, each consisting of a specific operating frequency and voltage. The first operating point represents the normal operating condition for the processor. Under this condition, the core-frequency-to-system-bus multiplier utilized by the processor is that contained in the IA32_FLEX_BRVID_SEL MSR and the VID is that specified in Table 2-9. These parameters represent normal system operation. The second point consists of both a lower operating frequency and voltage. When the TCC is activated, the processor automatically transitions to the new frequency. This transition occurs very rapidly (on the order of 5 microseconds). During the frequency transition, the processor is unable to service any bus requests, and consequently, all bus traffic is blocked. Edge-triggered interrupts will be latched and kept pending until the processor resumes operation at the new frequency. Once the new operating frequency is engaged, the processor will transition to the new core operating voltage by issuing a new VID code to the voltage regulator. The voltage regulator must support dynamic VID steps in order to support Thermal Monitor 2. During the voltage change, it will be necessary to transition through multiple VID codes to reach the target operating voltage. Each step will be one VID table entry (see Table 2-9). The processor continues to execute instructions during the voltage transition. Operation at the lower voltage reduces the power consumption of the processor. A small amount of hysteresis has been included to prevent rapid active/inactive transitions of the TCC when the processor temperature is near its maximum operating temperature. Once the temperature has dropped below the maximum operating temperature, and the hysteresis timer has expired, the operating frequency and voltage transition back to the normal system operating point. Transition of the VID code will occur first, in order to ensure proper operation once the processor reaches its normal operating frequency. Refer to Figure 7-3 for an illustration of this ordering. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 101 Thermal Specifications Figure 7-3. Thermal Monitor 2 Frequency and Voltage Ordering TTM2 Temperature fMAX fTM2 Frequency VNOM VTM2 Vcc Time T(hysteresis) The PROCHOT# signal is asserted when a high temperature situation is detected, regardless of whether Thermal Monitor or Thermal Monitor 2 is enabled. If a processor has its Thermal Control Circuit activated via a Thermal Monitor 2 event, and an Enhanced Intel SpeedStep“ technology transition to a higher target frequency (through the applicable MSR write) is attempted, this frequency transition will be delayed until the TCC is deactivated and the TM2 event is complete. 7.2.3 On-Demand Mode The processor provides an auxiliary mechanism that allows system software to force the processor to reduce its power consumption. This mechanism is referred to as “On-Demand” mode and is distinct from the Thermal Monitor and Thermal Monitor 2 features. On-Demand mode is intended as a means to reduce system level power consumption. Systems must not rely on software usage of this mechanism to limit the processor temperature. If bit 4 of the IA_32_CLOCK_MODULATION MSR is written to a ‘1’, the processor will immediately reduce its power consumption via modulation (starting and stopping) of the internal core clock, independent of the processor temperature. When using On-Demand mode, the duty cycle of the clock modulation is programmable via bits 3:1 of the same IA_32_CLOCK_MODULATION MSR. In On-Demand mode, the duty cycle can be programmed from 12.5% on / 87.5% off to 87.5% on / 12.5% off in 12.5% increments. On-Demand mode may be used in conjunction with the Thermal Monitor or Thermal Monitor 2. If Thermal Monitor is enabled and the system tries to enable On-Demand mode at the same time the TCC is engaged, the factory configured duty cycle of the TCC will override the duty cycle selected by the On-Demand mode. 102 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Thermal Specifications 7.2.4 PROCHOT# Signal Pin An external signal, PROCHOT# (processor hot), is asserted when the processor die temperature has reached its factory configured trip point. If the Thermal Monitor is enabled (note that the Thermal Monitor must be enabled for the processor to be operating within specification), the TCC will be active when PROCHOT# is asserted. The processor can be configured to generate an interrupt upon the assertion or de-assertion of PROCHOT#. Refer to the Intel Architecture Software Developer’s Manual and the Processor BIOS Writers Guide for specific register and programming details. PROCHOT# is designed to assert at or a few degrees higher than maximum TCASE (as specified by the thermal profile) when dissipating TDP power, and cannot be interpreted as an indication of processor case temperature. This temperature delta accounts for processor package, lifetime, and manufacturing variations and attempts to ensure the Thermal Control Circuit is not activated below maximum TCASE when dissipating TDP power. There is no defined or fixed correlation between the PROCHOT# trip temperature, the case temperature, or the thermal diode temperature. Thermal solutions must be designed to the processor specifications and cannot be adjusted based on experimental measurements of TCASE, PROCHOT#, or Tdiode on random processor samples. 7.2.5 FORCEPR# Signal Pin The FORCEPR# (force power reduction) input can be used by the platform to force the processor to activate the TCC. If the Thermal Monitor is enabled, the TCC will be activated upon the assertion of the FORCEPR# signal. The TCC will remain active until the system deasserts FORCEPR#. FORCEPR# is an asynchronous input. FORCEPR# can be used to thermally protect other system components. To use the voltage regulator (VR) as an example, when the FORCEPR# pin is asserted, the TCC in the processor will activate, reducing the current consumption of the processor and the corresponding temperature of the VR. It should be noted that assertion of FORCEPR# does not automatically assert PROCHOT#. As mentioned previously, the PROCHOT# signal is asserted when a high temperature situation is detected. A minimum pulse width of 500 microseconds is recommended when FORCEPR# is asserted by the system. Sustained activation of the FORCEPR# pin may cause noticeable platform performance degradation. 7.2.6 THERMTRIP# Signal Pin Regardless of whether or not Thermal Monitor or Thermal Monitor 2 is enabled, in the event of a catastrophic cooling failure, the processor will automatically shut down when the silicon has reached an elevated temperature (refer to the THERMTRIP# definition in Table 6-1). At this point, the system bus signal THERMTRIP# will go active and stay active as described in Table 6-1. THERMTRIP# activation is independent of processor activity and does not generate any bus cycles. If THERMTRIP# is asserted, processor core voltage (VCC) and cache voltage (VCACHE) must be removed within the timeframe defined in Table 2-23 and Figure 2-15. Intel also recommends removal of VTT. 7.2.7 TCONTROL and Fan Speed Reduction TCONTROL is a temperature specification based on a temperature reading from the thermal sensor. The value for TCONTROL will be calibrated in manufacturing and configured for each processor. The TCONTROL temperature for a given processor can be obtained by reading the 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 103 Thermal Specifications IA32_TEMPERATURE_TARGET MSR in the processor. The TCONTROL value that is read from the IA32_TEMPERATURE_TARGET MSR must be converted from Hexadecimal to Decimal and added to a base value of 50 ×C. The value of TCONTROL may vary from 0x00h to 0x1Eh. When TDIODE is above TCONTROL, then TCASE must be at or below TCASE_MAX as defined by the thermal profile (see Figure 7-1 and Table 7-2). Otherwise, the processor temperature can be maintained at TCONTROL. 7.2.8 Thermal Diode The processor incorporates an on-die thermal diode. A thermal sensor located on the processor package monitors the die temperature of the processor for thermal management/long term die temperature change purposes. The thermal diode is separate from the Thermal Monitor’s thermal sensor and cannot be used to predict the behavior of the Thermal Monitor. § 104 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 8 Features 8.1 Power-On Configuration Options Several configuration options can be set by hardware. The processor samples its hardware configuration at reset, on the active-to-inactive transition of RESET#. For specifications on these options, refer to Table 8-1. The sampled information configures the processor for subsequent operation. These configuration options can only be changed by another reset. All resets configure the processor. For reset purposes, the processor does not distinguish between a “warm” reset and a “power-on” reset. Table 8-1. Power-On Configuration Option Pins Pin1,2 Configuration Option SMI# Output tri state Execute BIST (Built-In Self Test) or A[39]# for Arb ID 3 (middle agent) A[36]# for Arb ID 0 (end agent) INIT# or A[3]# In Order Queue de-pipelining (set IOQ depth to 1) A[7]# Disable MCERR# observation A[9]# Disable BINIT# observation A[10]# APIC cluster ID Disable bus parking A[12:11]# A[15]# Core Frequency-to-Front Side Bus Multiplier A[21:16]# Symmetric agent arbitration ID BR[1:0]# Disable Hyper-Threading Technology (HT Technology) A[31]# NOTE: 1. Asserting this signal during RESET# selects the corresponding option. 2. Address pins not identified in this table as configuration options should not be asserted during RESET#. 8.2 Clock Control and Low Power States The processor allows the use of HALT and Stop-Grant states to reduce power consumption by stopping the clock to internal sections of the processor, depending on each particular state. See Figure 8-1 for a visual representation of the processor low power states. The processor adds support for Enhanced HALT power down state. Refer to Figure 8-1 and the following sections. The Stop-Grant state requires chipset and BIOS support on multiprocessor systems. In a multiprocessor system, all the STPCLK# signals are bussed together, thus all processors are affected in unison. The Hyper-Threading Technology feature adds the conditions that all logical processors share the same STPCLK# signal internally. When the STPCLK# signal is asserted, the 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 105 Features processor enters the Stop-Grant state, issuing a Stop-Grant Special Bus Cycle (SBC) for each processor or logical processor. The chipset needs to account for a variable number of processors asserting the Stop-Grant SBC on the bus before allowing the processor to be transitioned into one of the lower processor power states. Refer to the applicable chipset specification for more information. 8.2.1 Normal State This is the normal operating state for the processor. 8.2.2 HALT or Enhanced Power Down State The Enhanced HALT power down state is configured and enabled via the BIOS. If the Enhanced HALT state is not enabled, the default power down state entered will be HALT. Refer to the section below for details on HALT and Enhanced HALT states. 8.2.2.1 HALT Power Down State HALT is a low power state entered when all logical processors have executed the HALT or MWAIT instruction. When one of the logical processors executes the HALT or MWAIT instruction, that logical processor is halted; however, the other processor continues normal operation. The processor transitions to the Normal state upon the occurrence of SMI#, BINIT#, INIT#, LINT[1:0] (NMI, INTR), or an interrupt delivered over the front side bus. RESET# causes the processor to immediately initialize itself. The return from a System Management Interrupt (SMI) handler can be to either Normal Mode or the HALT Power Down state. See the IA-32 Intel® Architecture Software Developer's Manual, Volume 3: System Programming Guide for more information. The system can generate a STPCLK# while the processor is in the HALT Power Down state. When the system deasserts the STPCLK# interrupt, the processor returns execution to the HALT state. While in HALT Power Down state, the processor processes bus snoops and interrupts. 8.2.2.2 Enhanced HALT Power Down State Enhanced HALT state is a low power state entered when all logical processors have executed the HALT or MWAIT instructions and Enhanced HALT state has been enabled via the BIOS. When one of the logical processors executes the HALT instruction, that logical processor is halted; however, the other processor continues normal operation. The Enhanced HALT state is generally a lower power state than the Stop Grant state. The processor automatically transitions to a lower core frequency and voltage operating point before entering the Enhanced HALT state. Note that the processor FSB frequency is not altered; only the internal core frequency is changed. When entering the low power state, the processor first switches to the lower bus ratio and then transitions to the lower VID. While in the Enhanced HALT state, the processor processes bus snoops. The processor exits the Enhanced HALT state when a break event occurs. When the processor exits the Enhanced HALT state, it first transitions the VID to the original value and then changes the bus ratio back to the original value. 106 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Features Figure 8-1. Stop Clock State Machine HALT or MWAIT Instruction and HALT Bus Cycle Generated Normal State Normal execution S De TPC -a LK ss # er te d STPCLK# De-asserted S As TPC se L rte K# d STPCLK# Asserted INIT#, BINIT#, INTR, NMI, SMI#, RESET#, FSB interrupts Enhanced HALT or HALT State BCLK running Snoops and interrupts allowed Snoop Event Occurs Snoop Event Serviced Enhanced HALT Snoop or HALT Snoop State BCLK running Service snoops to caches Stop Grant State BCLK running Snoops and interrupts allowed 8.2.3 Snoop Event Occurs Snoop Event Serviced Stop Grant Snoop State BCLK running Service snoops to caches Stop-Grant State When the STPCLK# pin is asserted, the Stop-Grant state of the processor is entered 20 bus clocks after the response phase of the processor-issued Stop Grant Acknowledge special bus cycle. Both logical processors must be in the Stop-Grant state before the deassertion of STPCLK#. Since the AGTL+ signal pins receive power from the front side bus, these pins should not be driven (allowing the level to return to VTT) for minimum power drawn by the termination resistors in this state. In addition, all other input pins on the front side bus should be driven to the inactive state. BINIT# is not serviced while the processor is in Stop-Grant state. The event is latched and can be serviced by software upon exit from the Stop-Grant state. RESET# causes the processor to immediately initialize itself, but the processor will stay in StopGrant state. A transition back to the Normal state occurs with the deassertion of the STPCLK# signal. A transition to the Grant Snoop state occurs when the processor detects a snoop on the front side bus (see Section 8.2.4). 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 107 Features While in the Stop-Grant state, SMI#, INIT#, BINIT# and LINT[1:0] are latched by the processor, and only serviced when the processor returns to the Normal state. Only one occurrence of each event is recognized upon return to the Normal state. While in Stop-Grant state, the processor processes snoops on the front side bus and latches interrupts delivered on the front side bus. The PBE# signal can be driven when the processor is in Stop-Grant state. PBE# is asserted if there is any pending interrupt latched within the processor. Pending interrupts that are blocked by the EFLAGS.IF bit being clear still cause assertion of PBE#. Assertion of PBE# indicates to system logic that it should return the processor to the Normal state. 8.2.4 Enhanced HALT Snoop State or HALT Snoop State, Stop Grant Snoop State The Enhanced HALT Snoop state is used in conjunction with the Enhanced HALT state. If Enhanced HALT state is not enabled in the BIOS, the default Snoop state entered will be the HALT Snoop state. Refer to the sections below for details on HALT Snoop state, Grant Snoop state and Enhanced HALT Snoop state. 8.2.4.1 HALT Snoop State, Stop Grant Snoop State The processor responds to snoop or interrupt transactions on the front side bus while in Stop-Grant state or in HALT Power Down state. During a snoop or interrupt transaction, the processor enters the HALT/Grant Snoop state. The processor stays in this state until the snoop on the front side bus has been serviced (whether by the processor or another agent on the front side bus) or the interrupt has been latched. After the snoop is serviced or the interrupt is latched, the processor will return to the Stop-Grant state or HALT Power Down state, as appropriate. 8.2.4.2 Enhanced HALT Snoop State The Enhanced HALT Snoop state is the default Snoop state when the Enhanced HALT state is enabled via the BIOS. The processor remains in the lower bus ratio and VID operating point of the Enhanced HALT state. While in the Enhanced HALT Snoop state, snoops and interrupt transactions are handled the same way as in the HALT Snoop state. After the snoop is serviced or the interrupt is latched, the processor returns to the Enhanced HALT state. 8.3 Enhanced Intel SpeedStep® Technology Enhanced Intel SpeedStep technology enables the processor to switch between multiple frequency and voltage points, which may result in platform power savings. In order to support this technology, the system must support dynamic VID transitions. Switching between voltage/ frequency states is software controlled. 108 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Features Note: Not all processors are capable of supporting Enhanced Intel SpeedStep technology. More details on which processor frequencies will support this feature will be provided in future releases of the Specification Update. Enhanced Intel SpeedStep technology is a technology that creates processor performance states (P-states). P-states are power consumption and capability states within the Normal state. Enhanced Intel SpeedStep technology enables real-time dynamic switching between frequency and voltage points. It alters the performance of the processor by changing the bus to core frequency ratio and voltage. This allows the processor to run at different core frequencies and voltages to best serve the performance and power requirements of the processor and system. Note that the front side bus is not altered; only the internal core frequency is changed. In order to run at reduced power consumption, the voltage is altered in step with the bus ratio. The following are key features of Enhanced Intel SpeedStep technology: • Multiple voltage/frequency operating points provide optimal performance at reduced power consumption. • Voltage/frequency selection is software controlled by writing to processor MSR’s (Model Specific Registers), thus eliminating chipset dependency. — If the target frequency is higher than the current frequency, VCC is incremented in steps (+12.5 mV) by placing a new value on the VID signals and the processor shifts to the new frequency. Note that the top frequency for the processor can not be exceeded. — If the target frequency is lower than the current frequency, the processor shifts to the new frequency and VCC is then decremented in steps (-12.5 mV) by changing the target VID through the VID signals. 8.4 System Management Bus (SMBus) Interface The processor package includes an SMBus interface which allows access to a memory component with two sections (referred to as the Processor Information ROM and the Scratch EEPROM) and a thermal sensor on the substrate. The SMBus thermal sensor may be used to read the thermal diode mentioned in Section 7.2.8. These devices and their features are described below. The SMBus thermal sensor and its associated thermal diode are not related to and are completely independent of the precision, on-die temperature sensor and thermal control circuit (TCC) of the Thermal Monitor or Thermal Monitor 2 features discussed in Section 7.2.1. The processor SMBus implementation uses the clock and data signals of the System Management Bus (SMBus) Specification. It does not implement the SMBSUS# signal. For platforms which do not implement any of the SMBus features found on the processor, all of the SMBus connections, except SM_VCC, to the socket pins may be left unconnected (SM_ALERT#, SM_CLK, SM_DAT, SM_EP_A[2:0], SM_TS_A[1:0], SM_WP). 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 109 Features Figure 8-2. Logical Schematic of SMBus Circuitry SM_VCC SM_TS_A0 SM_TS_A1 VCC SM_EP_A0 A0 SM_EP_A1 A1 SM_EP_A2 A2 SM_WP WP VCC Processor Information ROM and Scratch EEPROM (1 Kbit each) CLK A0 DATA A1 CLK DATA Thermal Sensor STDBY# ALERT# VSS VSS SM_CLK SM_DAT SM_ALERT# NOTE: Actual implementation may vary. This figure is provided to offer a general understanding of the architecture. All SMBus pull-up and pull-down resistors are 10 kΩ and located on the processor. 8.4.1 Processor Information ROM (PIROM) The lower half (128 bytes) of the SMBus memory component is an electrically programmed readonly memory with information about the processor. This information is permanently writeprotected. Table 8-2 shows the data fields and formats provided in the Processor Information ROM (PIROM). Table 8-2. Processor Information ROM Format (Sheet 1 of 3) # of Bits Offset/Section Function Notes Header: 110 00h 8 Data Format Revision Two 4-bit hex digits 01 - 02h 16 EEPROM Size Size in bytes (MSB first) 03h 8 Processor Data Address Byte pointer, 00h if not present 04h 8 Processor Core Data Address Byte pointer, 00h if not present 05h 8 L3 Cache Data Address Byte pointer, 00h if not present 06h 8 Package Data Address Byte pointer, 00h if not present 07h 8 Part Number Data Address Byte pointer, 00h if not present 08h 8 Thermal Reference Data Address Byte pointer, 00h if not present 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Features Table 8-2. Processor Information ROM Format (Sheet 2 of 3) # of Bits Offset/Section 09h Function Notes 8 Feature Data Address Byte pointer, 00h if not present 0Ah 8 Other Data Address Byte pointer, 00h if not present 0Bh 16 Reserved Reserved 0Dh 8 Checksum 1 byte checksum Processor Data: 0E - 13h 48 S-spec/QDF Number Six 8-bit ASCII characters 14h 6 2 Reserved Sample/Production Reserved (most significant bits) 00b = Sample only, 01-11b = Production 15h 8 Checksum 1 byte checksum 2 Processor Core Type From CPUID 4 Processor Core Family From CPUID 4 Processor Core Model From CPUID 4 Processor Core Stepping From CPUID 2 Reserved Reserved for future use Processor Core Data: 16 - 17h 18 - 19h 16 Reserved Reserved for future use 1A - 1Bh 16 Front Side Bus Speed 16-bit binary number (in MHz) 1Ch 2 6 Multiprocessor Support Reserved 00b = UP,01b = DP,10b = RSVD,11b = MP Reserved 1D - 1Eh 16 Maximum Core Frequency 16-bit binary number (in MHz) 1F - 20h 16 Processor Core VID VCC requested by VID outputs in mV 21 - 22h 16 Core Voltage, Minimum Minimum processor DC VCC spec in mV 23h 8 TCASE Maximum Maximum case temperature spec in °C 24h 8 Checksum 1 byte checksum 25 - 26h 16 Reserved Reserved for future use 27 - 28h 16 L2 Cache Size 16-bit binary number (in KB) 29 - 2Ah 16 L3 Cache Size 16-bit binary number (in KB) Cache Data: 2B - 2Ch 16 Processor Cache VID VCACHE requested by CVID outputs in mV 2D - 2Eh 16 Cache Voltage, Minimum Minimum processor DC VCACHE spec in mV 2F - 30h 16 Reserved Reserved 31h 8 Checksum 1 byte checksum 32 - 35h 32 Package Revision Four 8-bit ASCII characters 36h 8 Reserved Reserved for future use 37h 8 Checksum 1 byte checksum Package Data: 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 111 Features Table 8-2. Processor Information ROM Format (Sheet 3 of 3) # of Bits Offset/Section Function Notes Part Number Data: 38 - 3Eh 56 Processor Part Number Seven 8-bit ASCII characters 3F - 4Ch 112 Reserved Reserved 4D - 54h 64 Processor Electronic Signature 64-bit identification number 55 - 6Eh 208 Reserved Reserved 8 Checksum 1 byte checksum 70h 8 Reserved Reserved 71 - 72h 16 Reserved Reserved 73h 8 Reserved Reserved 32 Processor Core Feature Flags From CPUID function 1, EDX contents 6Fh Thermal Ref. Data: Feature Data: 74 - 77h 78h 8 Processor Feature Flags [7] = Reserved [6] = Serial Signature [5] = Electronic Signature Present [4] = Thermal Sense Device Present [3] = Reserved [2] = OEM EEPROM Present [1] = Core VID Present [0] = L3 Cache Present 79-7Bh 24 Additional Processor Feature Flags All bits Reserved 7Ch 8 Reserved Reserved 7Dh 8 Checksum 1 byte checksum 16 Reserved Reserved Other Data: 7E - 7Fh 8.4.2 Scratch EEPROM Also available in the memory component on the processor SMBus is an EEPROM which may be used for other data at the system or processor vendor’s discretion. The data in this EEPROM, once programmed, can be write-protected by asserting the active-high SM_WP signal. This signal has a weak pull-down (10 kW) to allow the EEPROM to be programmed in systems with no implementation of this signal. The Scratch EEPROM resides in the upper half of the memory component (addresses 80 - FFh). The lower half comprises the Processor Information ROM (addresses 00 - 7Fh), which is permanently write-protected by Intel. 112 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Features 8.4.3 PIROM and Scratch EEPROM Supported SMBus Transactions The Processor Information ROM (PIROM) responds to two SMBus packet types: Read Byte and Write Byte. However, since the PIROM is write-protected, it will acknowledge a Write Byte command but ignore the data. The Scratch EEPROM responds to Read Byte and Write Byte commands. Table 8-3 diagrams the Read Byte command. Table 8-4 diagrams the Write Byte command. Following a write cycle to the scratch ROM, software must allow a minimum of 10 ms before accessing either ROM of the processor. In the tables, ‘S’ represents the SMBus start bit, ‘P’ represents a stop bit, ‘R’ represents a read bit, ‘W’ represents a write bit, ‘A’ represents an acknowledge (ACK), and ‘///’ represents a negative acknowledge (NACK). The shaded bits are transmitted by the Processor Information ROM or Scratch EEPROM, and the bits that aren’t shaded are transmitted by the SMBus host controller. In the tables, the data addresses indicate 8 bits. The SMBus host controller should transmit 8 bits with the most significant bit indicating which section of the EEPROM is to be addressed: the Processor Information ROM (MSB = 0) or the Scratch EEPROM (MSB = 1). Table 8-3. Read Byte SMBus Packet S Slave Address Write A Command Code A S Slave Address Read A Data /// P 1 7-bits 1 1 8-bits 1 1 7-bits 1 1 8-bits 1 1 Table 8-4. Write Byte SMBus Packet 8.4.4 S Slave Address Write A Command Code A Data A P 1 7-bits 1 1 8-bits 1 8-bits 1 1 SMBus Thermal Sensor The processor’s SMBus thermal sensor provides a means of acquiring thermal data from the processor. The thermal sensor is composed of control logic, SMBus interface logic, a precision analog-to-digital converter, and a precision current source. The sensor drives a small current through the p-n junction of a thermal diode located on the processor core. The forward bias voltage generated across the thermal diode is sensed and the precision A/D converter derives a single byte of thermal reference data, or a “thermal byte reading.” The nominal precision of the least significant bit of a thermal byte is 1° Celsius. The processor incorporates the SMBus thermal sensor onto the processor package. Upper and lower thermal reference thresholds can be individually programmed for the SMBus thermal sensor. Comparator circuits sample the register where the single byte of thermal data (thermal byte reading) is stored. These circuits compare the single-byte result against programmable threshold bytes. If enabled, the alert signal on the processor SMBus (SM_ALERT#) will be asserted when the sensor detects that either threshold is reached or crossed. Analysis of SMBus thermal sensor data may be useful in detecting changes in the system environment that may require attention. The processor SMBus thermal sensor may be used to monitor long term temperature trends, but can not be used to manage the short term temperature of the processor or predict the activation of the thermal control circuit. As mentioned earlier, the processor’s high thermal ramp rates make this infeasible. Refer to the thermal design guidelines listed in Section 1.2 for more details. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 113 Features The SMBus thermal sensor feature in the processor cannot be used to measure TCASE. The TCASE specification in Section 7 must be met regardless of the reading of the processor's thermal sensor in order to ensure adequate cooling for the entire processor. The SMBus thermal sensor feature is only available while VCC and SM_VCC are at valid levels and the processor is not in a low-power state. 8.4.5 Thermal Sensor Supported SMBus Transactions The thermal sensor responds to five of the SMBus packet types: Write Byte, Read Byte, Send Byte, Receive Byte, and Alert Response Address (ARA). The Send Byte packet can be used for sending one-shot commands. The Receive Byte packet accesses the register commanded by the last Read Byte packet and can be used to continuously read from a register. If a Receive Byte packet was preceded by a Write Byte or send Byte packet more recently than a Read Byte packet, then the behavior is undefined. Table 8-5 through Table 8-9 diagram the five packet types. In these figures, ‘S’ represents the SMBus start bit, ‘P’ represents a stop bit, ‘Ack’ represents an acknowledge, and ‘///’ represents a negative acknowledge (NACK). The shaded bits are transmitted by the thermal sensor, and the bits that aren’t shaded are transmitted by the SMBus host controller. Table 8-10 shows the encoding of the command byte. Table 8-5. Write Byte SMBus Packet S Slave Address Write Ack Command Code Ack Data Ack P 1 7-bits 0 1 8-bits 1 8-bits 1 1 Table 8-6. Read Byte SMBus Packet S Slave Address Write Ack Command Code Ack S Slave Address Read Ack Data /// P 1 7-bits 0 1 8-bits 1 1 7-bits 1 1 8bits 1 1 Table 8-7. Send Byte SMBus Packet S Slave Address Write Ack Command Code Ack P 1 7-bits 0 1 8-bits 1 1 Table 8-8. Receive Byte SMBus Packet S Slave Address Read Ack Data /// P 1 7-bits 1 1 8-bits 1 1 Table 8-9. ARA SMBus Packet S ARA Read Ack Address /// P 1 0001 100 1 1 Device Address1 1 1 NOTE: 1. This is an 8-bit field. The device which sent the alert will respond to the ARA Packet with its address in the seven most significant bits. The least significant bit is undefined and may return as a ‘1’ or ‘0’. See Section 8.4.8 for details on the Thermal Sensor Device addressing. 114 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Features 2. The shaded bits are transmitted by the thermal sensor, and the bits that aren’t shaded are transmitted by the SMBus host controller. Table 8-10. SMBus Thermal Sensor Command Byte Bit Assignments Register Command Reset State RESERVED 00h RESERVED TRR 01h 0000 0000 Function Reserved for future use Read processor core thermal diode RS 02h N/A RC 03h 00XX XXXX Read status byte (flags, busy signal) RCR 04h 0000 0010 RESERVED 05h RESERVED Reserved for future use RESERVED 06h RESERVED Reserved for future use RRHL 07h 0111 1111 Read processor core thermal diode THIGH limit RRLL 08h 1100 1001 Read processor core thermal diode TLOW limit WC 09h N/A Write configuration byte WCR 0Ah N/A Write conversion rate byte RESERVED 0Bh RESERVED Reserved for future use RESERVED 0Ch RESERVED Reserved for future use WRHL 0Dh N/A Write processor core thermal diode THIGH limit WRLL 0Eh N/A Write processor core thermal diode TLOW limit OSHT 0Fh N/A One shot command (use send byte packet) RESERVED 10h – FFh N/A Reserved for future use Read configuration byte Read conversion rate byte All of the commands in Table 8-10 are for reading or writing registers in the SMBus thermal sensor, except the one-shot command (OSHT) register. The one-shot command forces the immediate start of a new conversion cycle. If a conversion is in progress when the one-shot command is received, then the command is ignored. If the thermal sensor is in stand-by mode when the one-shot command is received, a conversion is performed and the sensor returns to stand-by mode. The one-shot command is not supported when the thermal sensor is in auto-convert mode. Note: Writing to a read-command register or reading from a write-command register will produce invalid results. The default command after reset is to a reserved value (00h). After reset, Receive Byte SMBus packets will return invalid data until another command is sent to the thermal sensor. 8.4.6 SMBus Thermal Sensor Registers 8.4.6.1 Thermal Value Registers Once the SMBus thermal sensor reads the processor thermal diode, it performs an analog to digital conversion and stores the results in the Thermal Reference Register (TRR). The supported range is +127 to 0 decimal and is expressed as an eight-bit number representing temperature in degrees 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 115 Features Celsius. This eight-bit value consists of seven bits of data and a sign bit (MSB) where the sign is always positive (sign = 0) and is shown in Table 8-11. The values shown are also used to program the Thermal Limit Registers. The values of these registers should be treated as saturating values. Values above 127 are represented at 127 decimal, and values of zero and below may be represented as 0 to -127 decimal. If the device returns a value where the sign bit is set (1) and the data is 000_0000 through 111_1110, the temperature should be interpreted as 0° Celsius. Table 8-11. Thermal Value Register Encoding Temperature (°C) 8.4.6.2 Register Value (binary) +127 0 111 1111 +126 0 111 1110 +100 0 110 0100 +50 0 011 0010 +25 0 001 1001 +1 0 000 0001 0 0 000 0000 Thermal Limit Registers The SMBus thermal sensor has four Thermal Limit Registers: • • • • RRHL — used to read the high limit RRLL — used to read for the low limit WRHL — used to write the high limit WRLL — used to write the low limit These registers allow the user to define high and low limits for the processor core thermal diode reading. The encoding for these registers is the same as for the TRR shown in Table 8-11. If the processor thermal diode reading equals or exceeds one of these limits, then the alarm bit (RHIGH or RLOW) in the Thermal Sensor Status Register is triggered. 8.4.6.3 Status Register The Status Register shown in Table 8-12 indicates which, if any, thermal value thresholds for the processor core thermal diode have been exceeded. It also indicates whether a conversion is in progress or an open circuit has been detected in the processor core thermal diode connection. Once set, alarm bits stay set until they are cleared by a Status Register read. A successful read to the Status Register will clear any alarm bits that may have been set (unless the alarm condition persists). If the SM_ALERT# signal is enabled via the Thermal Sensor Configuration Register and a thermal diode threshold is exceeded, an alert will be sent to the platform via the SM_ALERT# signal. This register is read by accessing the RS Command Register. 116 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Features Table 8-12. SMBus Thermal Sensor Status Register 8.4.6.4 Bit Name Reset State Function 7 (MSB) BUSY N/A If set, indicates that the device’s analog to digital converter is busy. 6 RESERVED RESERVED Reserved for future use 5 RESERVED RESERVED Reserved for future use 4 RHIGH 0 If set, indicates the processor core thermal diode high temperature alarm has activated. 3 RLOW 0 If set, indicates the processor core thermal diode low temperature alarm has activated. 2 OPEN 0 If set, indicates an open fault in the connection to the processor core diode. 1 RESERVED RESERVED Reserved for future use. 0 (LSB) RESERVED RESERVED Reserved for future use. Configuration Register The Configuration Register controls the operating mode (stand-by vs. auto-convert) of the SMBus thermal sensor. Table 8-13 shows the format of the Configuration Register. If the RUN/STOP bit is set (high) then the thermal sensor immediately stops converting and enters stand-by mode. The thermal sensor will still perform analog to digital conversions in stand-by mode when it receives a one-shot command. If the RUN/STOP bit is clear (low), then the thermal sensor enters autoconversion mode. This register is accessed by using the thermal sensor Command Register. The RC command register is used for read commands and the WC command register is used for write commands. See Table 8-10. Table 8-13. SMBus Thermal Sensor Configuration Register Bit 7 (MSB) Name MASK Reset State Function 0 Mask SM_ALERT# bit. Clear the bit to allow interrupts via SM_ALERT# and allow the thermal sensor to respond to the ARA command when an alarm is active. Set the bit to disable interrupt mode. The bit is not used to clear the state of the SM_ALERT# output. An ARA command may not be recognized if the mask is enabled. Stand-by mode control bit. If set, the device immediately stops converting, and enters standby mode. If cleared, the device converts in either one-shot mode or automatically updates on a timed basis. 6 RUN/STOP 0 5:0 RESERVED RESERVED Reserved for future use. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 117 Features 8.4.6.5 Conversion Rate Registers The contents of the Conversion Rate Registers determine the nominal rate at which analog-todigital conversions happen when the SMBus thermal sensor is in auto-convert mode. There are two Conversion Rate Registers: RCR for reading the conversion rate value; and WCR for writing the value. Table 8-14 shows the mapping between Conversion Rate Register values and the conversion rate. As indicated in Table 8-10, the Conversion Rate Register is set to its default state of 02h (0.25 Hz nominally) when the thermal sensor is powered up. There is a ±30% error tolerance between the conversion rate indicated in the conversion rate register and the actual conversion rate. Table 8-14. SMBus Thermal Sensor Conversion Rate Registers 8.4.7 Register Value Conversion Rate (Hz) 00h 0.0625 01h 0.125 02h 0.25 03h 0.5 04h 1.0 05h 2.0 06h 4.0 07h 8.0 08h to FFh Reserved for future use SMBus Thermal Sensor Alert Interrupt The SMBus thermal sensor located on the processor includes the ability to interrupt the SMBus when a fault condition exists. The fault conditions consist of: 1. a processor thermal diode value measurement that exceeds a user-defined high or low threshold programmed into the Command Register; or 2. disconnection of the processor thermal diode from the thermal sensor. The interrupt can be enabled and disabled via the thermal sensor Configuration Register and is delivered to the system board via the SM_ALERT# open drain output. Once latched, the SM_ALERT# should only be cleared by reading the Alert Response byte from the Alert Response Address of the thermal sensor. The Alert Response Address is a special slave address shown in Table 8-9. The SM_ALERT# will be cleared once the SMBus master device reads the slave ARA unless the fault condition persists. Reading the Status Register or setting the mask bit within the Configuration Register does not clear the interrupt. 8.4.8 SMBus Device Addressing Of the addresses broadcast across the SMBus, the memory component claims those of the form “1010XXXZb”. The “XXX” bits are defined by pull-up and pull-down resistors on the system baseboard. These address pins are pulled down weakly (10 kW) on the processor substrate to ensure that the memory components are in a known state in systems which do not support the SMBus (or only support a partial implementation). The “Z” bit is the read/write bit for the serial bus transaction. 118 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Features The thermal sensor internally decodes one of three upper address patterns from the bus of the form “0011XXXZb”, “1001XXXZb”, or “0101XXXZb”. The device’s addressing, as implemented, uses the SM_TS_A[1:0] pins in either the HI, LO, or Hi-Z state. Therefore, the thermal sensor supports nine unique addresses. To set either pin for the Hi-Z state, the pin must be left floating. As before, the “Z” bit is the read/write bit for the serial transaction. Note that addresses of the form “0000XXXXb” are Reserved and should not be generated by an SMBus master. The thermal sensor samples and latches the SM_TS_A[1:0] signals at power-up and at the starting point of every conversion. System designers should ensure that these signals are at valid VIH, VIL, or floating input levels prior to or while the thermal sensor’s SM_VCC supply powers up. This should be done by pulling the pins to SM_VCC or VSS via a 1 kW or smaller resistor, or leaving the pins floating to achieve the Hi-Z state. If the system designer wants to drive the SM_TS_A[1:0] pins with logic, the designer must still ensure that the pins are at valid input levels prior to or while the SM_VCC supply ramps up. The system designer must also ensure that their particular implementation does not add excessive capacitance to the address inputs. Excess capacitance at the address inputs may cause address recognition problems. Figure 8-2 shows a logical diagram of the pin connections. Table 8-15 and Table 8-16 describe the address pin connections and how they affect the addressing of the devices. Table 8-15. Thermal Sensor SMBus Addressing Address (Hex) Upper Address1 Device Select 8-bit Address Word on Serial Bus SM_TS_A1 SM_TS_A0 b[7:0] 3Xh 0011 0 Z2 1 0 0 0 0011000Xb 0011001Xb 0011010Xb 5Xh 0101 0 Z2 1 Z2 Z2 Z2 0101001Xb 0101010Xb 0101011Xb 9Xh 1001 0 Z2 1 1 1 1 1001100Xb 1001101Xb 1001110Xb NOTES: 1. Upper address bits are decoded in conjunction with the device select pins. 2. A tri-state or “Z” state on this pin is achieved by leaving this pin unconnected. Note: System management software must be aware of the processor dependent addresses for the thermal sensor. Table 8-16. Memory Device SMBus Addressing (Sheet 1 of 2) Address (Hex) Upper Address1 Device Select R/W bits 7-4 SM_EP_A2 bit 3 SM_EP_A1 bit 2 SM_EP_A0 bit 1 bit 0 A0h/A1h 1010 0 0 0 X A2h/A3h 1010 0 0 1 X A4h/A5h 1010 0 1 0 X 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 119 Features Table 8-16. Memory Device SMBus Addressing (Sheet 2 of 2) Address (Hex) Upper Address1 Device Select R/W bits 7-4 SM_EP_A2 bit 3 SM_EP_A1 bit 2 SM_EP_A0 bit 1 bit 0 1010 0 1 1 X A6h/A7h A8h/A9h 1010 1 0 0 X AAh/ABh 1010 1 0 1 X ACh/ADh 1010 1 1 0 X AEh/AFh 1010 1 1 1 X NOTE: 1. This addressing scheme will support up to 8 processors on a single SMBus. 8.4.9 Managing Data in the PIROM The PIROM consists of the following sections: • • • • • • • • • Header Processor Data Processor Core Data Cache Data Package Data Part Number Data Thermal Reference Data Feature Data Other Data Details on each of these sections are described below. Note: 8.4.9.1 Reserved fields or bits SHOULD be programmed to zeros. However, OEMs should not rely on this model. Header To maintain backward compatibility, the Header defines the starting address for each subsequent section of the PIROM. Software should check for the offset before reading data from a particular section of the ROM. Example: Code looking for the cache data of a processor would read offset 05h to find a value of 25h. 25h is the first address within the 'Cache Data' section of the PIROM. The Header also includes the data format revision at offset 0h and the EEPROM size (formatted in hex bytes) at offset 01-02h. The data format revision is used whenever fields within the PIROM are redefined. Normally the revision would begin at a value of 1. If a field, or bit assignment within a field, is changed such that software needs to discern between the old and new definition, then the data format revision field should be incremented. 120 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Features The EEPROM size provides the size of the PIROM in hex bytes. The PIROM is 128 bytes; thus, offset 01 - 02h would be programmed to 80h. 8.4.9.2 Processor Data This section contains two pieces of data: • The S-spec/QDF of the part in ASCII format • (1) 2-bit field to declare if the part is a pre-production sample or a production unit The S-spec/QDF field is six ASCII characters wide and is programmed with the same S-spec/QDF value as marked on the processor. If the value is less than six characters in length, leading spaces (20h) are programmed in this field. Example: A processor with a QDF mark of QEU5 contains the following in field 0E-13h: 20, 20, 51, 45, 55, 35h. This data consists of two blanks at 0Eh and 0Fh followed by the ASCII codes for QEU5 in locations 10 - 13h. Offset 14h contains the sample/production field, which is a two-bit field and is LSB aligned. All Qspec material will use a value of 00b. All S-spec material will use a value of 01b. All other values are reserved. Example: A processor with a Qxxx mark (engineering sample) will have offset 14h set to 00b. A processor with an Sxxxx mark (production unit) will use 01b at offset 14h. 8.4.9.3 Processor Core Data This section contains core silicon-related data. 8.4.9.3.1 CPUID The CPUID field is a copy of the results in EAX[13:0] from Function 1 of the CPUID instruction. Note: The field is not aligned on a byte boundary since the first two bits of the offset are reserved. Thus, the data must be shifted right by two in order to obtain the same results. Example: The CPUID of a C-0 stepping 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache is 0F41h. The value programmed into offset 16 - 17h of the PIROM is 3D04h. Note: The first two bits of the PIROM are reserved, as highlighted in the example below. CPUID instruction results 0000 1111 0100 0001 (0F41h) PIROM content 0011 1101 0000 0100 (3D04h) 8.4.9.3.2 Front Side Bus Frequency Offset 1A - 1Bh provides front side bus frequency information. Systems may need to read this offset to decide if all installed processors support the same front side bus speed. Because the Intel NetBurst microarchitecture bus is described as a 4X data bus, the frequency given in this field is currently 667 MHz. The data provided is the speed, rounded to a whole number, and reflected in hex. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 121 Features Example: The processor supports a 667 MHz front side bus. Therefore, offset 1A - 1Bh has a value of 029Bh. 8.4.9.3.3 Multi-Processor Support Offset 1Ch has 2 bits defined for representing the supported number of physical processors on the bus. These two bits are MSB aligned where 00b equates to single-processor operation, 01b is a dual-processor operation, and 11b represents multi-processor operation. Normally, only values of 01 and 11b are used. The remaining six bits in this field are reserved for the future use. 8.4.9.3.4 Maximum Core Frequency Offset 1D - 1Eh provides the maximum core frequency for the processor. The frequency should equate to the markings on the processor and/or the QDF/S-spec speed even if the parts are not limited or locked to the intended speed. Format of this field is in MHz, rounded to a whole number, and encoded in hex format. Example: A 3.00 GHz processor will have a value of 0BB8h, which equates to 3000 decimal. 8.4.9.3.5 Core Voltage There are two areas defined in the PIROM for the core voltages associated with the processor. Offset 1F - 20h is the Processor Core VID (Voltage Identification) field and contains the voltage requested via the VID pins. In the case of the 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache, this is 1.3875 V. This field, rounded to the next thousandth, is in mV and is reflected in hex. This data is also in Table 2-9. Some systems read this offset to determine if all processors support the same default VID setting. Minimum core voltage is reflected in offset 21 - 22h. This field is in mV and reflected in hex. The minimum VCC reflected in this field is the minimum allowable voltage assuming the FMB maximum current draw. Note: The minimum core voltage value in offset 21 - 22h is a single value that assumes the FMB maximum current draw. Refer to Table 2-10 for the minimum core voltage specifications based on actual real-time current draw. Example: The specifications for a 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache at FMB are 1.3875 V VID and 1.229 V minimum voltage. Offset 1F - 20h would contain 056Ch (1388 decimal) and offset 21 - 22h would contain 04D2h (1234 decimal). 8.4.9.3.6 TCASE Maximum The last field within Processor Core Data is the TCASE Maximum field. The field reflects temperature in degrees Celsius in hex format. This data can be found in the Table 7-1. The thermal specifications are specified at the case (IHS). 8.4.9.4 Cache Data This section contains cache-related data. 122 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Features 8.4.9.4.7 L2/L3 Cache Size Offset 27 - 28h is the L2 cache size field. The field reflects the size of the level two cache in kilobytes. Offset 29 - 3Ah is the L3 cache size field and also reflects size in kilobytes. Both fields are in hex format. Example: The 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache has a 1 MB (1024 KB) L2 cache and either 4 MB (4096 KB) or 8 MB (8192 KB) L3 cache. Thus, offset 27 28h will contain 0400h, and offset 29 - 2Ah will contain 1000h (for 4 MB) or 2000h (for 8 MB). 8.4.9.4.8 Cache Voltage There are two areas defined in the PIROM for the L3 cache voltages associated with the processor. Offset 2B - 2Ch is the Processor Cache VID (Cache Voltage Identification), or CVID, field and contains the voltage requested via the CVID pins. In the case of the 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache, this is 1.275 V. This field is in mV and is reflected in hex. This data is also in Table 2-9. Some systems read this offset to determine if all processors support the same default CVID setting. Minimum L3 cache voltage specifications are reflected in offset 2D - 2Eh. This field is in mV and reflected in hex. This data is also in Table 2-9. For processors that follow a load line DC specification, the minimum VCACHE reflected in this field should reflect the minimum allowable voltage at maximum current. Example: The specifications for a 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache are 1.275 V CVID and 1.125 V minimum voltage (at maximum current). Offset 2B - 2Ch would contain 04FBh (1275 decimal) and offset 2D - 2Eh would contain 0465h (1125 decimal). 8.4.9.5 Package Data This section describes the package revision location at offset 32 - 35h. This field tracks the highest level revision. It is provided in ASCII format of four characters (8 bits x 4 characters = 32 bits). The package is documented as 1.0, 2.0, etc. Because this only consumes three ASCII characters, a leading space is provided in the data field. Example: The C-0 stepping of the 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache utilizes revision 1.0 of the FC-mPGA package. Thus, at offset 32-35h, the data is a space followed by 1.0. In hex, this would be 20, 31, 2E, 30. 8.4.9.6 Part Number Data This section provides traceability. There are 208 available bytes in this section for future use. 8.4.9.6.9 Processor Part Number Offset 38 - 3Eh contains seven ASCII characters reflecting the Intel part number for the processor. This information is typically marked on the outside of the processor. If the part number is less than 7 characters, a leading space is inserted into the value. The part number should match the information found in the marking specification found in Section 4. Example: A processor with a part number of 80546KF will have data found at offset 38 - 3Eh is 38, 30, 35, 34, 36, 4B, 46. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 123 Features 8.4.9.6.10 Processor Electronic Signature Offset 4D - 54h contains a 64-bit identification number. Intel does not guarantee that each processor will have a unique value in this field. 8.4.9.7 Feature Data This section provides information on key features that the platform may need to understand without powering on the processor. 8.4.9.7.11 Processor Core Feature Flags Offset 74 - 77h contains a copy of results in EDX[31:0] from Function 1 of the CPUID instruction. These details provide instruction and feature support by product family. A decode of these bits is found in the AP-485 Intel® Processor Identification and CPUID Instruction application note. 8.4.9.7.12 Processor Feature Flags Offset 78h provides additional feature information from the processor. This field is defined as follows: Table 8-17. Offset 78h Definitions Bit Definition 7 Reserved 6 Serial signature (set if there is a serial signature at offset 4D - 54h) 5 Electronic signature present (set if there is a electronic signature at 4D - 54h) 4 Thermal Sense Device present (set if an SMBus thermal sensor on package) 3 Reserved 2 OEM EEPROM present (set if there is a scratch ROM at offset 80 - FFh) 1 Core VID present (set if there is a VID provided by the processor) 0 L3 Cache present (set if there is a level 3 cache on the processor) 8.4.9.7.13 Additional Processor Feature Flags All bits of this field are reserved at this time. The field resides at offset 79 - 7Bh. 8.4.9.8 Other Data Addresses 7E - 7F are listed as reserved. 124 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Features 8.4.9.9 Checksums The PIROM includes multiple checksums. Table 8-18 includes the checksum values for each section defined in the 128 byte ROM, except Other Data. Table 8-18. 128 Byte ROM Checksum Values Section Checksum Address Header 0Dh Processor Data 15h Processor Core Data 24h Cache Data 31h Package Data 37h Part Number Data 6Fh Feature Data 7Dh Other Data None Defined Checksums are automatically calculated and programmed by Intel. The first step in calculating the checksum is to add each byte from the field to the next subsequent byte. This result is then negated to provide the checksum. Example: For a byte string of AA445Ch, the resulting checksum will be B6h. AA = 10101010 44 = 01000100 5C = 0101100 AA + 44 + 5C = 01001010 Negate the sum: 10110101 +1 = 101101 (B6h) § 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 125 Features 126 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 9 Boxed Processor Specifications 9.1 Introduction The 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache processor may also be offered as an Intel boxed processor. Intel boxed processors are intended for system integrators who build systems from components available through distribution channels. The boxed thermal solution is still under development and subject to change. This section is meant to provide some insight into the current direction of the thermal solution. Future revisions may have solutions that differ from those discussed here. The current thermal solution plan for the boxed 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache processor is to include an unattached passive heatsink. This solution is currently targeted at chassis which are 3U and above in height. This section documents baseboard and platform requirements for the thermal solution, supplied with the boxed 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache processor. This section is particularly important to companies that design and manufacture baseboards, chassis and complete systems. Figure 9-1 shows the conceptual drawing of the boxed processor thermal solution. Drawings in this section reflect only the specifications on the Intel boxed processor product. These dimensions should not be used as a generic keep-out zone for all cooling solutions. It is the system designer’s responsibility to consider their proprietary cooling solution when designing to the required keep-out zone on their system platform and chassis. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 127 Boxed Processor Specifications Figure 9-1. Passive Processor Thermal Solution (3U and larger) NOTE: 1. The heatsink in this image is for reference only. 2. This drawing shows the retention scheme for the boxed processor. 9.2 Mechanical Specifications This section documents the mechanical specifications of the boxed processor passive heatsink. 9.2.1 Boxed Processor Heatsink Dimensions The boxed processor is shipped with an unattached passive heatsink. Clearance is required around the heatsink to ensure unimpeded airflow for proper cooling. The physical space requirements and dimensions for the boxed processor and assembled heatsink are shown in the following figures. 128 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Boxed Processor Specifications Figure 9-2. Top Side Board Keep-Out Zones (Part 1) 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 129 Boxed Processor Specifications Figure 9-3. Top Side Board Keep-Out Zones (Part 2) 130 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Boxed Processor Specifications Figure 9-4. Bottom Side Board Keep-Out Zones 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 131 Boxed Processor Specifications Figure 9-5. Board Mounting-Hole Keep-Out Zones 132 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Boxed Processor Specifications Figure 9-6. Thermal Solution Volumetric 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 133 Boxed Processor Specifications Figure 9-7. Recommended Processor Layout and Pitch 134 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet Boxed Processor Specifications 9.2.2 Boxed Processor Heatsink Weight The boxed processor heatsink weight is approximately 530 grams. See Section 4 of this document for details on the processor weight. 9.2.3 Boxed Processor Retention Mechanism and Heatsink Supports Baseboards and chassis’s designed for use by system integrators should include holes that are in proper alignment with each other to support the boxed processor. See Figure 9-7 for example of processor pitch and layout. Figure 9-1 illustrates the new retention solution. This is designed to extend air-cooling capability through the use of larger heatsinks with minimal airflow blockage and minimal bypass. These retention mechanisms can allow the use of much heavier heatsink masses compared to legacy solution limitations by using a load path attached to the chassis pan. The cooling solution spring (Figure 9-1 labeled as “CEK SPRING”) on the under side of the baseboard provides the necessary compressive load for the thermal interface material. The baseboard is intended to be isolated such that the dynamic loads from the heatsink are transferred to the chassis pan via the heatsink screws and heatsink standoffs. This reduces the risk of package pullout and solder joint failures in a shock and vibe situation. The assembly requires larger diameter holes to compensate for the cooling solution spring embosses. See Figure 9-2 and Figure 9-3 for processor mounting through holes. 9.3 Thermal Specifications This section describes the cooling requirements of the heatsink solution utilized by the boxed processor. 9.3.1 Boxed Processor Cooling Requirements The boxed processor will be cooled by forcing ducted chassis fan airflow through the passive heat sink solution. Meeting the processor’s temperature specifications is a function of the thermal design of the entire system, and ultimately the responsibility of the system integrator. The processor temperature specification is found in Section 7 of this document. For the boxed processor passive heatsink to operate properly, chassis air movement devices are required. Necessary airflow and associated flow impedance is 29 cfm at 0.10” H2O. In addition, the processor pitch should be 3.25 inches, or slightly more, when placed in side by side orientation. Figure 9-7 illustrates the side by side orientation and pitch. Note that the heatsinks are interleaved to reduce air bypass. It is also recommended that the ambient air temperature outside of the chassis be kept at or below 35°C. The air passing directly over the processor heatsink should not be preheated by other system components (such as another processor), and should be kept at or below 40°C. Again, meeting the processor’s temperature specification is the responsibility of the system integrator. 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 135 Boxed Processor Specifications 9.3.2 Boxed Processor Contents The boxed processor will include the following items: • • • • 64-bit Intel® Xeon™ processor MP with up to 8MB L3 cache processor Unattached Passive Heatsink with captive screws Thermal Interface Material (pre-attached) Warranty / Installation manual with Intel Inside logo The other items listed in Figure 9-1, required with this thermal solution should be shipped with either the chassis or the mainboard. They include: • Cooling Solution Spring (typically included with mainboard) • Chassis Standoffs • System fans § 136 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 10 Debug Tools Specifications 10.1 Logic Analyzer Interface (LAI) Intel is working with two logic analyzer vendors to provide logic analyzer interfaces (LAIs) for use in debugging systems. Tektronix and Agilent should be contacted to get specific information about their logic analyzer interfaces. The following information is general in nature. Specific information must be obtained from the logic analyzer vendor. Due to the complexity of multiprocessor systems, the LAI is critical in providing the ability to probe and capture front side bus signals. There are two sets of considerations to keep in mind when designing systems that can make use of an LAI: mechanical and electrical. 10.1.1 Mechanical Considerations The LAI is installed between the processor socket and the processor. The LAI pins plug into the socket, while the processor pins plug into a socket on the LAI. Cabling that is part of the LAI egresses the system to allow an electrical connection between the processor and a logic analyzer. The maximum volume occupied by the LAI, known as the keepout volume, as well as the cable egress restrictions, should be obtained from the logic analyzer vendor. System designers must make sure that the keepout volume remains unobstructed inside the system. Note that it is possible that the keepout volume reserved for the LAI may differ from the space normally occupied by the processor heatsink. If this is the case, the logic analyzer vendor will provide a cooling solution as part of the LAI. 10.1.2 Electrical Considerations The LAI will also affect the electrical performance of the front side bus; therefore, it is critical to obtain electrical load models from each of the logic analyzer vendors to be able to run system level simulations to prove that their tool will work in the system. Contact the logic analyzer vendor for electrical specifications and load models for the LAI solution they provide. § 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet 137 Debug Tools Specifications 138 64-bit Intel® Xeon™ Processor MP with up to 8MB L3 Cache Datasheet