Application Brief Data Handling, Throughput, and Image Processing: Key Features for Machine Vision Designs AMD Solutions for Machine Vision Applications AMD Embedded R-Series APUs (Accelerated Processing Units) and G-Series SOCs (Systemson-Chip) are helping enable another leap forward for machine vision technology. The silicon-level integration of a low-power x86 CPU and the parallel processing performance of a programmable, discrete-class general-purpose graphics processing unit (GPGPU) in a single device drives the high-speed processing that is essential for achieving high-performance machine vision. Combining a GPU core on the same die as the CPU enables the system to offload computation-intensive pixel data processing from the CPU to the multi-core GPU, distributing the processing workload across available processor cores in parallel to help improve the real-time performance of the whole system. This can yield an order of magnitude increase in image processing performance versus serial task execution on a CPU alone and offers the simplified hardware architecture, data handling, and connectivity of a standard PC platform. While traditional benefits of these solutions include long-term availability and low power consumption, applications built around the AMD G-Series SOCs and R-Series APUs also offer software compatibility support that can broadly scale across machine vision solutions. AMD Embedded Solutions Feature a Variety of Attributes for Machine Vision Applications Machine vision technology is evolving quickly, fueled by dramatic gains in processing performance through innovative heterogeneous architectures, which accelerate image processing functions and handle data transfer and I/O respectively. The relatively recent arrival of PC-based ”smart cameras” that forego conventional DSP and FPGA-based processing platforms heralds another significant advance in intelligent vision system technology as the industry shifts away from specialized legacy processors and narrowly supported imaging software in favor of the more versatile x86 platform. • PC-compatible solutions backed by an ecosystem of software and tools to help deliver a full solution, easily integrate with enterprise networks, and shorten design cycles – Software compatibility between smart cameras and machine vision PCs – Broad ecosystem support to help simplify development – Support for Microsoft® Windows® and Linux® operating systems –E asy networking with IP-based industrial and enterprise networks • Range of price, power, and performance: APU and SOC solutions for lowpower smart cameras or high-performance PC-based designs – Ability for same design to scale from smart camera to compact vision system – Support for multiple product categories with a single software architecture • Excellent performance on highly integrated and low-power APU and SOC solutions to help simplify design into small form factor and fanless applications – Low power and high integration, well suited for compact smart camera designs – Combination of an image processing engine, full complement of I/O, and excellent data management capabilities on a single device – Support for 2-4 USB 3.0 ports and accessible interface to multiple-Gb Ethernet ports through PCI Express® • OpenCL™ acceleration delivers high compute performance on the GPU – From 46 to 563 SP GFLOPS1, 2 compute performance – Royalty-free open standard – Availability of more than 70 open source functions for accelerated machine vision applications3 • Support for remote management across families – Increased system reliability and reduced downtime – AMD DAS 1.0 with out-of-band remote management solution support, to remotely manage devices even when the OS is not running Application Brief: Machine Vision AMD Embedded Solutions for Machine Vision Applications FOR SMART CAMERAS FOR MACHINE VISION SERVERS OR COMPACT IMAGING SYSTEMS • Recommended: AMD GX-210HA SOC - Low-power at only 4.1W average power4 - Highly Integrated Dual-core SOC - Good image processing performance - H igher performance alternative: AMD GX-415GA SOC • Recommended: AMD R-464L APU - Excellent graphics performance - High x86 performance and data throughput - 563 SP GFLOPS5 image processing performance - L ower power/cost scalable alternative: AMD GX-420CA SOC (10.8W Average Power)5 CPLD PCIe® or USB 3.0 CPU 2 CPU 3 CPU 1 PC-based Smart Camera with AMD G-Series SOC CPU 0 SDIO/Card HDMI™/DP PCIe®/Display Memory (Interpolation and color correction/deserializer) Direct® 11 GPU Image Sensor MIPI or proprietary (Security) AMD G-Series SOC SDIO/Card SATA v3.x (Storage) Boot Flash 12-48VDC HD Audio PCIe® IR USB 2.0/3.0 LPC, SPI GbE GbE (Opt) Optical Isolation 8X DI/8X DO eSATA 8X USB 2.0 2X USB 3.0 UART RS-232/485 Or PoE Remote Management For more information, or to discuss which AMD-based solution might be right for you, please contact your local AMD Embedded Solutions sales representative: www.amd.com/embeddedsales or see www.amd.com/industrial 1 Calculated SP GFLOPS = (# of x86 cores x (128-bit (FPUs) / 32-bit (SP Operation)) * CPU Base Frequency) + (# of shader units * (64-bit (shader) / 32-bit (SP Operation)) * GPU Max Frequency). EMB-21 2 The calculated single precision GFLOPs for the AMD G-T16R is 46 and for the Intel Atom D2700 is 8.52. Calculated SP GFLOPs = (# of x86 cores x (# of FP bits per core / 32 bit (SP Operation)) * CPU Frequency) + (# of OpenCL enabled GPU cores * (# of FP bits per GPU core / 32 bit (SP Operation)) * GPU Frequency). EMB-31 3 Accelerated functions in OpenCV 2.4.3 (updated in 2.4.4) downloadable under free BSD license from http://opencv.org. 4 The average power for the AMD GX-210HA SOC is 4.1 watts, determined by averaging the results of the measured average power of the SOC running the following benchmarks; 3DMark® 11, AMD Sys Stress Test CPU, AMD Sys Stress Test CPU & GPU, AMD Sys Stress Test GPU, WinBench® 99, CoreMark, Game: Meat Boy, PCMark® 7, POV-Ray, Sandra 2011, Game: Street Fighter. Testing was performed on an AMD E1-2100 (Rev A1) that is equivalent to the AMD GX-210HA SOC. System configuration: AMD E1-2100 @ 70°C, “Larne" development platform, 4GB RAM, Windows® 7 Ultimate. Please see AMD publication ID 53395A for more information. EMB-41 5 The average power for the AMD GX-420CASOC is 10.8 Watts, determined by averaging the results of the measured average power of the SOC running the following benchmarks; 3DMark® 11, AMD Sys Stress Test CPU, AMD Sys Stress Test CPU & GPU, AMD Sys Stress Test GPU, Winbench® 99, CoreMark(Multi-thread), Game:Meat Boy, PCMark® 7, POV-Ray(all), Sandra 2011, Game: Street Fighter. System configuration: AMD GX-420CA @ 60°C, “Larne" development platform, 4GB RAM, Windows 7 Ultimate. Please see AMD Publication ID 53395B for more information. DISCLAIMER The information contained herein is for informational purposes only, and is subject to change without notice. While every precaution has been taken in the preparation of this document, it may contain technical inaccuracies, omissions and typographical errors, and AMD is under no obligation to update or otherwise correct this information. Advanced Micro Devices, Inc. makes no representations or warranties with respect to the accuracy or completeness of the contents of this document, and assumes no liability of any kind, including the implied warranties of non-infringement, merchantability or fitness for particular purposes, with respect to the operation or use of AMD hardware, software or other products described herein. No license, including implied or arising by estoppel, to any intellectual property rights is granted by this document. Terms and limitations applicable to the purchase or use of AMD’s products are as set forth in a signed agreement between the parties or in AMD's Standard Terms and Conditions of Sale. ©2013 AMD, the AMD Arrow logo, and combinations thereof are trademarks of Advanced Micro Devices. 3DMark and PCMark are registered trademarks of Futuremark Corporation. Linux is a registered trademark of Linus Torvalds. Microsoft and Windows are registered trademarks of Microsoft Corporation. OpenCL is a trademark of Apple Inc. used by permission by Khronos. PCI Express is a registered trademark of PCI-SIG Corporation. WinBench is a registered trademark of Ziff Davis, Inc. Other names are for informational purposes only and may be trademarks of their respective owners. PID 54071-A