Adtech Global worked with AMD FirePro to develop a High Speed Cluster solution based on GPGPU technology to support efforts in quantum chromodynamics research for the King Abdualziz City for Science and Technology (KACST) and Frankfurt Institute for Advanced Studies (FIAS).
ACST and FIAS were striving to not only maximize performance, but also to optimize this performance for thermal dissipation and power draw efficiency. The challenge in building a supercomputer is no longer the money you’re willing to spend on processors. The challenge now lies within getting enough power and cooling to the supercomputer itself. The SANAM supercomputer was able to break these barriers. It is currently the world’s No. 52 most powerful supercomputer, and the world’s 2nd most power-efficient supercomputer.
Critical to the power efficiency was the use of (420) AMD FirePro S10000 dual-GPU server graphics cards. The SANAM supercomputer is capable of sustaining 421.2 TFLOPS, providing a system energy efficiency of over 2.3 GFLOPS per watt and performing 2,351 million calculations per second per watt.
The video below gives a really insightful tour of the collaboration between Adtech and AMD, as well as some clear visuals on the actual SANAM supercomputer. You can also read the case study which details the custom computing solution they implemented with the help of Adtech Global.
SPECapc today released their SolidWorks 2013 workstation benchmark. The benchmark includes 10 new models, each of which is exercised with 12 tests (10 graphics and 2 CPU). The models were selected from the SolidWorks user community, with an emphasis on real-world datasets that can be distributed in the public domain.
The 10 graphics-oriented tests in SPECapc for SolidWorks 2013 use three SolidWorks view settings - RealView, ambient occlusion (AO) and shadows - in combination with shaded and shaded-with-edges display styles.
The video below compares the AMD FirePro W7000 to the Nvidia Quadro K5000 using the SPECapc for SolidWorks 2013 benchmark. As you can see, the FirePro W7000 finished rendering the models across the RealView, AO, edge and shadows settings in 9:06, while the same content on the Quadro K5000 isn’t complete until 13:48. What I find important to note is that these are real-world datasets running real-world scenarios (not some synthetic or artificial benchmark).
Note: I posted this video last week, not realizing this was the new 2013 benchmark. Now that I know, this is the latest benchmark, I am even more excited.
This video compares rendering performance in SolidWorks 2013 using the AMD FirePro W7000 ($899 MSRP) vs the Nvidia Quadro K5000 ($2249 MSRP).
These are videos of identical scripted manipulations with real-world models under real-world rendering and lighting conditions using SPECapc for SolidWorks 2013 (designed to represent a day in the life of a typical SolidWorks user, i.e. not an artificial benchmark).
Cut to the chase: the FirePro W7000 is finished rendering the content at 9:06, while the same content on the Quadro K5000 isn’t complete until 13:48 – in other words, the FirePro W7000 is about 30% faster in SolidWorks 2013 rendering.
If you are a SolidWorks user, this comparison might be a bit of an eyeopener.
In addition to rendering performance, the FirePro card also benefits from PCI Express 3.0 offering less overhead and up to twice as much data throughput. Also the FirePro card can simultaneously handle rendering and compute (e.g. FEA/CAE). The Nvidia K series is PCI Express 2.0 and provides rendering only.
Specs for the test were as follows:
Hewlett-Packard HP Z600 Quad-Core Workstation running Win 7 Professional, w/ 8GB RAM
AMD Driver : 9.003.3
Nvidia Driver : 310.90
Screen Resolution : 1280 x 720
Benchmark: SPECapc for SolidWorks x64 SP0.0
Models Used : Audi R8 S.SLDASM , Digger.SLDASM, Supercar.SLDASM, Rally Car.SLDASM, BRERA_ASS.SLDASM, Tower Assembly 02.SLDASM
Updated March 13 to reference SPECapc for SolidWorks 2013.
Develop3D tests and compares the new AMD FirePro W5000, W7000 and W8000 along with the Nvidia Quadro K5000, specifically for CAD applications and price / performance. The cards were tested across three applications relevant to CAD-centric workflows.
Creo 2.0 - SPECapc benchmark which tests for wireframe, shaded and aggregates the two (composite). SolidWorks 2013 - tested with two different models — an engineering model with shading and edges, and a camera model with RealView, the real time rendering mode. 3ds Max 2011 - SPECapc benchmark, which features a wide variety of models to test for interactive graphics and GPU shaders.
Conclusion: The FirePro W8000 and the Quadro K5000 high-end card are overkill for most CAD-based workflows, with both the FirePro W5000 and W7000 offering excellent price/performance.
“The fact is that 3D performance in many CAD applications is still limited by the GHz of the CPU, which means high-end cards like the Quadro K5000 and FirePro W8000 are not able to fulfil their true potential.”
“There are glimpses of what more power can offer when a heavier load is placed on the GPU. The shaded (Creo), RealView (SolidWorks) and GPU shaders (3ds Max) tests all show a significant jump when moving from the FirePro W5000 to the W7000.”
Desktop Engineering has written up their first impressions of the new AMD FirePro W series: the mid-range W5000, the high-end W7000 and W8000, and the ultra-high-end W9000.
This line of boards are optimized for heterogeneous computing. With other graphics boards, if you want to use the GPU for computing, you have to hold off on graphics until computations were completed—or use two graphics cards, one for computing and one for graphics (e.g. Nvidia Tesla + Quadro). The GCN technology in the new FirePro W series enables execution of a graphics thread and up to two compute threads per clock cycle at the same time on the GPU.
Desktop Engineering ran the SPECviewperf 11 benchmarks. Although these are only synthetic benchmarks and not real world application use, the results are still interesting. Most noteworthy: the FirePro W5000 outpace the older, high-end V7900 and the new FirePro W7000 outperformed the V8800.
The Green500 List ranks of the world’s most energy-efficient supercomputers. This year, the AMD FirePro S10000 has taken the SANAM supercomputer to #2.
AMD worked with the University of Frankfurt’s Institute for Advanced Studies (FIAS) to research quantum chromodynamics. Powered by 420 AMD FirePro S10000 dual-GPU server graphics cards, the SANAM supercomputer can sustain 420 TFLOPS, providing a system energy efficiency of over 2.3 GFLOPS per watt and performing 2,351 million calculations per second per watt.
Overall, the performance of machines in the Green500 List has increased at a higher rate compared to power consumption. “That’s why the machines’ efficiencies are going up,” says Feng. “We are more performance for the same amount of power.” For machines based on commodity components—machines built with off-the-shelf components—coprocessors and GPUs are attributing a great deal to the efficiency gains. So much so that they are keeping pace and in the latest list even outpacing purpose built systems like IBM’s Blue Gene/Q.
Consuming a total of 3.6 kilowatts, the folks down in the basement lab at AMD have built the fastest multi-GPU compute server using eight of their flagship compute powerhouse boards, the new AMD FirePro S10000!
This 16 GPU (eight S10000s) Exxact Computing Server provides more than 8 TFLOPS of real world double precision computing performance. While these are early drivers, this still means you are still seeing around 70% efficiency of the theoretical peak double precision floating point performance of 11.84 TFLOPS (47.28 TFLOPS peak single precision performance!).
Here’s a look at the DGEMM result and a glimpse at what’s under the hood of the compute server:
Today both AMD and Nvidia announced their new flagship HPC server GPU solutions. AMD released the FirePro S10000 (pdf spec sheet), a dual-GPU card that combines both compute and 3D visualization on a single card. Nvdia announced their Tesla K20 line which target compute only (you need a Quadro card for the 3D visualization capabilities). Below is a chart comparing the specs and performance of the new cards, as well as comparisons to Nvidia's Tesla K10 and compute-specific M2090 cards
AMD FirePro S10000
Nvidia Tesla K10
Nvidia Tesla K20
Nvidia Tesla K20X
Nvidia Tesla M2090
Core Clock Speed
Peak Single Precision
Peak Double Precision
FirePro S10000 is a single card that delivers both compute and professional 3D graphics, (as opposed to needing a Quadro plus Tesla). The FirePro S10000 can handle 2 compute & 1 graphics operation simultaneously
FirePro S10000 is up to 7.8X as fast as the K10 for peak double precision and 1.3X as fast for peak single precision
FirePro S10000 has up to 1.5X the memory bandwidth as Tesla K10
FirePro S10000 can drive up to 5 high resolutions displays vs Tesla (oops you need a Quadro to drive any displays)
FireUser.com is a community resource for CAD, visualization, 3D, video and engineering professionals to learn about the latest acceleration and display technologies and news with a focus on the AMD FirePro workstation graphics line.