ºÝºÝߣ

ºÝºÝߣShare a Scribd company logo

NAMD Molecular Dynamics on GPU

?
?
Devang Sachdev
Devang Sachdev

1) GPU accelerated computing provides significant performance benefits over CPU-only systems, with NAMD simulations running faster on GPU systems in all tests. 2) GPU acceleration results in a large performance boost for a small additional price of the GPU hardware. 3) Energy usage is reduced by half when using GPU acceleration compared to CPU-only systems.

1 of 11
NAMD 2.9
Summary/Conclusions Benefits of GPU Accelerated Computing Faster than CPU only systems in all tests Large performance boost with small marginal price increase Energy usage cut in half GPUs scale very well within a node and over multiple nodes Tesla K20 GPU is our fastest and lowest power high performance GPU to date Try GPU accelerated NAMD for free ¨C www.nvidia.com/GPUTestDrive
Kepler - Our Fastest Family of GPUs Yet 4.50 ApoA1 Running NAMD version 2.9 4.00 4.00 The blue node contains Dual E5-2687W CPUs 3.57 (8 Cores per CPU). 3.45 3.50 The green nodes contain Dual E5-2687W CPUs (8 2.9x Cores per CPU) and either 1x NVIDIA M2090, 1x K10 3.00 or 1x K20 for the GPU Nanoseconds/Day 2.63 2.6x 2.50 2.5x 2.00 1.50 1.37 1.9x 1.00 0.50 0.00 1 CPU Node 1 CPU Node + 1 CPU Node + K10 1 CPU Node + K20 1 CPU Node + K20X Apolipoprotein A1 M2090 GPU speedup/throughput increased from 1.9x (with M2090) to 2.9x (with K20X) when compared to a CPU only node 3 NAMD Benchmark Report, Revision 2.0, dated Nov. 5, 2012
Run NAMD 2.5x Faster with GPUs 3 Running NAMD 2.9 with CUDA 4.0 ECC Off 2.7 2.6 The blue node contains 2x Intel E5-2687W CPUs 2.5 2.4 (8 Cores per CPU) Speedup Compared to CPU Only Each green node contains 2x Intel E5-2687W 2 CPUs (8 Cores per CPU) plus 1x NVIDIA K20 GPUs 1.5 1 0.5 0 CPU All Molecules ApoA1 F1-ATPase STMV Apolipoprotein A1 Gain 2.5x throughput/performance by adding just 1 GPU when compared to dual CPU performance 4 NAMD Benchmark Report, Revision 2.0, dated Nov. 5, 2012
Kepler ¨C Universally Faster 6 Running NAMD version 2.9 The CPU Only node contains Dual E5-2687W CPUs 5 (8 Cores per CPU). Speedup Compared to CPU Only 5.1x The Kepler nodes contain Dual E5-2687W CPUs (8 4 4.7x Cores per CPU) and 1 or two NVIDIA K10, K20, or K20X GPUs. 4.3x F1-ATPase 3 ApoA1 STMV 2.9x 2 2.6x 2.4x 1 0 CPU Only 1x K10 1x K20 1x K20X 2x K10 2x K20 2x K20X F1-ATPase | Kepler nodes use Dual CPUs | The Kepler GPUs accelerate all simulations, up to 5x Average acceleration printed in bars
Outstanding Strong Scaling with Multi-STMV Running NAMD version 2.9 Each blue XE6 CPU node contains 1x AMD 100 STMV on Hundreds of Nodes 1600 Opteron (16 Cores per CPU). 1.2 Fermi XK6 Each green XK6 CPU+GPU node contains 1x AMD 1600 Opteron (16 Cores per CPU) 1 and an additional 1x NVIDIA X2090 GPU. CPU XK6 2.7x Nanoseconds / Day 0.8 2.9x 0.6 0.4 0.2 3.6x 3.8x Concatenation of 100 0 Satellite Tobacco Mosaic Virus 32 64 128 256 512 640 768 # of Nodes Accelerate your science by 2.7-3.8x when compared to CPU-based supercomputers
Replace 3 Nodes with 1 2090 GPU Running NAMD version 2.9 Each blue node contains 2x Intel Xeon X5550 CPUs F1-ATPase (4 Cores, $1000 per CPU). 4 CPU Nodes 0.8 9000 0.74 The green node contains 2x Intel Xeon X5550 CPUs $8,000 1 CPU Node +8000 (4 Cores, $1000 per CPU) and 1x NVIDIA M2090 GPU 0.7 1x M2090 GPUs 0.63 ($2000 each) 7000 0.6 6000 0.5 5000 0.4 $4,000 4000 0.3 3000 0.2 2000 0.1 1000 0 0 F1-ATPase Nanoseconds/Day Cost Speedup of 1.2x for 50% the cost
K20 - Greener: Twice The Science Per Watt 1200000 Energy Used in Simulating 1 Nanosecond of ApoA1 Running NAMD version 2.9 1000000 Each blue node contains Dual E5-2687W CPUs (95W, 4 Cores per CPU). Each green node contains 2x Intel Xeon X5550 Energy Expended (kJ) 800000 CPUs (95W, 4 Cores per CPU) and 2x NVIDIA Lower is better K20 GPUs (225W per GPU) 600000 Energy Expended 400000 = Power x Time 200000 0 1 Node 1 Node + 2x K20 Cut down energy usage by ? with GPUs 8 NAMD Benchmark Report, Revision 2.0, dated Nov. 5, 2012
Kepler - Greener: Twice The Science/Joule Energy used in simulating 1 ns of SMTV 250000 Running NAMD version 2.9 The blue node contains Dual E5-2687W CPUs 200000 (150W each, 8 Cores per CPU). Energy Expended (kJ) Lower is better The green nodes contain Dual E5-2687W CPUs (8 Cores per CPU) and 2x NVIDIA K10, K20, or 150000 K20X GPUs (235W each). Energy Expended 100000 = Power x Time 50000 0 CPU Only CPU + 2 K10s CPU + 2 K20s CPU + 2 K20Xs Cut down energy usage by ? with GPUs Satellite Tobacco Mosaic Virus
Recommended GPU Node Configuration for NAMD Computational Chemistry Workstation or Single Node Configuration # of CPU sockets 2 Cores per CPU socket 6+ CPU speed (Ghz) 2.66+ System memory per socket (GB) 32 Kepler K10, K20, K20X GPUs Fermi M2090, M2075, C2075 # of GPUs per CPU socket 1-2 GPU memory preference (GB) 6 GPU to CPU connection PCIe 2.0 or higher Server storage 500 GB or higher Network configuration Gemini, InfiniBand 10 Scale to multiple nodes with same single node configuration NAMD Benchmark Report, Revision 2.0, dated Nov. 5, 2012
Summary/Conclusions Benefits of GPU Accelerated Computing Faster than CPU only systems in all tests Large performance boost with small marginal price increase Energy usage cut in half GPUs scale very well within a node and over multiple nodes Tesla K20 GPU is our fastest and lowest power high performance GPU to date Try GPU accelerated NAMD for free ¨C www.nvidia.com/GPUTestDrive 11 NAMD Benchmark Report, Revision 2.0, dated Nov. 5, 2012

More Related Content

NAMD Molecular Dynamics on GPU

  • 2. Summary/Conclusions Benefits of GPU Accelerated Computing Faster than CPU only systems in all tests Large performance boost with small marginal price increase Energy usage cut in half GPUs scale very well within a node and over multiple nodes Tesla K20 GPU is our fastest and lowest power high performance GPU to date Try GPU accelerated NAMD for free ¨C www.nvidia.com/GPUTestDrive
  • 3. Kepler - Our Fastest Family of GPUs Yet 4.50 ApoA1 Running NAMD version 2.9 4.00 4.00 The blue node contains Dual E5-2687W CPUs 3.57 (8 Cores per CPU). 3.45 3.50 The green nodes contain Dual E5-2687W CPUs (8 2.9x Cores per CPU) and either 1x NVIDIA M2090, 1x K10 3.00 or 1x K20 for the GPU Nanoseconds/Day 2.63 2.6x 2.50 2.5x 2.00 1.50 1.37 1.9x 1.00 0.50 0.00 1 CPU Node 1 CPU Node + 1 CPU Node + K10 1 CPU Node + K20 1 CPU Node + K20X Apolipoprotein A1 M2090 GPU speedup/throughput increased from 1.9x (with M2090) to 2.9x (with K20X) when compared to a CPU only node 3 NAMD Benchmark Report, Revision 2.0, dated Nov. 5, 2012
  • 4. Run NAMD 2.5x Faster with GPUs 3 Running NAMD 2.9 with CUDA 4.0 ECC Off 2.7 2.6 The blue node contains 2x Intel E5-2687W CPUs 2.5 2.4 (8 Cores per CPU) Speedup Compared to CPU Only Each green node contains 2x Intel E5-2687W 2 CPUs (8 Cores per CPU) plus 1x NVIDIA K20 GPUs 1.5 1 0.5 0 CPU All Molecules ApoA1 F1-ATPase STMV Apolipoprotein A1 Gain 2.5x throughput/performance by adding just 1 GPU when compared to dual CPU performance 4 NAMD Benchmark Report, Revision 2.0, dated Nov. 5, 2012
  • 5. Kepler ¨C Universally Faster 6 Running NAMD version 2.9 The CPU Only node contains Dual E5-2687W CPUs 5 (8 Cores per CPU). Speedup Compared to CPU Only 5.1x The Kepler nodes contain Dual E5-2687W CPUs (8 4 4.7x Cores per CPU) and 1 or two NVIDIA K10, K20, or K20X GPUs. 4.3x F1-ATPase 3 ApoA1 STMV 2.9x 2 2.6x 2.4x 1 0 CPU Only 1x K10 1x K20 1x K20X 2x K10 2x K20 2x K20X F1-ATPase | Kepler nodes use Dual CPUs | The Kepler GPUs accelerate all simulations, up to 5x Average acceleration printed in bars
  • 6. Outstanding Strong Scaling with Multi-STMV Running NAMD version 2.9 Each blue XE6 CPU node contains 1x AMD 100 STMV on Hundreds of Nodes 1600 Opteron (16 Cores per CPU). 1.2 Fermi XK6 Each green XK6 CPU+GPU node contains 1x AMD 1600 Opteron (16 Cores per CPU) 1 and an additional 1x NVIDIA X2090 GPU. CPU XK6 2.7x Nanoseconds / Day 0.8 2.9x 0.6 0.4 0.2 3.6x 3.8x Concatenation of 100 0 Satellite Tobacco Mosaic Virus 32 64 128 256 512 640 768 # of Nodes Accelerate your science by 2.7-3.8x when compared to CPU-based supercomputers
  • 7. Replace 3 Nodes with 1 2090 GPU Running NAMD version 2.9 Each blue node contains 2x Intel Xeon X5550 CPUs F1-ATPase (4 Cores, $1000 per CPU). 4 CPU Nodes 0.8 9000 0.74 The green node contains 2x Intel Xeon X5550 CPUs $8,000 1 CPU Node +8000 (4 Cores, $1000 per CPU) and 1x NVIDIA M2090 GPU 0.7 1x M2090 GPUs 0.63 ($2000 each) 7000 0.6 6000 0.5 5000 0.4 $4,000 4000 0.3 3000 0.2 2000 0.1 1000 0 0 F1-ATPase Nanoseconds/Day Cost Speedup of 1.2x for 50% the cost
  • 8. K20 - Greener: Twice The Science Per Watt 1200000 Energy Used in Simulating 1 Nanosecond of ApoA1 Running NAMD version 2.9 1000000 Each blue node contains Dual E5-2687W CPUs (95W, 4 Cores per CPU). Each green node contains 2x Intel Xeon X5550 Energy Expended (kJ) 800000 CPUs (95W, 4 Cores per CPU) and 2x NVIDIA Lower is better K20 GPUs (225W per GPU) 600000 Energy Expended 400000 = Power x Time 200000 0 1 Node 1 Node + 2x K20 Cut down energy usage by ? with GPUs 8 NAMD Benchmark Report, Revision 2.0, dated Nov. 5, 2012
  • 9. Kepler - Greener: Twice The Science/Joule Energy used in simulating 1 ns of SMTV 250000 Running NAMD version 2.9 The blue node contains Dual E5-2687W CPUs 200000 (150W each, 8 Cores per CPU). Energy Expended (kJ) Lower is better The green nodes contain Dual E5-2687W CPUs (8 Cores per CPU) and 2x NVIDIA K10, K20, or 150000 K20X GPUs (235W each). Energy Expended 100000 = Power x Time 50000 0 CPU Only CPU + 2 K10s CPU + 2 K20s CPU + 2 K20Xs Cut down energy usage by ? with GPUs Satellite Tobacco Mosaic Virus
  • 10. Recommended GPU Node Configuration for NAMD Computational Chemistry Workstation or Single Node Configuration # of CPU sockets 2 Cores per CPU socket 6+ CPU speed (Ghz) 2.66+ System memory per socket (GB) 32 Kepler K10, K20, K20X GPUs Fermi M2090, M2075, C2075 # of GPUs per CPU socket 1-2 GPU memory preference (GB) 6 GPU to CPU connection PCIe 2.0 or higher Server storage 500 GB or higher Network configuration Gemini, InfiniBand 10 Scale to multiple nodes with same single node configuration NAMD Benchmark Report, Revision 2.0, dated Nov. 5, 2012
  • 11. Summary/Conclusions Benefits of GPU Accelerated Computing Faster than CPU only systems in all tests Large performance boost with small marginal price increase Energy usage cut in half GPUs scale very well within a node and over multiple nodes Tesla K20 GPU is our fastest and lowest power high performance GPU to date Try GPU accelerated NAMD for free ¨C www.nvidia.com/GPUTestDrive 11 NAMD Benchmark Report, Revision 2.0, dated Nov. 5, 2012

Editor's Notes

  • #4: ns/dayDual E5-2687W CPUs 1.370Dual E5-2687W CPUs + M2090 2.632Dual E5-2687W CPUs + K10 3.448Dual E5-2687W CPUs + K20 3.571Dual E5-2687W CPUs + K20X 4.000
  • #5: cpu ns/day gpu ns/dayApoA1 1.370 3.571F1-ATPase 0.461 1.124STMV 0.116 0.314ECC off
  • #6: All #s are days/ns apoa1atpasestmvCPU Only0.732.178.641x K100.291.043.51x K200.28 0.893.181x K20X0.250.82.872x K100.160.561.932x K200.150.491.772x K20X0.140.451.63
  • #7: 32 64 128 256 512 640 768s/step GPU XK6 1.2414 0.660887 0.342743 0.199465 0.10837 0.089752 0.0774948s/step CPU XK6 4.62633 2.36707 1.19722 0.609124 0.314745 0.255016 0.209511ns/day Fermi XK6 0.069599 0.13073339 0.252084 0.433159 0.797269 0.962655 1.114913517ns/day CPU XK6 0.018676 0.03650082 0.072167 0.141843 0.274508 0.338802 0.412388848
  • #9: Config: TDP sec/ns energy 2x E5-2687W 150 63,072.0 9,460,800.0 2x E5-2687W+ 2x K20 600 24,192.0 14,515,200 TDP = Thermal Design Power
  • #10: ns/day tdp energyCpu .115 300 223kK10s .518 770 128kK20s .565 770 117kK20xs .613 770 108k