Slide 1: AMBER 16 on V100s
AMBER 16 on V100s October 2017
Slide 2: PME-Cellulose_NPT on V100s PCIe
PME-Cellulose_NPT on V100s PCIe 50 47.67 40 24.6X (Untuned on Volta) Running AMBER version 16.8 30 ns/day The blue node contains Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] 20 (Broadwell) CPUs The green nodes contain Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz 10 Turbo] (Broadwell) CPUs + Tesla V
Slide 3: PME-Cellulose_NPT on V100s SXM2
PME-Cellulose_NPT on V100s SXM2 60 54.74 55.52 50 28.2X 28.6X 40 (Untuned on Volta) Running AMBER version 16.8 ns/day 30 The blue node contains Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs 20 The green nodes contain Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CP
Slide 4: PME-Cellulose_NVE on V100s PCIe
PME-Cellulose_NVE on V100s PCIe 60 54.08 50 40 27.6X (Untuned on Volta) Running AMBER version 16.8 ns/day 30 The blue node contains Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] (Broadwell) CPUs 20 The green nodes contain Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] (Broadwell) CPUs + 10 Tesl
Slide 5: PME-Cellulose_NVE on V100s SXM2
PME-Cellulose_NVE on V100s SXM2 70 65.02 63.04 60 50 33.2X (Untuned on Volta) 40 32.2X Running AMBER version 16.8 ns/day The blue node contains Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] 30 (Broadwell) CPUs 20 The green nodes contain Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell)
Slide 6: PME-FactorIX_NPT on V100s PCIe
PME-FactorIX_NPT on V100s PCIe 250 200 193.16 150 20.7X (Untuned on Volta) Running AMBER version 16.8 ns/day The blue node contains Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] 100 (Broadwell) CPUs The green nodes contain Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz 50 Turbo] (Broadwell) CPUs + Tes
Slide 7: PME-FactorIX_NPT on V100s SXM2
PME-FactorIX_NPT on V100s SXM2 250 224.23 217.95 200 24.0X 150 23.4X (Untuned on Volta) Running AMBER version 16.8 ns/day The blue node contains Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] 100 (Broadwell) CPUs The green nodes contain Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz 50 Turbo] (Broadwel
Slide 8: PME-FactorIX_NVE on V100s PCIe
PME-FactorIX_NVE on V100s PCIe 250 217.95 200 22.7X (Untuned on Volta) Running AMBER version 16.8 150 ns/day The blue node contains Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] 100 (Broadwell) CPUs The green nodes contain Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz 50 Turbo] (Broadwell) CPUs + Tes
Slide 9: PME-FactorIX_NVE on V100s SXM2
PME-FactorIX_NVE on V100s SXM2 300 261.19 249.63 250 200 27.2X (Untuned on Volta) Running AMBER version 16.8 26.0X ns/day 150 The blue node contains Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs 100 The green nodes contain Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwe
Slide 10: PME-JAC_NPT on V100s PCIe
PME-JAC_NPT on V100s PCIe 500 439.87 400 12.8X (Untuned on Volta) Running AMBER version 16.8 300 ns/day The blue node contains Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] 200 (Broadwell) CPUs The green nodes contain Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz 100 Turbo] (Broadwell) CPUs + Tesla V
Slide 11: PME-JAC_NPT on V100s SXM2
PME-JAC_NPT on V100s SXM2 600 515.36 500 481.75 400 15.0X (Untuned on Volta) Running AMBER version 16.8 14.0X ns/day 300 The blue node contains Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs 200 The green nodes contain Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) C
Slide 12: PME-JAC_NVE on V100s PCIe
PME-JAC_NVE on V100s PCIe 600 490.77 500 400 13.4X (Untuned on Volta) Running AMBER version 16.8 ns/day 300 The blue node contains Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] (Broadwell) CPUs 200 The green nodes contain Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] (Broadwell) CPUs + 100 Tes
Slide 13: PME-JAC_NVE on V100s SXM2
PME-JAC_NVE on V100s SXM2 700 600 583.33 539.78 500 (Untuned on Volta) 400 16.0X Running AMBER version 16.8 ns/day 300 14.8X The blue node contains Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs 200 The green nodes contain Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwel
Slide 14: PME-JAC_NPT_4fs on V100s PCIe
PME-JAC_NPT_4fs on V100s PCIe 900 863.80 750 600 13.1X (Untuned on Volta) Running AMBER version 16.8 ns/day 450 The blue node contains Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] (Broadwell) CPUs 300 The green nodes contain Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] (Broadwell) CPUs + 150
Slide 15: PME-JAC_NPT_4fs on V100s SXM2
PME-JAC_NPT_4fs on V100s SXM2 1200 1006.32 1000 946.57 800 15.3X (Untuned on Volta) Running AMBER version 16.8 14.4X ns/day 600 The blue node contains Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs 400 The green nodes contain Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broad
Slide 16: PME-JAC_NVE_4fs on V100s PCIe
PME-JAC_NVE_4fs on V100s PCIe 1050 940.32 900 750 26.0X (Untuned on Volta) Running AMBER version 16.8 600 ns/day The blue node contains Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] 450 (Broadwell) CPUs 300 The green nodes contain Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] (Broadwell) CPUs
Slide 17: PME-JAC_NVE_4fs on V100s SXM2
PME-JAC_NVE_4fs on V100s SXM2 1200 1123.40 1027.44 1000 800 16.7X (Untuned on Volta) Running AMBER version 16.8 15.3X ns/day 600 The blue node contains Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs 400 The green nodes contain Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broa
Slide 18: PME-STMV_NPT_4fs on V100s PCIe
PME-STMV_NPT_4fs on V100s PCIe 35 33.21 30 25 31.3X (Untuned on Volta) Running AMBER version 16.8 20 ns/day The blue node contains Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] 15 (Broadwell) CPUs 10 The green nodes contain Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] (Broadwell) CPUs + 5 Tes
Slide 19: PME-STMV_NPT_4fs on V100s SXM2
PME-STMV_NPT_4fs on V100s SXM2 40 37.24 35 30 25 35.1X (Untuned on Volta) Running AMBER version 16.8 ns/day 20 The blue node contains Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs 15 The green nodes contain Dual Intel 10 Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs + 5
Slide 20: GB-Myoglobin on V100s PCIe
GB-Myoglobin on V100s PCIe 750 699.21 600 31.4X (Untuned on Volta) Running AMBER version 16.8 450 ns/day The blue node contains Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] 300 (Broadwell) CPUs The green nodes contain Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz 150 Turbo] (Broadwell) CPUs + Tesla
Slide 21: GB-Myoglobin on V100s SXM2
GB-Myoglobin on V100s SXM2 800 750.76 700 600 33.7X (Untuned on Volta) 500 Running AMBER version 16.8 ns/day 400 The blue node contains Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs 300 The green nodes contain Dual Intel 200 Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs
Slide 22: GB-Nucleosome on V100s PCIe
GB-Nucleosome on V100s PCIe 85 78.39 68 49.14 252.9X (Untuned on Volta) Running AMBER version 16.8 51 ns/day The blue node contains Dual Intel Xeon 34 158.5X E5-2690 v4@2.6GHz [3.5GHz Turbo] (Broadwell) CPUs The green nodes contain Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz 17 Turbo] (Broadwell) CPUs
Slide 23: GB-Nucleosome on V100s SXM2
GB-Nucleosome on V100s SXM2 100 92.46 75 (Untuned on Volta) 52.89 298.3X Running AMBER version 16.8 ns/day 50 The blue node contains Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs 170.6X The green nodes contain Dual Intel 25 Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs +
Slide 24: Rubisco on V100s PCIe
Rubisco on V100s PCIe 8 7 6.78 6 5 5.22 678.0X (Untuned on Volta) Running AMBER version 16.8 ns/day 4 The blue node contains Dual Intel Xeon 522.0X E5-2690 v4@2.6GHz [3.5GHz Turbo] (Broadwell) CPUs 3 2.79 The green nodes contain Dual Intel 2 Xeon E5-2690 v4@2.6GHz [3.5GHz 1 279.0X Turbo] (Broadwell)
Slide 25: Rubisco on V100s SXM2
Rubisco on V100s SXM2 8 7.00 7 5.96 700.0X 6 (Untuned on Volta) 5 Running AMBER version 16.8 ns/day 4 The blue node contains Dual Intel Xeon 596.0X E5-2698 v4@2.2GHz [3.6GHz Turbo] 3.00 (Broadwell) CPUs 3 The green nodes contain Dual Intel 2 Xeon E5-2698 v4@2.2GHz [3.6GHz 1 300.0X Turbo] (Broadwell)
Slide 26: Recommended GPU Node Configuration for AMBER Computational Chemistry
Recommended GPU Node Configuration for AMBER Computational Chemistry Workstation or Single Node Configuration # of CPU sockets 2 Cores per CPU socket 6+ (1 CPU core drives 1 GPU) CPU speed (Ghz) 2.66+ System memory per node (GB) 16 GPUs P100, V100 1-4 # of GPUs per CPU socket GPU memory preference (