gpu-server-2u 2 x Intel Xeon E5-2697 v4 testing with a Dell PowerEdge R730 [0WCJNT] (2.19.0 BIOS) and NVIDIA GA102GL [RTX A5000] 24GB on Ubuntu 22.04 via the Phoronix Test Suite.
HTML result view exported from: https://openbenchmarking.org/result/2404153-NE-GPUSERVER09&sor&grr .
gpu-server-2u Processor Motherboard Chipset Memory Disk Graphics Audio Network OS Kernel Display Driver Vulkan Compiler File-System Screen Resolution OpenCL GPU-SERVER-2U-2xA5000 2 x Intel Xeon E5-2697 v4 - NVIDIA GA102GL [RTX NVIDIA GA102GL 2 x Intel Xeon E5-2697 v4 @ 3.60GHz (36 Cores / 72 Threads) Dell PowerEdge R730 [0WCJNT] (2.19.0 BIOS) Intel Xeon E7 v4/Xeon 8 x 32 GB DDR4-2400MT/s M393A4K40CB1-CRC 2 x 1920GB SAMSUNG MZ7LH1T9 NVIDIA GA102GL [RTX A5000] 24GB NVIDIA GA102 HD Audio 2 x Intel 10-Gigabit X540-AT2 + 2 x Intel I350 Ubuntu 22.04 5.15.0-102-generic (x86_64) NVIDIA 1.3.277 GCC 12.3.0 ext4 1024x768 OpenCL 3.0 CUDA 12.4.125 OpenBenchmarking.org Kernel Details - Transparent Huge Pages: madvise Processor Details - GPU-SERVER-2U-2xA5000: Scaling Governor: intel_cpufreq schedutil - CPU Microcode: 0xb000040 - 2 x Intel Xeon E5-2697 v4 - NVIDIA GA102GL [RTX: Scaling Governor: intel_cpufreq performance - CPU Microcode: 0xb000040 - NVIDIA GA102GL: Scaling Governor: intel_cpufreq performance - CPU Microcode: 0xb000040 Python Details - GPU-SERVER-2U-2xA5000, 2 x Intel Xeon E5-2697 v4 - NVIDIA GA102GL [RTX: Python 3.10.12 Security Details - gather_data_sampling: Not affected + itlb_multihit: KVM: Mitigation of VMX disabled + l1tf: Mitigation of PTE Inversion; VMX: conditional cache flushes SMT vulnerable + mds: Mitigation of Clear buffers; SMT vulnerable + meltdown: Mitigation of PTI + mmio_stale_data: Mitigation of Clear buffers; SMT vulnerable + retbleed: Not affected + spec_rstack_overflow: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines IBPB: conditional IBRS_FW STIBP: conditional RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Mitigation of Clear buffers; SMT vulnerable Graphics Details - NVIDIA GA102GL: BAR1 / Visible vRAM Size: 256 MiB - vBIOS Version: 94.02.6d.00.0d
gpu-server-2u pytorch: CPU - 1 - ResNet-152 pytorch: CPU - 32 - ResNet-50 pytorch: CPU - 16 - ResNet-50 pytorch: CPU - 64 - ResNet-50 pytorch: NVIDIA CUDA GPU - 512 - ResNet-152 pytorch: CPU - 1 - ResNet-50 hashcat: SHA-512 ai-benchmark: GPU-SERVER-2U-2xA5000 2 x Intel Xeon E5-2697 v4 - NVIDIA GA102GL [RTX NVIDIA GA102GL 10.39 22.57 23.10 23.84 33.72 27.48 3986400000 OpenBenchmarking.org
PyTorch Device: CPU - Batch Size: 1 - Model: ResNet-152 OpenBenchmarking.org batches/sec, More Is Better PyTorch 2.2.1 Device: CPU - Batch Size: 1 - Model: ResNet-152 2 x Intel Xeon E5-2697 v4 - NVIDIA GA102GL [RTX 3 6 9 12 15 SE +/- 0.10, N = 3 10.39 MIN: 7.26 / MAX: 11.03
PyTorch Device: CPU - Batch Size: 32 - Model: ResNet-50 OpenBenchmarking.org batches/sec, More Is Better PyTorch 2.2.1 Device: CPU - Batch Size: 32 - Model: ResNet-50 2 x Intel Xeon E5-2697 v4 - NVIDIA GA102GL [RTX 5 10 15 20 25 SE +/- 0.27, N = 3 22.57 MIN: 16.36 / MAX: 23.62
PyTorch Device: CPU - Batch Size: 16 - Model: ResNet-50 OpenBenchmarking.org batches/sec, More Is Better PyTorch 2.2.1 Device: CPU - Batch Size: 16 - Model: ResNet-50 2 x Intel Xeon E5-2697 v4 - NVIDIA GA102GL [RTX 6 12 18 24 30 SE +/- 0.19, N = 3 23.10 MIN: 16.52 / MAX: 24.05
PyTorch Device: CPU - Batch Size: 64 - Model: ResNet-50 OpenBenchmarking.org batches/sec, More Is Better PyTorch 2.2.1 Device: CPU - Batch Size: 64 - Model: ResNet-50 2 x Intel Xeon E5-2697 v4 - NVIDIA GA102GL [RTX 6 12 18 24 30 SE +/- 0.29, N = 3 23.84 MIN: 17.15 / MAX: 24.64
PyTorch Device: NVIDIA CUDA GPU - Batch Size: 512 - Model: ResNet-152 OpenBenchmarking.org batches/sec, More Is Better PyTorch 2.2.1 Device: NVIDIA CUDA GPU - Batch Size: 512 - Model: ResNet-152 2 x Intel Xeon E5-2697 v4 - NVIDIA GA102GL [RTX 8 16 24 32 40 SE +/- 0.10, N = 3 33.72 MIN: 29.31 / MAX: 34.08
PyTorch Device: CPU - Batch Size: 1 - Model: ResNet-50 OpenBenchmarking.org batches/sec, More Is Better PyTorch 2.2.1 Device: CPU - Batch Size: 1 - Model: ResNet-50 2 x Intel Xeon E5-2697 v4 - NVIDIA GA102GL [RTX 6 12 18 24 30 SE +/- 0.37, N = 3 27.48 MIN: 11.6 / MAX: 29.11
Hashcat Benchmark: SHA-512 OpenBenchmarking.org H/s, More Is Better Hashcat 6.2.4 Benchmark: SHA-512 NVIDIA GA102GL 900M 1800M 2700M 3600M 4500M SE +/- 5750652.14, N = 3 3986400000
Phoronix Test Suite v10.8.5