pytorch001 AMD Ryzen Threadripper PRO 7965WX 24-Cores testing with a ASUS Pro WS WRX90E-SAGE SE (0404 BIOS) and NVIDIA GeForce RTX 4090 24GB on Ubuntu 22.04 via the Phoronix Test Suite.
HTML result view exported from: https://openbenchmarking.org/result/2403301-NE-PYTORCH0019 .
pytorch001 Processor Motherboard Chipset Memory Disk Graphics Audio Monitor Network OS Kernel Desktop Display Server Display Driver Compiler File-System Screen Resolution test001 AMD Ryzen Threadripper PRO 7965WX 24-Cores @ 7.30GHz (24 Cores / 48 Threads) ASUS Pro WS WRX90E-SAGE SE (0404 BIOS) AMD Device 14a4 128GB 2000GB Sabrent SB-ROCKET-NVMe4-2TB + 0GB Virtual HDisk0 NVIDIA GeForce RTX 4090 24GB NVIDIA Device 22ba DELL U2723QE 2 x Intel X710 for 10GBASE-T Ubuntu 22.04 6.5.0-26-generic (x86_64) GNOME Shell 42.9 X Server 1.21.1.4 NVIDIA 545.23.08 GCC 11.4.0 + CUDA 12.3 ext4 4480x2160 OpenBenchmarking.org - Transparent Huge Pages: madvise - Scaling Governor: amd-pstate-epp powersave (EPP: performance) - CPU Microcode: 0xa108105 - Python 3.11.8 - gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: Mitigation of Safe RET + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced / Automatic IBRS IBPB: conditional STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected
pytorch001 pytorch: CPU - 32 - ResNet-50 pytorch: CPU - 32 - ResNet-152 pytorch: CPU - 32 - Efficientnet_v2_l pytorch: NVIDIA CUDA GPU - 32 - ResNet-50 pytorch: NVIDIA CUDA GPU - 32 - ResNet-152 pytorch: NVIDIA CUDA GPU - 32 - Efficientnet_v2_l test001 46.00 17.84 9.92 364.38 132.78 66.63 OpenBenchmarking.org
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 test001 10 20 30 40 50 SE +/- 0.09, N = 3 46.00 MIN: 40.27 / MAX: 46.99
PyTorch Device: CPU - Batch Size: 32 - Model: ResNet-152 OpenBenchmarking.org batches/sec, More Is Better PyTorch 2.2.1 Device: CPU - Batch Size: 32 - Model: ResNet-152 test001 4 8 12 16 20 SE +/- 0.12, N = 3 17.84 MIN: 17.03 / MAX: 18.28
PyTorch Device: CPU - Batch Size: 32 - Model: Efficientnet_v2_l OpenBenchmarking.org batches/sec, More Is Better PyTorch 2.2.1 Device: CPU - Batch Size: 32 - Model: Efficientnet_v2_l test001 3 6 9 12 15 SE +/- 0.00, N = 3 9.92 MIN: 8.18 / MAX: 10.32
PyTorch Device: NVIDIA CUDA GPU - Batch Size: 32 - Model: ResNet-50 OpenBenchmarking.org batches/sec, More Is Better PyTorch 2.2.1 Device: NVIDIA CUDA GPU - Batch Size: 32 - Model: ResNet-50 test001 80 160 240 320 400 SE +/- 0.60, N = 3 364.38 MIN: 330.44 / MAX: 371.21
PyTorch Device: NVIDIA CUDA GPU - Batch Size: 32 - Model: ResNet-152 OpenBenchmarking.org batches/sec, More Is Better PyTorch 2.2.1 Device: NVIDIA CUDA GPU - Batch Size: 32 - Model: ResNet-152 test001 30 60 90 120 150 SE +/- 1.16, N = 3 132.78 MIN: 114.16 / MAX: 136.93
PyTorch Device: NVIDIA CUDA GPU - Batch Size: 32 - Model: Efficientnet_v2_l OpenBenchmarking.org batches/sec, More Is Better PyTorch 2.2.1 Device: NVIDIA CUDA GPU - Batch Size: 32 - Model: Efficientnet_v2_l test001 15 30 45 60 75 SE +/- 0.60, N = 3 66.63 MIN: 56.43 / MAX: 68.66
Phoronix Test Suite v10.8.4