pytorch 2.2.1 AMD EPYC Siena

AMD EPYC 8534P 64-Core testing with a AMD Cinnabar (RCB1009C BIOS) and ASPEED on Ubuntu 23.10 via the Phoronix Test Suite.

Compare your own system(s) to this result file with the Phoronix Test Suite by running the command: phoronix-test-suite benchmark 2403269-NE-PYTORCH2202
Jump To Table - Results

View

Do Not Show Noisy Results
Do Not Show Results With Incomplete Data
Do Not Show Results With Little Change/Spread
List Notable Results
Show Result Confidence Charts
Allow Limiting Results To Certain Suite(s)

Statistics

Show Overall Harmonic Mean(s)
Show Overall Geometric Mean
Show Wins / Losses Counts (Pie Chart)
Normalize Results
Remove Outliers Before Calculating Averages

Graph Settings

Force Line Graphs Where Applicable
Convert To Scalar Where Applicable
Prefer Vertical Bar Graphs

Multi-Way Comparison

Condense Multi-Option Tests Into Single Result Graphs

Table

Show Detailed System Result Table

Run Management

Highlight
Result
Toggle/Hide
Result
Result
Identifier
View Logs
Performance Per
Dollar
Date
Run
  Test
  Duration
a
March 26
  51 Minutes
b
March 26
  51 Minutes
c
March 26
  51 Minutes
Invert Behavior (Only Show Selected Data)
  51 Minutes

Only show results where is faster than
Only show results matching title/arguments (delimit multiple options with a comma):
Do not show results matching title/arguments (delimit multiple options with a comma):

a

Kernel Notes: Transparent Huge Pages: madvise
Processor Notes: Scaling Governor: acpi-cpufreq performance (Boost: Enabled) - CPU Microcode: 0xaa00212
Python Notes: Python 3.11.6
Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + reg_file_data_sampling: 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

b

c

Processor: AMD EPYC 8534P 64-Core @ 2.30GHz (64 Cores / 128 Threads), Motherboard: AMD Cinnabar (RCB1009C BIOS), Chipset: AMD Device 14a4, Memory: 6 x 32GB DRAM-4800MT/s Samsung M321R4GA0BB0-CQKMG, Disk: 3201GB Micron_7450_MTFDKCB3T2TFS + 2000GB Corsair MP700, Graphics: ASPEED, Network: 2 x Broadcom NetXtreme BCM5720 PCIe

OS: Ubuntu 23.10, Kernel: 6.8.1-060801-generic (x86_64), Desktop: GNOME Shell 45.2, Display Server: X Server 1.21.1.7, Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1920x1200

pytorch 2.2.1 AMD EPYC SienaOpenBenchmarking.orgPhoronix Test SuiteAMD EPYC 8534P 64-Core @ 2.30GHz (64 Cores / 128 Threads)AMD Cinnabar (RCB1009C BIOS)AMD Device 14a46 x 32GB DRAM-4800MT/s Samsung M321R4GA0BB0-CQKMG3201GB Micron_7450_MTFDKCB3T2TFS + 2000GB Corsair MP700ASPEED2 x Broadcom NetXtreme BCM5720 PCIeUbuntu 23.106.8.1-060801-generic (x86_64)GNOME Shell 45.2X Server 1.21.1.7GCC 13.2.0ext41920x1200ProcessorMotherboardChipsetMemoryDiskGraphicsNetworkOSKernelDesktopDisplay ServerCompilerFile-SystemScreen ResolutionPytorch 2.2.1 AMD EPYC Siena BenchmarksSystem Logs- Transparent Huge Pages: madvise- Scaling Governor: acpi-cpufreq performance (Boost: Enabled) - CPU Microcode: 0xaa00212 - Python 3.11.6- gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + reg_file_data_sampling: 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

abcResult OverviewPhoronix Test Suite100%101%102%102%103%PyTorchPyTorchPyTorchPyTorchPyTorchPyTorchPyTorchPyTorchPyTorchPyTorchPyTorchPyTorchPyTorchPyTorchPyTorchPyTorchPyTorchPyTorchCPU - 256 - ResNet-152CPU - 32 - ResNet-50CPU - 64 - ResNet-152CPU - 1 - ResNet-152CPU - 1 - Efficientnet_v2_lCPU - 32 - Efficientnet_v2_lCPU - 512 - ResNet-152CPU - 1 - ResNet-50CPU - 64 - ResNet-50CPU - 16 - ResNet-152CPU - 256 - ResNet-50CPU - 16 - ResNet-50CPU - 64 - Efficientnet_v2_lCPU - 32 - ResNet-152CPU - 512 - ResNet-50CPU - 16 - Efficientnet_v2_lCPU - 512 - Efficientnet_v2_lCPU - 256 - Efficientnet_v2_l

pytorch 2.2.1 AMD EPYC Sienapytorch: CPU - 16 - Efficientnet_v2_lpytorch: CPU - 64 - Efficientnet_v2_lpytorch: CPU - 32 - Efficientnet_v2_lpytorch: CPU - 512 - Efficientnet_v2_lpytorch: CPU - 256 - Efficientnet_v2_lpytorch: CPU - 64 - ResNet-152pytorch: CPU - 32 - ResNet-152pytorch: CPU - 512 - ResNet-152pytorch: CPU - 16 - ResNet-152pytorch: CPU - 256 - ResNet-152pytorch: CPU - 1 - Efficientnet_v2_lpytorch: CPU - 1 - ResNet-152pytorch: CPU - 32 - ResNet-50pytorch: CPU - 16 - ResNet-50pytorch: CPU - 512 - ResNet-50pytorch: CPU - 256 - ResNet-50pytorch: CPU - 64 - ResNet-50pytorch: CPU - 1 - ResNet-50abc5.655.675.645.685.6512.9613.0713.2413.1912.978.6415.1532.9433.3033.4233.6733.8641.115.655.635.625.655.6712.9913.0313.0913.0613.108.7814.8833.6533.3133.5133.3433.5140.675.625.665.695.655.6813.2212.9913.1013.1013.398.6415.0333.3633.5433.6233.5433.6540.75OpenBenchmarking.org

PyTorch

This is a benchmark of PyTorch making use of pytorch-benchmark [https://github.com/LukasHedegaard/pytorch-benchmark]. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 16 - Model: Efficientnet_v2_lcba1.27132.54263.81395.08526.35655.625.655.65MIN: 4.63 / MAX: 5.72MIN: 4.71 / MAX: 5.76MIN: 4.48 / MAX: 5.76

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 64 - Model: Efficientnet_v2_lcba1.27582.55163.82745.10326.3795.665.635.67MIN: 4.52 / MAX: 5.76MIN: 4.88 / MAX: 5.72MIN: 5.32 / MAX: 5.76

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 32 - Model: Efficientnet_v2_lcba1.28032.56063.84095.12126.40155.695.625.64MIN: 4.72 / MAX: 5.78MIN: 5.06 / MAX: 5.71MIN: 4.7 / MAX: 5.75

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 512 - Model: Efficientnet_v2_lcba1.2782.5563.8345.1126.395.655.655.68MIN: 4.41 / MAX: 5.74MIN: 5.38 / MAX: 5.73MIN: 4.6 / MAX: 5.77

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 256 - Model: Efficientnet_v2_lcba1.2782.5563.8345.1126.395.685.675.65MIN: 5.4 / MAX: 5.78MIN: 4.55 / MAX: 5.76MIN: 4.62 / MAX: 5.73

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 64 - Model: ResNet-152cba369121513.2212.9912.96MIN: 13.02 / MAX: 13.32MIN: 12.86 / MAX: 13.09MIN: 12.23 / MAX: 13.06

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 32 - Model: ResNet-152cba369121512.9913.0313.07MIN: 8.89 / MAX: 13.09MIN: 12.7 / MAX: 13.14MIN: 8.78 / MAX: 13.2

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 512 - Model: ResNet-152cba369121513.1013.0913.24MIN: 12.96 / MAX: 13.2MIN: 8.81 / MAX: 13.18MIN: 12.34 / MAX: 13.35

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 16 - Model: ResNet-152cba369121513.1013.0613.19MIN: 12.7 / MAX: 13.21MIN: 8.77 / MAX: 13.16MIN: 13.07 / MAX: 13.3

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 256 - Model: ResNet-152cba369121513.3913.1012.97MIN: 8.95 / MAX: 13.48MIN: 13 / MAX: 13.2MIN: 8.8 / MAX: 13.09

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 1 - Model: Efficientnet_v2_lcba2468108.648.788.64MIN: 8.5 / MAX: 8.72MIN: 6.74 / MAX: 8.86MIN: 8.55 / MAX: 8.71

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 1 - Model: ResNet-152cba4812162015.0314.8815.15MIN: 9.88 / MAX: 15.14MIN: 14.71 / MAX: 15MIN: 14.65 / MAX: 15.33

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 32 - Model: ResNet-50cba81624324033.3633.6532.94MIN: 31.61 / MAX: 33.8MIN: 32.2 / MAX: 34.12MIN: 14.77 / MAX: 33.36

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 16 - Model: ResNet-50cba81624324033.5433.3133.30MIN: 18.28 / MAX: 34MIN: 32.5 / MAX: 33.78MIN: 16.55 / MAX: 33.78

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 512 - Model: ResNet-50cba81624324033.6233.5133.42MIN: 32.63 / MAX: 34.14MIN: 32.71 / MAX: 33.92MIN: 32.29 / MAX: 33.81

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 256 - Model: ResNet-50cba81624324033.5433.3433.67MIN: 16.1 / MAX: 34.1MIN: 32.81 / MAX: 33.77MIN: 32.41 / MAX: 34.14

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 64 - Model: ResNet-50cba81624324033.6533.5133.86MIN: 32.6 / MAX: 34.13MIN: 15.58 / MAX: 33.94MIN: 32.86 / MAX: 34.37

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 1 - Model: ResNet-50cba91827364540.7540.6741.11MIN: 39.84 / MAX: 41.56MIN: 39.57 / MAX: 41.58MIN: 39.95 / MAX: 41.96

a

Kernel Notes: Transparent Huge Pages: madvise
Processor Notes: Scaling Governor: acpi-cpufreq performance (Boost: Enabled) - CPU Microcode: 0xaa00212
Python Notes: Python 3.11.6
Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + reg_file_data_sampling: 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

Testing initiated at 26 March 2024 20:15 by user phoronix.

b

Kernel Notes: Transparent Huge Pages: madvise
Processor Notes: Scaling Governor: acpi-cpufreq performance (Boost: Enabled) - CPU Microcode: 0xaa00212
Python Notes: Python 3.11.6
Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + reg_file_data_sampling: 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

Testing initiated at 26 March 2024 21:08 by user phoronix.

c

Processor: AMD EPYC 8534P 64-Core @ 2.30GHz (64 Cores / 128 Threads), Motherboard: AMD Cinnabar (RCB1009C BIOS), Chipset: AMD Device 14a4, Memory: 6 x 32GB DRAM-4800MT/s Samsung M321R4GA0BB0-CQKMG, Disk: 3201GB Micron_7450_MTFDKCB3T2TFS + 2000GB Corsair MP700, Graphics: ASPEED, Network: 2 x Broadcom NetXtreme BCM5720 PCIe

OS: Ubuntu 23.10, Kernel: 6.8.1-060801-generic (x86_64), Desktop: GNOME Shell 45.2, Display Server: X Server 1.21.1.7, Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1920x1200

Kernel Notes: Transparent Huge Pages: madvise
Processor Notes: Scaling Governor: acpi-cpufreq performance (Boost: Enabled) - CPU Microcode: 0xaa00212
Python Notes: Python 3.11.6
Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + reg_file_data_sampling: 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

Testing initiated at 26 March 2024 22:40 by user phoronix.