nlp-benchmarks

AWS EC2 Amazon Linux 2023 Benchmarking

HTML result view exported from: https://openbenchmarking.org/result/2402089-NE-2402012NE50&rdt&grs.

nlp-benchmarksProcessorMotherboardChipsetMemoryDiskNetworkOSKernelCompilerFile-SystemSystem Layerc6i.2xlargem7i-flex.2xlargec7a.2xlargeIntel Xeon Platinum 8375C (4 Cores / 8 Threads)Amazon EC2 c6i.2xlarge (1.0 BIOS)Intel 440FX 82441FX PMC1 x 16GB DDR4-3200MT/s215GB Amazon Elastic Block StoreAmazon ElasticAmazon Linux 20236.1.61-85.141.amzn2023.x86_64 (x86_64)GCC 11.4.1 20230605xfsamazonIntel Xeon Platinum 8488C (4 Cores / 8 Threads)Amazon EC2 m7i-flex.2xlarge (1.0 BIOS)1 x 32GB 4800MT/s6.1.72-96.166.amzn2023.x86_64 (x86_64)AMD EPYC 9R14 (8 Cores)Amazon EC2 c7a.2xlarge (1.0 BIOS)1 x 16GB 4800MT/sOpenBenchmarking.orgKernel Details- Transparent Huge Pages: madviseCompiler Details- --build=x86_64-amazon-linux --disable-libunwind-exceptions --enable-__cxa_atexit --enable-bootstrap --enable-cet --enable-checking=release --enable-gnu-indirect-function --enable-gnu-unique-object --enable-initfini-array --enable-languages=c,c++,fortran,lto --enable-multilib --enable-offload-targets=nvptx-none --enable-plugin --enable-shared --enable-threads=posix --mandir=/usr/share/man --with-arch_32=x86-64 --with-arch_64=x86-64-v2 --with-gcc-major-version-only --with-linker-hash-style=gnu --with-tune=generic --without-cuda-driver Processor Details- c6i.2xlarge: CPU Microcode: 0xd0003a5- m7i-flex.2xlarge: CPU Microcode: 0x2b000571- c7a.2xlarge: CPU Microcode: 0xa10113ePython Details- Python 3.11.6Security Details- c6i.2xlarge: gather_data_sampling: Unknown: Dependent on hypervisor status + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Mitigation of Clear buffers; SMT Host state unknown + retbleed: Not affected + spec_rstack_overflow: Not affected + 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 IBRS IBPB: conditional RSB filling PBRSB-eIBRS: SW sequence + srbds: Not affected + tsx_async_abort: Not affected - m7i-flex.2xlarge: 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: Not affected + 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 IBRS IBPB: conditional RSB filling PBRSB-eIBRS: SW sequence + srbds: Not affected + tsx_async_abort: Not affected - c7a.2xlarge: 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 no microcode + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines IBPB: conditional IBRS_FW STIBP: disabled RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected

nlp-benchmarksopenvino: Face Detection FP16 - CPUopenvino: Face Detection FP16 - CPUonednn: Deconvolution Batch shapes_1d - bf16bf16bf16 - CPUopenvino: Face Detection FP16-INT8 - CPUopenvino: Face Detection FP16-INT8 - CPUonednn: Deconvolution Batch shapes_1d - f32 - CPUopenvino: Machine Translation EN To DE FP16 - CPUopenvino: Machine Translation EN To DE FP16 - CPUpytorch: CPU - 32 - Efficientnet_v2_lpytorch: CPU - 16 - Efficientnet_v2_lpytorch: CPU - 32 - ResNet-152pytorch: CPU - 16 - ResNet-152onednn: Deconvolution Batch shapes_1d - u8s8f32 - CPUonednn: Deconvolution Batch shapes_3d - bf16bf16bf16 - CPUonednn: Convolution Batch Shapes Auto - bf16bf16bf16 - CPUpytorch: CPU - 16 - ResNet-50pytorch: CPU - 1 - ResNet-152pytorch: CPU - 1 - ResNet-50onednn: Deconvolution Batch shapes_3d - u8s8f32 - CPUonednn: Recurrent Neural Network Inference - bf16bf16bf16 - CPUonednn: Recurrent Neural Network Inference - f32 - CPUonednn: Recurrent Neural Network Inference - u8s8f32 - CPUonednn: Deconvolution Batch shapes_3d - f32 - CPUnumpy: pytorch: CPU - 1 - Efficientnet_v2_lpybench: Total For Average Test Timesonednn: Convolution Batch Shapes Auto - f32 - CPUonednn: Convolution Batch Shapes Auto - u8s8f32 - CPUpytorch: CPU - 32 - ResNet-50c6i.2xlargem7i-flex.2xlargec7a.2xlarge1.772252.0446.69866.53610.5912.661822.26179.534.044.066.386.362.0383634.716633.192015.9610.5726.781.772862501.502496.842492.298.09803374.997.9910005.936696.2439115.818.43474.461.7635016.57241.218.1000753.8774.205.475.387.557.361.0163273.110753.6899318.8412.1029.891.0032692318.312382.462389.647.94878438.258.997365.842617.1946218.345.16774.236.902009.76409.785.0143754.5173.348.718.7513.0312.951.094293.123393.2824531.8920.0050.161.264051478.511482.351480.935.10330590.1011.748877.351017.7461931.29OpenBenchmarking.org

OpenVINO

Model: Face Detection FP16 - Device: CPU

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2023.2.devModel: Face Detection FP16 - Device: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge246810SE +/- 0.01, N = 3SE +/- 0.10, N = 3SE +/- 0.00, N = 31.778.435.161. (CXX) g++ options: -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -pie

OpenVINO

Model: Face Detection FP16 - Device: CPU

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2023.2.devModel: Face Detection FP16 - Device: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge5001000150020002500SE +/- 3.74, N = 3SE +/- 5.97, N = 3SE +/- 0.28, N = 32252.04474.46774.23MIN: 2202.18 / MAX: 2317.24MIN: 427.39 / MAX: 558.54MIN: 767.78 / MAX: 795.491. (CXX) g++ options: -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -pie

oneDNN

Harness: Deconvolution Batch shapes_1d - Data Type: bf16bf16bf16 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.3Harness: Deconvolution Batch shapes_1d - Data Type: bf16bf16bf16 - Engine: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge1122334455SE +/- 0.01032, N = 3SE +/- 0.00827, N = 3SE +/- 0.00810, N = 346.698601.763506.90200MIN: 46.421. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl

OpenVINO

Model: Face Detection FP16-INT8 - Device: CPU

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2023.2.devModel: Face Detection FP16-INT8 - Device: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge48121620SE +/- 0.04, N = 3SE +/- 0.16, N = 3SE +/- 0.00, N = 36.5316.579.761. (CXX) g++ options: -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -pie

OpenVINO

Model: Face Detection FP16-INT8 - Device: CPU

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2023.2.devModel: Face Detection FP16-INT8 - Device: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge130260390520650SE +/- 3.28, N = 3SE +/- 2.35, N = 3SE +/- 0.07, N = 3610.59241.21409.78MIN: 497.95 / MAX: 643.84MIN: 91.81 / MAX: 374.94MIN: 407.35 / MAX: 424.191. (CXX) g++ options: -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -pie

oneDNN

Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.3Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge3691215SE +/- 0.07954, N = 3SE +/- 0.01302, N = 3SE +/- 0.00229, N = 312.661808.100075.01437MIN: 10.8MIN: 6.7MIN: 4.931. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl

OpenVINO

Model: Machine Translation EN To DE FP16 - Device: CPU

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2023.2.devModel: Machine Translation EN To DE FP16 - Device: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge1224364860SE +/- 0.06, N = 3SE +/- 0.19, N = 3SE +/- 0.02, N = 322.2653.8754.511. (CXX) g++ options: -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -pie

OpenVINO

Model: Machine Translation EN To DE FP16 - Device: CPU

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2023.2.devModel: Machine Translation EN To DE FP16 - Device: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge4080120160200SE +/- 0.44, N = 3SE +/- 0.25, N = 3SE +/- 0.02, N = 3179.5374.2073.34MIN: 100.26 / MAX: 344MIN: 34.84 / MAX: 96.18MIN: 65.17 / MAX: 80.621. (CXX) g++ options: -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -pie

PyTorch

Device: CPU - Batch Size: 32 - Model: Efficientnet_v2_l

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.1Device: CPU - Batch Size: 32 - Model: Efficientnet_v2_lc6i.2xlargem7i-flex.2xlargec7a.2xlarge246810SE +/- 0.02, N = 3SE +/- 0.01, N = 3SE +/- 0.02, N = 34.045.478.71MIN: 3.5 / MAX: 4.36MIN: 2.31 / MAX: 6.19MIN: 5.53 / MAX: 8.82

PyTorch

Device: CPU - Batch Size: 16 - Model: Efficientnet_v2_l

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.1Device: CPU - Batch Size: 16 - Model: Efficientnet_v2_lc6i.2xlargem7i-flex.2xlargec7a.2xlarge246810SE +/- 0.03, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 34.065.388.75MIN: 3.29 / MAX: 4.32MIN: 2.12 / MAX: 6.08MIN: 5.72 / MAX: 8.84

PyTorch

Device: CPU - Batch Size: 32 - Model: ResNet-152

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.1Device: CPU - Batch Size: 32 - Model: ResNet-152c6i.2xlargem7i-flex.2xlargec7a.2xlarge3691215SE +/- 0.01, N = 3SE +/- 0.02, N = 3SE +/- 0.05, N = 36.387.5513.03MIN: 3.7 / MAX: 6.57MIN: 2.93 / MAX: 8.62MIN: 10.5 / MAX: 13.2

PyTorch

Device: CPU - Batch Size: 16 - Model: ResNet-152

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.1Device: CPU - Batch Size: 16 - Model: ResNet-152c6i.2xlargem7i-flex.2xlargec7a.2xlarge3691215SE +/- 0.05, N = 3SE +/- 0.08, N = 5SE +/- 0.09, N = 36.367.3612.95MIN: 5.43 / MAX: 6.61MIN: 2.24 / MAX: 8.66MIN: 4.2 / MAX: 13.21

oneDNN

Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.3Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge0.45860.91721.37581.83442.293SE +/- 0.002826, N = 3SE +/- 0.013619, N = 15SE +/- 0.002971, N = 32.0383601.0163271.094290MIN: 1.97MIN: 0.79MIN: 1.081. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl

oneDNN

Harness: Deconvolution Batch shapes_3d - Data Type: bf16bf16bf16 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.3Harness: Deconvolution Batch shapes_3d - Data Type: bf16bf16bf16 - Engine: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge816243240SE +/- 0.00790, N = 3SE +/- 0.01261, N = 3SE +/- 0.00414, N = 334.716603.110753.12339MIN: 34.61. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl

oneDNN

Harness: Convolution Batch Shapes Auto - Data Type: bf16bf16bf16 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.3Harness: Convolution Batch Shapes Auto - Data Type: bf16bf16bf16 - Engine: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge816243240SE +/- 0.00337, N = 3SE +/- 0.03984, N = 3SE +/- 0.00275, N = 333.192003.689933.28245MIN: 33.131. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl

PyTorch

Device: CPU - Batch Size: 16 - Model: ResNet-50

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.1Device: CPU - Batch Size: 16 - Model: ResNet-50c6i.2xlargem7i-flex.2xlargec7a.2xlarge714212835SE +/- 0.12, N = 3SE +/- 0.17, N = 3SE +/- 0.24, N = 315.9618.8431.89MIN: 11.65 / MAX: 17.13MIN: 4.37 / MAX: 21.77MIN: 23.54 / MAX: 32.58

PyTorch

Device: CPU - Batch Size: 1 - Model: ResNet-152

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.1Device: CPU - Batch Size: 1 - Model: ResNet-152c6i.2xlargem7i-flex.2xlargec7a.2xlarge510152025SE +/- 0.01, N = 3SE +/- 0.10, N = 3SE +/- 0.05, N = 310.5712.1020.00MIN: 9.04 / MAX: 10.77MIN: 2.89 / MAX: 13.92MIN: 15.6 / MAX: 20.36

PyTorch

Device: CPU - Batch Size: 1 - Model: ResNet-50

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.1Device: CPU - Batch Size: 1 - Model: ResNet-50c6i.2xlargem7i-flex.2xlargec7a.2xlarge1122334455SE +/- 0.13, N = 3SE +/- 0.05, N = 3SE +/- 0.32, N = 326.7829.8950.16MIN: 13.67 / MAX: 27.8MIN: 7.96 / MAX: 34.56MIN: 33.27 / MAX: 51.43

oneDNN

Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.3Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge0.39890.79781.19671.59561.9945SE +/- 0.013170, N = 3SE +/- 0.011545, N = 4SE +/- 0.004187, N = 31.7728601.0032691.264050MIN: 1.73MIN: 0.84MIN: 1.251. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl

oneDNN

Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.3Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge5001000150020002500SE +/- 2.60, N = 3SE +/- 15.72, N = 3SE +/- 1.00, N = 32501.502318.311478.51MIN: 2476.86MIN: 2205.71MIN: 1472.491. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl

oneDNN

Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.3Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge5001000150020002500SE +/- 7.01, N = 3SE +/- 26.45, N = 4SE +/- 1.52, N = 32496.842382.461482.35MIN: 2465.66MIN: 2219.69MIN: 1476.161. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl

oneDNN

Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.3Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge5001000150020002500SE +/- 2.42, N = 3SE +/- 24.80, N = 3SE +/- 2.09, N = 32492.292389.641480.93MIN: 2460.06MIN: 2260.99MIN: 1474.241. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl

oneDNN

Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.3Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge246810SE +/- 0.01712, N = 3SE +/- 0.08711, N = 3SE +/- 0.00261, N = 38.098037.948785.10330MIN: 8.02MIN: 6.79MIN: 5.051. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl

Numpy Benchmark

OpenBenchmarking.orgScore, More Is BetterNumpy Benchmarkc6i.2xlargem7i-flex.2xlargec7a.2xlarge130260390520650SE +/- 1.37, N = 3SE +/- 3.53, N = 3SE +/- 1.02, N = 3374.99438.25590.10

PyTorch

Device: CPU - Batch Size: 1 - Model: Efficientnet_v2_l

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.1Device: CPU - Batch Size: 1 - Model: Efficientnet_v2_lc6i.2xlargem7i-flex.2xlargec7a.2xlarge3691215SE +/- 0.02, N = 3SE +/- 0.09, N = 12SE +/- 0.02, N = 37.998.9911.74MIN: 6.59 / MAX: 8.31MIN: 3.08 / MAX: 10.39MIN: 8.95 / MAX: 11.9

PyBench

Total For Average Test Times

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyBench 2018-02-16Total For Average Test Timesc6i.2xlargem7i-flex.2xlargec7a.2xlarge2004006008001000SE +/- 0.33, N = 3SE +/- 3.18, N = 3SE +/- 0.67, N = 31000736887

oneDNN

Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.3Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge246810SE +/- 0.03862, N = 3SE +/- 0.02435, N = 3SE +/- 0.01338, N = 35.936695.842617.35101MIN: 5.66MIN: 5MIN: 7.221. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl

oneDNN

Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.3Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPUc6i.2xlargem7i-flex.2xlargec7a.2xlarge246810SE +/- 0.08834, N = 3SE +/- 0.03815, N = 3SE +/- 0.01540, N = 36.243917.194627.74619MIN: 5.78MIN: 6.66MIN: 7.551. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl

PyTorch

Device: CPU - Batch Size: 32 - Model: ResNet-50

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.1Device: CPU - Batch Size: 32 - Model: ResNet-50c6i.2xlargem7i-flex.2xlargec7a.2xlarge714212835SE +/- 0.17, N = 4SE +/- 0.29, N = 15SE +/- 0.42, N = 1515.8118.3431.29MIN: 9.11 / MAX: 17.08MIN: 4.11 / MAX: 22.15MIN: 20.73 / MAX: 33.27

Phoronix Test Suite v10.8.5