icelake onx onednn

Intel Core i7-1065G7 testing with a Dell 06CDVY (1.0.9 BIOS) and Intel Iris Plus ICL GT2 3GB on Ubuntu 21.10 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 2204012-NE-ICELAKEON62
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

Limit displaying results to tests within:

HPC - High Performance Computing 2 Tests
Machine Learning 2 Tests

Statistics

Show Overall Harmonic Mean(s)
Show Overall Geometric Mean
Show Geometric Means Per-Suite/Category
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
Hide
Result
Result
Identifier
View Logs
Performance Per
Dollar
Date
Run
  Test
  Duration
A
March 31 2022
  2 Hours, 56 Minutes
B
March 31 2022
  3 Hours, 51 Minutes
C
March 31 2022
  4 Hours, 11 Minutes
Invert Hiding All Results Option
  3 Hours, 40 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):


icelake onx onednnOpenBenchmarking.orgPhoronix Test SuiteIntel Core i7-1065G7 @ 3.90GHz (4 Cores / 8 Threads)Dell 06CDVY (1.0.9 BIOS)Intel Ice Lake-LP DRAM16GBToshiba KBG40ZPZ512G NVMe 512GBIntel Iris Plus ICL GT2 3GB (1100MHz)Realtek ALC289Intel Ice Lake-LP PCH CNVi WiFiUbuntu 21.105.13.0-35-generic (x86_64)GNOME Shell 40.5X Server + Wayland4.6 Mesa 21.2.21.2.182GCC 11.2.0ext41920x1200ProcessorMotherboardChipsetMemoryDiskGraphicsAudioNetworkOSKernelDesktopDisplay ServerOpenGLVulkanCompilerFile-SystemScreen ResolutionIcelake Onx Onednn BenchmarksSystem Logs- Transparent Huge Pages: madvise- --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-bootstrap --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,brig,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-link-serialization=2 --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-targets=nvptx-none=/build/gcc-11-ZPT0kp/gcc-11-11.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-11-ZPT0kp/gcc-11-11.2.0/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-build-config=bootstrap-lto-lean --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -v - Scaling Governor: intel_pstate powersave (EPP: balance_performance) - CPU Microcode: 0xa6 - Thermald 2.4.6 - Python 3.9.7- itlb_multihit: KVM: Mitigation of VMX disabled + l1tf: Not affected + mds: Not affected + meltdown: 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 Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected

ABCResult OverviewPhoronix Test Suite100%112%123%135%147%oneDNNoneDNNoneDNNoneDNNoneDNNONNX RuntimeONNX RuntimeoneDNNONNX RuntimeoneDNNoneDNNONNX RuntimeONNX RuntimeONNX RuntimeONNX RuntimeONNX RuntimeoneDNNoneDNNONNX RuntimeoneDNNONNX RuntimeONNX RuntimeONNX RuntimeoneDNNoneDNNoneDNNoneDNNoneDNNoneDNNoneDNNoneDNNoneDNNoneDNNoneDNNoneDNNoneDNNIP Shapes 3D - bf16bf16bf16 - CPUIP Shapes 3D - u8s8f32 - CPUIP Shapes 1D - bf16bf16bf16 - CPUIP Shapes 1D - u8s8f32 - CPUC.B.S.A - f32 - CPUArcFace ResNet-100 - CPU - Standardyolov4 - CPU - StandardD.B.s - f32 - CPUsuper-resolution-10 - CPU - StandardD.B.s - bf16bf16bf16 - CPUD.B.s - f32 - CPUGPT-2 - CPU - Standardbertsquad-12 - CPU - Parallelfcn-resnet101-11 - CPU - Standardsuper-resolution-10 - CPU - ParallelGPT-2 - CPU - ParallelD.B.s - u8s8f32 - CPUIP Shapes 3D - f32 - CPUbertsquad-12 - CPU - StandardD.B.s - bf16bf16bf16 - CPUfcn-resnet101-11 - CPU - Parallelyolov4 - CPU - ParallelArcFace ResNet-100 - CPU - ParallelIP Shapes 1D - f32 - CPUR.N.N.T - u8s8f32 - CPUR.N.N.T - bf16bf16bf16 - CPUD.B.s - u8s8f32 - CPUC.B.S.A - u8s8f32 - CPUR.N.N.I - bf16bf16bf16 - CPUR.N.N.I - u8s8f32 - CPUR.N.N.T - f32 - CPUR.N.N.I - f32 - CPUM.M.B.S.T - f32 - CPUC.B.S.A - bf16bf16bf16 - CPUM.M.B.S.T - u8s8f32 - CPUM.M.B.S.T - bf16bf16bf16 - CPU

icelake onx onednnonednn: IP Shapes 1D - f32 - CPUonednn: IP Shapes 3D - f32 - CPUonednn: IP Shapes 1D - u8s8f32 - CPUonednn: IP Shapes 3D - u8s8f32 - CPUonednn: IP Shapes 1D - bf16bf16bf16 - CPUonednn: IP Shapes 3D - bf16bf16bf16 - CPUonednn: Convolution Batch Shapes Auto - f32 - CPUonednn: Deconvolution Batch shapes_1d - f32 - CPUonednn: Deconvolution Batch shapes_3d - f32 - CPUonednn: Convolution Batch Shapes Auto - u8s8f32 - CPUonednn: Deconvolution Batch shapes_1d - u8s8f32 - CPUonednn: Deconvolution Batch shapes_3d - u8s8f32 - CPUonednn: Recurrent Neural Network Training - f32 - CPUonednn: Recurrent Neural Network Inference - f32 - CPUonednn: Recurrent Neural Network Training - u8s8f32 - CPUonednn: Convolution Batch Shapes Auto - bf16bf16bf16 - CPUonednn: Deconvolution Batch shapes_1d - bf16bf16bf16 - CPUonednn: Deconvolution Batch shapes_3d - bf16bf16bf16 - CPUonednn: Recurrent Neural Network Inference - u8s8f32 - CPUonednn: Matrix Multiply Batch Shapes Transformer - f32 - CPUonednn: Recurrent Neural Network Training - bf16bf16bf16 - CPUonednn: Recurrent Neural Network Inference - bf16bf16bf16 - CPUonednn: Matrix Multiply Batch Shapes Transformer - u8s8f32 - CPUonednn: Matrix Multiply Batch Shapes Transformer - bf16bf16bf16 - CPUonnx: GPT-2 - CPU - Parallelonnx: GPT-2 - CPU - Standardonnx: yolov4 - CPU - Parallelonnx: yolov4 - CPU - Standardonnx: bertsquad-12 - CPU - Parallelonnx: bertsquad-12 - CPU - Standardonnx: fcn-resnet101-11 - CPU - Parallelonnx: fcn-resnet101-11 - CPU - Standardonnx: ArcFace ResNet-100 - CPU - Parallelonnx: ArcFace ResNet-100 - CPU - Standardonnx: super-resolution-10 - CPU - Parallelonnx: super-resolution-10 - CPU - StandardABC9.847076.373952.118122.5340623.07687.1520812.852420.406812.670011.23363.942923.3731311554.25885.8411553.664.782683.831046.05385941.114.7797211761.36052.081.9532815.3859337943881341451602282224398519124314799.483676.376842.914393.6669932.963210.4823615.414622.887113.882311.62714.081573.5947611926.46050.4611980.665.835088.537850.93446104.084.8404312010.36101.821.9425515.4670337943481361661702232322411535128815589.486816.784552.697973.6924532.500310.4988215.058823.065712.685811.56153.987553.4966311573.35906.7411587.765.102384.193446.04455905.074.7221411588.05901.881.9362315.360936214756140166175237232441446713351379OpenBenchmarking.org

oneDNN

This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of Intel oneAPI. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: IP Shapes 1D - Data Type: f32 - Engine: CPUABC3691215SE +/- 0.01786, N = 3SE +/- 0.07263, N = 3SE +/- 0.03560, N = 39.847079.483679.48681MIN: 8.52MIN: 8.22MIN: 8.421. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: IP Shapes 1D - Data Type: f32 - Engine: CPUABC3691215Min: 9.82 / Avg: 9.85 / Max: 9.88Min: 9.36 / Avg: 9.48 / Max: 9.61Min: 9.45 / Avg: 9.49 / Max: 9.561. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: IP Shapes 3D - Data Type: f32 - Engine: CPUABC246810SE +/- 0.01595, N = 3SE +/- 0.02663, N = 3SE +/- 0.01646, N = 36.373956.376846.78455MIN: 6.06MIN: 6.07MIN: 6.471. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: IP Shapes 3D - Data Type: f32 - Engine: CPUABC3691215Min: 6.34 / Avg: 6.37 / Max: 6.39Min: 6.32 / Avg: 6.38 / Max: 6.41Min: 6.76 / Avg: 6.78 / Max: 6.821. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: IP Shapes 1D - Data Type: u8s8f32 - Engine: CPUABC0.65571.31141.96712.62283.2785SE +/- 0.03031, N = 3SE +/- 0.17277, N = 12SE +/- 0.16535, N = 122.118122.914392.69797MIN: 1.88MIN: 1.82MIN: 1.771. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: IP Shapes 1D - Data Type: u8s8f32 - Engine: CPUABC246810Min: 2.06 / Avg: 2.12 / Max: 2.15Min: 2 / Avg: 2.91 / Max: 3.33Min: 2.01 / Avg: 2.7 / Max: 3.341. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPUABC0.83081.66162.49243.32324.154SE +/- 0.00331, N = 3SE +/- 0.09454, N = 12SE +/- 0.08757, N = 152.534063.666993.69245MIN: 2.4MIN: 2.39MIN: 2.491. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPUABC246810Min: 2.53 / Avg: 2.53 / Max: 2.54Min: 2.63 / Avg: 3.67 / Max: 3.79Min: 2.6 / Avg: 3.69 / Max: 3.851. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: IP Shapes 1D - Data Type: bf16bf16bf16 - Engine: CPUABC816243240SE +/- 0.13, N = 3SE +/- 0.85, N = 12SE +/- 0.78, N = 1523.0832.9632.50MIN: 22.24MIN: 22.82MIN: 22.841. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: IP Shapes 1D - Data Type: bf16bf16bf16 - Engine: CPUABC714212835Min: 22.86 / Avg: 23.08 / Max: 23.3Min: 23.59 / Avg: 32.96 / Max: 34.1Min: 23.79 / Avg: 32.5 / Max: 33.761. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: IP Shapes 3D - Data Type: bf16bf16bf16 - Engine: CPUABC3691215SE +/- 0.05916, N = 15SE +/- 0.21402, N = 12SE +/- 0.19453, N = 127.1520810.4823610.49882MIN: 5.68MIN: 5.93MIN: 6.561. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: IP Shapes 3D - Data Type: bf16bf16bf16 - Engine: CPUABC3691215Min: 6.44 / Avg: 7.15 / Max: 7.29Min: 8.23 / Avg: 10.48 / Max: 11.16Min: 8.36 / Avg: 10.5 / Max: 10.81. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPUABC48121620SE +/- 0.02, N = 3SE +/- 0.18, N = 12SE +/- 0.25, N = 1512.8515.4115.06MIN: 12.43MIN: 11.77MIN: 11.721. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPUABC48121620Min: 12.81 / Avg: 12.85 / Max: 12.88Min: 13.41 / Avg: 15.41 / Max: 15.9Min: 12.66 / Avg: 15.06 / Max: 15.61. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPUABC612182430SE +/- 0.55, N = 15SE +/- 0.24, N = 3SE +/- 0.20, N = 820.4122.8923.07MIN: 14.79MIN: 19.9MIN: 19.731. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPUABC510152025Min: 15.91 / Avg: 20.41 / Max: 22.92Min: 22.41 / Avg: 22.89 / Max: 23.18Min: 21.79 / Avg: 23.07 / Max: 23.741. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPUABC48121620SE +/- 0.04, N = 3SE +/- 0.21, N = 12SE +/- 0.02, N = 312.6713.8812.69MIN: 12.39MIN: 12.05MIN: 12.411. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPUABC48121620Min: 12.61 / Avg: 12.67 / Max: 12.73Min: 12.44 / Avg: 13.88 / Max: 14.57Min: 12.65 / Avg: 12.69 / Max: 12.721. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPUABC3691215SE +/- 0.02, N = 3SE +/- 0.15, N = 3SE +/- 0.13, N = 411.2311.6311.56MIN: 10.98MIN: 11.05MIN: 11.011. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPUABC3691215Min: 11.2 / Avg: 11.23 / Max: 11.26Min: 11.33 / Avg: 11.63 / Max: 11.84Min: 11.19 / Avg: 11.56 / Max: 11.721. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPUABC0.91841.83682.75523.67364.592SE +/- 0.04445, N = 15SE +/- 0.03306, N = 9SE +/- 0.04434, N = 53.942924.081573.98755MIN: 2.89MIN: 3.27MIN: 3.261. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPUABC246810Min: 3.41 / Avg: 3.94 / Max: 4.03Min: 3.86 / Avg: 4.08 / Max: 4.2Min: 3.81 / Avg: 3.99 / Max: 4.041. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPUABC0.80881.61762.42643.23524.044SE +/- 0.06649, N = 15SE +/- 0.05792, N = 15SE +/- 0.05316, N = 153.373133.594763.49663MIN: 2.84MIN: 2.84MIN: 2.851. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPUABC246810Min: 3.01 / Avg: 3.37 / Max: 3.62Min: 3.04 / Avg: 3.59 / Max: 3.93Min: 3.03 / Avg: 3.5 / Max: 3.611. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPUABC3K6K9K12K15KSE +/- 9.62, N = 3SE +/- 96.19, N = 3SE +/- 21.52, N = 311554.211926.411573.3MIN: 11456.1MIN: 11490.5MIN: 11441.51. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPUABC2K4K6K8K10KMin: 11542.9 / Avg: 11554.17 / Max: 11573.3Min: 11765.5 / Avg: 11926.43 / Max: 12098.2Min: 11530.3 / Avg: 11573.33 / Max: 11595.71. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPUABC13002600390052006500SE +/- 2.20, N = 3SE +/- 23.35, N = 3SE +/- 4.21, N = 35885.846050.465906.74MIN: 5806.32MIN: 5857.07MIN: 5822.821. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPUABC11002200330044005500Min: 5883.4 / Avg: 5885.84 / Max: 5890.23Min: 6024.33 / Avg: 6050.46 / Max: 6097.04Min: 5898.33 / Avg: 5906.74 / Max: 5911.481. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPUABC3K6K9K12K15KSE +/- 9.35, N = 3SE +/- 66.98, N = 3SE +/- 3.03, N = 311553.611980.611587.7MIN: 11444.3MIN: 11507.8MIN: 11494.61. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPUABC2K4K6K8K10KMin: 11534.9 / Avg: 11553.6 / Max: 11563.4Min: 11849.7 / Avg: 11980.6 / Max: 12070.7Min: 11581.7 / Avg: 11587.73 / Max: 11591.21. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Convolution Batch Shapes Auto - Data Type: bf16bf16bf16 - Engine: CPUABC1530456075SE +/- 1.77, N = 12SE +/- 1.62, N = 12SE +/- 1.80, N = 1264.7865.8465.10MIN: 44.54MIN: 44.54MIN: 44.541. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Convolution Batch Shapes Auto - Data Type: bf16bf16bf16 - Engine: CPUABC1326395265Min: 45.37 / Avg: 64.78 / Max: 66.68Min: 48.54 / Avg: 65.83 / Max: 71.4Min: 45.28 / Avg: 65.1 / Max: 67.111. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Deconvolution Batch shapes_1d - Data Type: bf16bf16bf16 - Engine: CPUABC20406080100SE +/- 1.19, N = 3SE +/- 0.70, N = 15SE +/- 0.94, N = 383.8388.5484.19MIN: 74.16MIN: 77.04MIN: 76.261. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Deconvolution Batch shapes_1d - Data Type: bf16bf16bf16 - Engine: CPUABC20406080100Min: 81.45 / Avg: 83.83 / Max: 85.11Min: 82.28 / Avg: 88.54 / Max: 93.59Min: 82.33 / Avg: 84.19 / Max: 85.371. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Deconvolution Batch shapes_3d - Data Type: bf16bf16bf16 - Engine: CPUABC1122334455SE +/- 0.14, N = 3SE +/- 0.60, N = 15SE +/- 0.08, N = 346.0550.9346.04MIN: 45.49MIN: 45.44MIN: 45.441. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Deconvolution Batch shapes_3d - Data Type: bf16bf16bf16 - Engine: CPUABC1020304050Min: 45.79 / Avg: 46.05 / Max: 46.25Min: 46.13 / Avg: 50.93 / Max: 55.59Min: 45.91 / Avg: 46.04 / Max: 46.181. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPUABC13002600390052006500SE +/- 19.05, N = 3SE +/- 56.07, N = 3SE +/- 5.63, N = 35941.116104.085905.07MIN: 5805.39MIN: 5822.26MIN: 5827.861. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPUABC11002200330044005500Min: 5905.7 / Avg: 5941.11 / Max: 5970.97Min: 6023.05 / Avg: 6104.08 / Max: 6211.74Min: 5893.99 / Avg: 5905.07 / Max: 5912.331. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPUABC1.08912.17823.26734.35645.4455SE +/- 0.01354, N = 3SE +/- 0.05235, N = 4SE +/- 0.00888, N = 34.779724.840434.72214MIN: 3.62MIN: 3.75MIN: 3.681. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPUABC246810Min: 4.76 / Avg: 4.78 / Max: 4.8Min: 4.71 / Avg: 4.84 / Max: 4.96Min: 4.71 / Avg: 4.72 / Max: 4.741. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPUABC3K6K9K12K15KSE +/- 66.05, N = 3SE +/- 124.39, N = 5SE +/- 2.48, N = 311761.312010.311588.0MIN: 11453.7MIN: 10693.2MIN: 11492.61. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPUABC2K4K6K8K10KMin: 11629.2 / Avg: 11761.3 / Max: 11827.7Min: 11670.1 / Avg: 12010.26 / Max: 12294.1Min: 11583.7 / Avg: 11588 / Max: 11592.31. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPUABC13002600390052006500SE +/- 17.19, N = 3SE +/- 24.13, N = 3SE +/- 5.21, N = 36052.086101.825901.88MIN: 5839.98MIN: 5858.61MIN: 5818.551. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPUABC11002200330044005500Min: 6021.56 / Avg: 6052.08 / Max: 6081.06Min: 6053.61 / Avg: 6101.82 / Max: 6127.69Min: 5894.61 / Avg: 5901.88 / Max: 5911.981. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPUABC0.43950.8791.31851.7582.1975SE +/- 0.01181, N = 3SE +/- 0.01138, N = 3SE +/- 0.00369, N = 31.953281.942551.93623MIN: 1.62MIN: 1.62MIN: 1.621. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPUABC246810Min: 1.93 / Avg: 1.95 / Max: 1.97Min: 1.93 / Avg: 1.94 / Max: 1.96Min: 1.93 / Avg: 1.94 / Max: 1.941. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Matrix Multiply Batch Shapes Transformer - Data Type: bf16bf16bf16 - Engine: CPUABC48121620SE +/- 0.06, N = 3SE +/- 0.02, N = 3SE +/- 0.02, N = 315.3915.4715.36MIN: 13.63MIN: 13.36MIN: 13.521. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.6Harness: Matrix Multiply Batch Shapes Transformer - Data Type: bf16bf16bf16 - Engine: CPUABC48121620Min: 15.32 / Avg: 15.39 / Max: 15.51Min: 15.42 / Avg: 15.47 / Max: 15.5Min: 15.33 / Avg: 15.36 / Max: 15.41. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl

ONNX Runtime

ONNX Runtime is developed by Microsoft and partners as a open-source, cross-platform, high performance machine learning inferencing and training accelerator. This test profile runs the ONNX Runtime with various models available from the ONNX Zoo. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: GPT-2 - Device: CPU - Executor: ParallelABC8001600240032004000SE +/- 31.81, N = 12SE +/- 34.69, N = 5SE +/- 27.27, N = 113379337936211. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: GPT-2 - Device: CPU - Executor: ParallelABC6001200180024003000Min: 3268.5 / Avg: 3379.17 / Max: 3643Min: 3315 / Avg: 3378.5 / Max: 3513Min: 3560.5 / Avg: 3620.82 / Max: 38191. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: GPT-2 - Device: CPU - Executor: StandardABC10002000300040005000SE +/- 25.37, N = 3SE +/- 53.21, N = 11SE +/- 25.46, N = 34388434847561. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: GPT-2 - Device: CPU - Executor: StandardABC8001600240032004000Min: 4357 / Avg: 4387.67 / Max: 4438Min: 3821 / Avg: 4347.77 / Max: 4448.5Min: 4719 / Avg: 4756.33 / Max: 48051. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: yolov4 - Device: CPU - Executor: ParallelABC306090120150SE +/- 0.87, N = 3SE +/- 1.00, N = 3SE +/- 0.17, N = 31341361401. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: yolov4 - Device: CPU - Executor: ParallelABC306090120150Min: 132.5 / Avg: 134 / Max: 135.5Min: 134.5 / Avg: 135.5 / Max: 137.5Min: 139.5 / Avg: 139.83 / Max: 1401. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: yolov4 - Device: CPU - Executor: StandardABC4080120160200SE +/- 7.24, N = 12SE +/- 4.41, N = 12SE +/- 6.99, N = 121451661661. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: yolov4 - Device: CPU - Executor: StandardABC306090120150Min: 120 / Avg: 144.75 / Max: 170.5Min: 128 / Avg: 165.71 / Max: 175Min: 124.5 / Avg: 165.96 / Max: 1811. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: bertsquad-12 - Device: CPU - Executor: ParallelABC4080120160200SE +/- 1.61, N = 3SE +/- 2.32, N = 12SE +/- 2.35, N = 121601701751. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: bertsquad-12 - Device: CPU - Executor: ParallelABC306090120150Min: 157 / Avg: 159.5 / Max: 162.5Min: 162 / Avg: 170.08 / Max: 187Min: 168 / Avg: 174.54 / Max: 191.51. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: bertsquad-12 - Device: CPU - Executor: StandardABC50100150200250SE +/- 1.45, N = 3SE +/- 7.41, N = 12SE +/- 7.06, N = 102282232371. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: bertsquad-12 - Device: CPU - Executor: StandardABC4080120160200Min: 225 / Avg: 227.67 / Max: 230Min: 167.5 / Avg: 222.83 / Max: 236Min: 173.5 / Avg: 236.6 / Max: 249.51. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: fcn-resnet101-11 - Device: CPU - Executor: ParallelABC612182430SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.17, N = 32223231. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: fcn-resnet101-11 - Device: CPU - Executor: ParallelABC510152025Min: 21.5 / Avg: 21.5 / Max: 21.5Min: 22.5 / Avg: 22.5 / Max: 22.5Min: 22.5 / Avg: 22.83 / Max: 231. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: fcn-resnet101-11 - Device: CPU - Executor: StandardABC612182430SE +/- 0.94, N = 9SE +/- 1.28, N = 12SE +/- 1.38, N = 122422241. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: fcn-resnet101-11 - Device: CPU - Executor: StandardABC612182430Min: 17 / Avg: 24.06 / Max: 26Min: 17.5 / Avg: 22.42 / Max: 27Min: 17.5 / Avg: 24.21 / Max: 281. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: ArcFace ResNet-100 - Device: CPU - Executor: ParallelABC90180270360450SE +/- 3.38, N = 3SE +/- 0.93, N = 3SE +/- 3.18, N = 33984114141. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: ArcFace ResNet-100 - Device: CPU - Executor: ParallelABC70140210280350Min: 393.5 / Avg: 397.83 / Max: 404.5Min: 410 / Avg: 411.17 / Max: 413Min: 408 / Avg: 414.33 / Max: 4181. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: ArcFace ResNet-100 - Device: CPU - Executor: StandardABC120240360480600SE +/- 1.32, N = 3SE +/- 2.17, N = 3SE +/- 24.61, N = 125195354671. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: ArcFace ResNet-100 - Device: CPU - Executor: StandardABC90180270360450Min: 516 / Avg: 518.5 / Max: 520.5Min: 532.5 / Avg: 534.67 / Max: 539Min: 354 / Avg: 466.63 / Max: 532.51. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: super-resolution-10 - Device: CPU - Executor: ParallelABC30060090012001500SE +/- 16.80, N = 3SE +/- 4.62, N = 3SE +/- 5.36, N = 31243128813351. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: super-resolution-10 - Device: CPU - Executor: ParallelABC2004006008001000Min: 1217 / Avg: 1243.17 / Max: 1274.5Min: 1280 / Avg: 1288 / Max: 1296Min: 1327.5 / Avg: 1335.17 / Max: 1345.51. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: super-resolution-10 - Device: CPU - Executor: StandardABC30060090012001500SE +/- 17.45, N = 3SE +/- 8.31, N = 3SE +/- 66.66, N = 121479155813791. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Minute, More Is BetterONNX Runtime 1.11Model: super-resolution-10 - Device: CPU - Executor: StandardABC30060090012001500Min: 1459.5 / Avg: 1478.67 / Max: 1513.5Min: 1541.5 / Avg: 1558 / Max: 1568Min: 1165.5 / Avg: 1379.08 / Max: 16501. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt

36 Results Shown

oneDNN:
  IP Shapes 1D - f32 - CPU
  IP Shapes 3D - f32 - CPU
  IP Shapes 1D - u8s8f32 - CPU
  IP Shapes 3D - u8s8f32 - CPU
  IP Shapes 1D - bf16bf16bf16 - CPU
  IP Shapes 3D - bf16bf16bf16 - CPU
  Convolution Batch Shapes Auto - f32 - CPU
  Deconvolution Batch shapes_1d - f32 - CPU
  Deconvolution Batch shapes_3d - f32 - CPU
  Convolution Batch Shapes Auto - u8s8f32 - CPU
  Deconvolution Batch shapes_1d - u8s8f32 - CPU
  Deconvolution Batch shapes_3d - u8s8f32 - CPU
  Recurrent Neural Network Training - f32 - CPU
  Recurrent Neural Network Inference - f32 - CPU
  Recurrent Neural Network Training - u8s8f32 - CPU
  Convolution Batch Shapes Auto - bf16bf16bf16 - CPU
  Deconvolution Batch shapes_1d - bf16bf16bf16 - CPU
  Deconvolution Batch shapes_3d - bf16bf16bf16 - CPU
  Recurrent Neural Network Inference - u8s8f32 - CPU
  Matrix Multiply Batch Shapes Transformer - f32 - CPU
  Recurrent Neural Network Training - bf16bf16bf16 - CPU
  Recurrent Neural Network Inference - bf16bf16bf16 - CPU
  Matrix Multiply Batch Shapes Transformer - u8s8f32 - CPU
  Matrix Multiply Batch Shapes Transformer - bf16bf16bf16 - CPU
ONNX Runtime:
  GPT-2 - CPU - Parallel
  GPT-2 - CPU - Standard
  yolov4 - CPU - Parallel
  yolov4 - CPU - Standard
  bertsquad-12 - CPU - Parallel
  bertsquad-12 - CPU - Standard
  fcn-resnet101-11 - CPU - Parallel
  fcn-resnet101-11 - CPU - Standard
  ArcFace ResNet-100 - CPU - Parallel
  ArcFace ResNet-100 - CPU - Standard
  super-resolution-10 - CPU - Parallel
  super-resolution-10 - CPU - Standard