AMD Ryzen 9 7950X / Ryzen 9 9950X AVX-512 Comparison

AMD Ryzen 9 7950X and Ryzen 9 9950X AVX-512 on/off comparison benchmarks by Michael Larabel for a future article.

Compare your own system(s) to this result file with the Phoronix Test Suite by running the command: phoronix-test-suite benchmark 2408145-NE-RYZEN999549
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

Limit displaying results to tests within:

CPU Massive 5 Tests
Creator Workloads 6 Tests
HPC - High Performance Computing 12 Tests
Machine Learning 9 Tests
Molecular Dynamics 2 Tests
Multi-Core 8 Tests
NVIDIA GPU Compute 2 Tests
Intel oneAPI 6 Tests
Python Tests 5 Tests
Raytracing 2 Tests
Renderers 2 Tests
Scientific Computing 2 Tests
Server CPU Tests 3 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
Disable Color Branding
Prefer Vertical Bar Graphs
No Box Plots
On Line Graphs With Missing Data, Connect The Line Gaps

Multi-Way Comparison

Condense Comparison
Transpose Comparison

Table

Show Detailed System Result Table

Run Management

Highlight
Result
Hide
Result
Result
Identifier
Performance Per
Dollar
Date
Run
  Test
  Duration
Ryzen 9 9950X: AVX-512 On
August 11
  5 Hours, 44 Minutes
Ryzen 9 9950X: AVX-512 Off
August 12
  5 Hours, 22 Minutes
Ryzen 9 7950X: AVX-512 On
August 12
  5 Hours, 35 Minutes
Ryzen 9 7950X: AVX-512 Off
August 12
  5 Hours, 39 Minutes
Invert Hiding All Results Option
  5 Hours, 35 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):


AMD Ryzen 9 7950X / Ryzen 9 9950X AVX-512 ComparisonOpenBenchmarking.orgPhoronix Test SuiteAMD Ryzen 9 9950X 16-Core @ 5.75GHz (16 Cores / 32 Threads)AMD Ryzen 9 7950X 16-Core @ 5.88GHz (16 Cores / 32 Threads)ASUS ROG STRIX X670E-E GAMING WIFI (2124 BIOS)AMD Device 14d82 x 16GB DDR5-6000MT/s G Skill F5-6000J3038F16G2000GB Corsair MP700 PROAMD Radeon RX 7900 GRE 16GBAMD Navi 31 HDMI/DPDELL U2723QEIntel I225-V + Intel Wi-Fi 6EUbuntu 24.046.10.0-phx (x86_64)GNOME Shell 46.0X Server + Wayland4.6 Mesa 24.2~git2406040600.8112d4~oibaf~n (git-8112d44 2024-06-04 noble-oibaf-ppa) (LLVM 17.0.6 DRM 3.57)GCC 13.2.0ext43840x2160ProcessorsMotherboardChipsetMemoryDiskGraphicsAudioMonitorNetworkOSKernelDesktopDisplay ServerOpenGLCompilerFile-SystemScreen ResolutionAMD Ryzen 9 7950X / Ryzen 9 9950X AVX-512 Comparison BenchmarksSystem Logs- Transparent Huge Pages: madvise- --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-backtrace --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-13-uJ7kn6/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-uJ7kn6/gcc-13-13.2.0/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-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 - Ryzen 9 9950X: AVX-512 On: Scaling Governor: amd-pstate-epp powersave (EPP: balance_performance) - CPU Microcode: 0xb40401a- Ryzen 9 9950X: AVX-512 Off: Scaling Governor: amd-pstate-epp powersave (EPP: balance_performance) - CPU Microcode: 0xb40401a- Ryzen 9 7950X: AVX-512 On: Scaling Governor: amd-pstate-epp powersave (EPP: balance_performance) - CPU Microcode: 0xa601206- Ryzen 9 7950X: AVX-512 Off: Scaling Governor: amd-pstate-epp powersave (EPP: balance_performance) - CPU Microcode: 0xa601206- Python 3.12.3- Ryzen 9 9950X: AVX-512 On: 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: 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 / Automatic IBRS; IBPB: conditional; STIBP: always-on; RSB filling; PBRSB-eIBRS: Not affected; BHI: Not affected + srbds: Not affected + tsx_async_abort: Not affected - Ryzen 9 9950X: AVX-512 Off: 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: 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 / Automatic IBRS; IBPB: conditional; STIBP: always-on; RSB filling; PBRSB-eIBRS: Not affected; BHI: Not affected + srbds: Not affected + tsx_async_abort: Not affected - Ryzen 9 7950X: AVX-512 On: 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; BHI: Not affected + srbds: Not affected + tsx_async_abort: Not affected - Ryzen 9 7950X: AVX-512 Off: 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; BHI: Not affected + srbds: Not affected + tsx_async_abort: Not affected

openvino: Weld Porosity Detection FP16 - CPUopenvino: Weld Porosity Detection FP16 - CPUopenvino: Face Detection FP16 - CPUopenvino: Face Detection FP16 - CPUopenvino: Age Gender Recognition Retail 0013 FP16 - CPUopenvino: Age Gender Recognition Retail 0013 FP16 - CPUopenvino: Person Re-Identification Retail FP16 - CPUopenvino: Handwritten English Recognition FP16 - CPUopenvino: Handwritten English Recognition FP16 - CPUopenvino: Person Re-Identification Retail FP16 - CPUopenvino: Person Vehicle Bike Detection FP16 - CPUopenvino: Person Vehicle Bike Detection FP16 - CPUtensorflow: CPU - 64 - ResNet-50openvino: Face Detection Retail FP16 - CPUopenvino: Face Detection Retail FP16 - CPUnamd: ATPase with 327,506 Atomsnamd: STMV with 1,066,628 Atomsopenvino: Machine Translation EN To DE FP16 - CPUopenvino: Machine Translation EN To DE FP16 - CPUospray: gravity_spheres_volume/dim_512/scivis/real_timeospray: gravity_spheres_volume/dim_512/ao/real_timeonednn: IP Shapes 3D - CPUminibude: OpenMP - BM1minibude: OpenMP - BM1onednn: Deconvolution Batch shapes_3d - CPUonnx: ResNet50 v1-12-int8 - CPU - Standardopenvino: Vehicle Detection FP16 - CPUopenvino: Vehicle Detection FP16 - CPUonednn: IP Shapes 1D - CPUopenvino: Person Detection FP16 - CPUopenvino: Person Detection FP16 - CPUopenvino: Road Segmentation ADAS FP16-INT8 - CPUopenvino: Road Segmentation ADAS FP16-INT8 - CPUospray: gravity_spheres_volume/dim_512/pathtracer/real_timesimdjson: PartialTweetssimdjson: DistinctUserIDonednn: Recurrent Neural Network Training - CPUsimdjson: TopTweetonnx: super-resolution-10 - CPU - Standardonednn: Recurrent Neural Network Inference - CPUopenvino: Noise Suppression Poconet-Like FP16 - CPUopenvino: Noise Suppression Poconet-Like FP16 - CPUospray-studio: 1 - 4K - 1 - Path Tracer - CPUospray-studio: 1 - 4K - 16 - Path Tracer - CPUospray-studio: 1 - 4K - 32 - Path Tracer - CPUospray-studio: 3 - 4K - 1 - Path Tracer - CPUospray-studio: 3 - 4K - 32 - Path Tracer - CPUospray-studio: 3 - 4K - 16 - Path Tracer - CPUonednn: Deconvolution Batch shapes_1d - CPUsimdjson: Kostyalczero: Eigenopenvkl: vklBenchmarkCPU ISPCnumpy: onnx: CaffeNet 12-int8 - CPU - Standardonnx: fcn-resnet101-11 - CPU - Standardtensorflow: CPU - 1 - ResNet-50pytorch: CPU - 1 - ResNet-50embree: Pathtracer ISPC - Asian Dragonembree: Pathtracer ISPC - Asian Dragon Objpytorch: CPU - 64 - ResNet-50ospray: particle_volume/pathtracer/real_timesimdjson: LargeRandlczero: BLASembree: Pathtracer ISPC - Crowny-cruncher: 1Bmnn: resnet-v2-50gromacs: MPI CPU - water_GMX50_barey-cruncher: 500Monnx: ArcFace ResNet-100 - CPU - Standardxnnpack: QU8MobileNetV2xnnpack: QU8MobileNetV3Largeonnx: super-resolution-10 - CPU - Standardonnx: ResNet50 v1-12-int8 - CPU - Standardonnx: ArcFace ResNet-100 - CPU - Standardonnx: fcn-resnet101-11 - CPU - Standardonnx: CaffeNet 12-int8 - CPU - StandardRyzen 9 9950XRyzen 9 7950X AVX-512 On AVX-512 Off AVX-512 On AVX-512 Off6.822327.1823.29342.810.2555883.542.5916.56962.923027.683.722133.7651.741.534938.983.483380.9870745.10177.119.532779.727003.0753077.1901929.7531.36386571.4688.77908.960.664875109.2973.12638.6712.5011.030012.3813.05759.93412.95227.646404.6978.431867.603225540221057983829124596634761.788467.472207651054.92872.9404.8548816.8285.1441.351535.507157.89277.2802.0722036.651516.49010.0483.1357.57055.507878511544.392531.7494218.0181205.9801.1459523.57675.856.801168.940.7519062.587.6746.11346.431034.5410.06792.0338.993.922015.671.697570.48492102.8677.696.469206.706626.1672138.469961.7272.59761447.23815.29521.701.0585560.69131.60476.8816.749.083769.9910.441258.8010.33151.023635.73213.181201.733594599581177124235138397703642.411926.431626561010.59803.4423.6003212.9870.1439.532734.081447.05261.1561.9020534.627220.23011.3172.6699.07349.806585611776.621282.2354020.0769277.7511.2440010.971454.5214.33556.000.436493.603.8720.61775.122059.524.771667.9244.082.283357.783.058770.8258157.56138.797.937308.076463.1150853.1221328.0502.38830419.18610.02795.701.00661101.5178.75481.0416.589.179629.4710.281245.2810.48144.777629.5989.361680.424004676111316054735154740792712.753466.01177647794.79695.7313.5477516.2374.4634.711129.546950.23239.2971.8117133.766516.02910.8752.7407.69647.852580710946.906902.3849820.8958281.8671.4368223.67674.776.871158.010.8017640.847.6148.98326.301048.4910.08791.2619.733.942015.991.493540.42614104.1176.774.208254.442336.6029338.401960.0372.72463299.66916.50483.591.2119562.80127.26359.2622.226.357557.207.591303.167.64137.126664.15413.571166.245131852771668056020195460994752.802514.80142507731.83605.7713.4163612.2263.2331.037826.716943.87210.8501.5816828.782818.05712.6552.5928.39547.847387011307.292543.3363920.8995292.7071.65024OpenBenchmarking.org

OpenVINO

This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Weld Porosity Detection FP16 - Device: CPUAVX-512 OnAVX-512 Off612182430SE +/- 0.00, N = 3SE +/- 0.02, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 36.8223.5710.9723.671. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Weld Porosity Detection FP16 - Device: CPUAVX-512 OnAVX-512 Off5001000150020002500SE +/- 1.07, N = 3SE +/- 0.69, N = 3SE +/- 0.51, N = 3SE +/- 0.19, N = 32327.18675.851454.52674.771. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Face Detection FP16 - Device: CPUAVX-512 OnAVX-512 Off612182430SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 323.296.8014.336.871. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Face Detection FP16 - Device: CPUAVX-512 OnAVX-512 Off30060090012001500SE +/- 0.13, N = 3SE +/- 3.79, N = 3SE +/- 0.04, N = 3SE +/- 0.43, N = 3342.811168.94556.001158.011. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Age Gender Recognition Retail 0013 FP16 - Device: CPUAVX-512 OnAVX-512 Off0.180.360.540.720.9SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 30.250.750.400.801. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Age Gender Recognition Retail 0013 FP16 - Device: CPUAVX-512 OnAVX-512 Off12K24K36K48K60KSE +/- 40.07, N = 3SE +/- 18.06, N = 3SE +/- 34.76, N = 3SE +/- 30.09, N = 355883.5419062.5836493.6017640.841. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Person Re-Identification Retail FP16 - Device: CPUAVX-512 OnAVX-512 Off246810SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 32.597.673.877.611. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Handwritten English Recognition FP16 - Device: CPUAVX-512 OnAVX-512 Off1122334455SE +/- 0.02, N = 3SE +/- 0.04, N = 3SE +/- 0.13, N = 3SE +/- 0.09, N = 316.5646.1120.6148.981. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Handwritten English Recognition FP16 - Device: CPUAVX-512 OnAVX-512 Off2004006008001000SE +/- 1.39, N = 3SE +/- 0.29, N = 3SE +/- 4.65, N = 3SE +/- 0.62, N = 3962.92346.43775.12326.301. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Person Re-Identification Retail FP16 - Device: CPUAVX-512 OnAVX-512 Off6001200180024003000SE +/- 2.87, N = 3SE +/- 0.29, N = 3SE +/- 1.23, N = 3SE +/- 1.51, N = 33027.681034.542059.521048.491. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Person Vehicle Bike Detection FP16 - Device: CPUAVX-512 OnAVX-512 Off3691215SE +/- 0.01, N = 3SE +/- 0.04, N = 3SE +/- 0.03, N = 3SE +/- 0.11, N = 33.7210.064.7710.081. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Person Vehicle Bike Detection FP16 - Device: CPUAVX-512 OnAVX-512 Off5001000150020002500SE +/- 5.99, N = 3SE +/- 3.56, N = 3SE +/- 10.71, N = 3SE +/- 8.20, N = 32133.76792.031667.92791.261. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

TensorFlow

This is a benchmark of the TensorFlow deep learning framework using the TensorFlow reference benchmarks (tensorflow/benchmarks with tf_cnn_benchmarks.py). Note with the Phoronix Test Suite there is also pts/tensorflow-lite for benchmarking the TensorFlow Lite binaries if desired for complementary metrics. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 64 - Model: ResNet-50AVX-512 OnAVX-512 Off1224364860SE +/- 0.04, N = 3SE +/- 0.02, N = 3SE +/- 0.02, N = 3SE +/- 0.02, N = 351.7438.9944.0819.73

OpenVINO

This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Face Detection Retail FP16 - Device: CPUAVX-512 OnAVX-512 Off0.88651.7732.65953.5464.4325SE +/- 0.00, N = 3SE +/- 0.01, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 31.533.922.283.941. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Face Detection Retail FP16 - Device: CPUAVX-512 OnAVX-512 Off11002200330044005500SE +/- 7.46, N = 3SE +/- 7.10, N = 3SE +/- 2.51, N = 3SE +/- 4.63, N = 34938.982015.673357.782015.991. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

NAMD

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgns/day, More Is BetterNAMD 3.0b6Input: ATPase with 327,506 AtomsAVX-512 OnAVX-512 Off0.78381.56762.35143.13523.919SE +/- 0.02808, N = 3SE +/- 0.01059, N = 3SE +/- 0.02678, N = 3SE +/- 0.00219, N = 33.483381.697573.058771.49354

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgns/day, More Is BetterNAMD 3.0b6Input: STMV with 1,066,628 AtomsAVX-512 OnAVX-512 Off0.22210.44420.66630.88841.1105SE +/- 0.00150, N = 3SE +/- 0.00040, N = 3SE +/- 0.00092, N = 3SE +/- 0.00052, N = 30.987070.484920.825810.42614

OpenVINO

This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Machine Translation EN To DE FP16 - Device: CPUAVX-512 OnAVX-512 Off20406080100SE +/- 0.15, N = 3SE +/- 0.49, N = 3SE +/- 0.30, N = 3SE +/- 0.55, N = 345.10102.8657.56104.111. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Machine Translation EN To DE FP16 - Device: CPUAVX-512 OnAVX-512 Off4080120160200SE +/- 0.57, N = 3SE +/- 0.38, N = 3SE +/- 0.72, N = 3SE +/- 0.41, N = 3177.1177.69138.7976.771. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OSPRay

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 3.2Benchmark: gravity_spheres_volume/dim_512/scivis/real_timeAVX-512 OnAVX-512 Off3691215SE +/- 0.01315, N = 3SE +/- 0.01723, N = 3SE +/- 0.00679, N = 3SE +/- 0.00671, N = 39.532776.469207.937304.20825

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 3.2Benchmark: gravity_spheres_volume/dim_512/ao/real_timeAVX-512 OnAVX-512 Off3691215SE +/- 0.00263, N = 3SE +/- 0.00321, N = 3SE +/- 0.01996, N = 3SE +/- 0.00077, N = 39.727006.706628.076464.44233

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 the Intel oneAPI toolkit. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: IP Shapes 3D - Engine: CPUAVX-512 OnAVX-512 Off246810SE +/- 0.00113, N = 5SE +/- 0.00900, N = 5SE +/- 0.00401, N = 5SE +/- 0.00179, N = 53.075306.167213.115086.602931. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl

miniBUDE

MiniBUDE is a mini application for the the core computation of the Bristol University Docking Engine (BUDE). This test profile currently makes use of the OpenMP implementation of miniBUDE for CPU benchmarking. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgBillion Interactions/s, More Is BetterminiBUDE 20210901Implementation: OpenMP - Input Deck: BM1AVX-512 OnAVX-512 Off20406080100SE +/- 0.06, N = 5SE +/- 0.02, N = 3SE +/- 0.00, N = 4SE +/- 0.33, N = 377.1938.4753.1238.401. (CC) gcc options: -std=c99 -Ofast -ffast-math -fopenmp -march=native -lm

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgGFInst/s, More Is BetterminiBUDE 20210901Implementation: OpenMP - Input Deck: BM1AVX-512 OnAVX-512 Off400800120016002000SE +/- 1.54, N = 5SE +/- 0.52, N = 3SE +/- 0.05, N = 4SE +/- 8.36, N = 31929.75961.731328.05960.041. (CC) gcc options: -std=c99 -Ofast -ffast-math -fopenmp -march=native -lm

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 the Intel oneAPI toolkit. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: Deconvolution Batch shapes_3d - Engine: CPUAVX-512 OnAVX-512 Off0.6131.2261.8392.4523.065SE +/- 0.00309, N = 9SE +/- 0.01154, N = 9SE +/- 0.00055, N = 9SE +/- 0.00474, N = 91.363862.597612.388302.724631. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -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 Model Zoo. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: ResNet50 v1-12-int8 - Device: CPU - Executor: StandardAVX-512 OnAVX-512 Off120240360480600SE +/- 0.98, N = 3SE +/- 1.07, N = 3SE +/- 1.03, N = 3SE +/- 0.27, N = 3571.47447.24419.19299.671. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenVINO

This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Vehicle Detection FP16 - Device: CPUAVX-512 OnAVX-512 Off48121620SE +/- 0.01, N = 3SE +/- 0.07, N = 3SE +/- 0.08, N = 3SE +/- 0.05, N = 38.7715.2910.0216.501. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Vehicle Detection FP16 - Device: CPUAVX-512 OnAVX-512 Off2004006008001000SE +/- 0.91, N = 3SE +/- 2.33, N = 3SE +/- 6.59, N = 3SE +/- 1.66, N = 3908.96521.70795.70483.591. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

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 the Intel oneAPI toolkit. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: IP Shapes 1D - Engine: CPUAVX-512 OnAVX-512 Off0.27270.54540.81811.09081.3635SE +/- 0.006767, N = 6SE +/- 0.002635, N = 4SE +/- 0.000816, N = 4SE +/- 0.000909, N = 40.6648751.0585501.0066101.2119501. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl

OpenVINO

This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Person Detection FP16 - Device: CPUAVX-512 OnAVX-512 Off20406080100SE +/- 0.34, N = 3SE +/- 0.20, N = 3SE +/- 1.04, N = 3SE +/- 0.10, N = 3109.2960.69101.5162.801. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Person Detection FP16 - Device: CPUAVX-512 OnAVX-512 Off306090120150SE +/- 0.22, N = 3SE +/- 0.43, N = 3SE +/- 0.82, N = 3SE +/- 0.21, N = 373.12131.6078.75127.261. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Road Segmentation ADAS FP16-INT8 - Device: CPUAVX-512 OnAVX-512 Off140280420560700SE +/- 3.09, N = 3SE +/- 3.66, N = 3SE +/- 1.08, N = 3SE +/- 0.48, N = 3638.67476.88481.04359.261. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Road Segmentation ADAS FP16-INT8 - Device: CPUAVX-512 OnAVX-512 Off510152025SE +/- 0.06, N = 3SE +/- 0.13, N = 3SE +/- 0.04, N = 3SE +/- 0.03, N = 312.5016.7416.5822.221. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OSPRay

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 3.2Benchmark: gravity_spheres_volume/dim_512/pathtracer/real_timeAVX-512 OnAVX-512 Off3691215SE +/- 0.00324, N = 3SE +/- 0.00447, N = 3SE +/- 0.01877, N = 3SE +/- 0.00125, N = 311.030009.083769.179626.35755

simdjson

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgGB/s, More Is Bettersimdjson 3.10Throughput Test: PartialTweetsAVX-512 OnAVX-512 Off3691215SE +/- 0.12, N = 15SE +/- 0.08, N = 9SE +/- 0.02, N = 3SE +/- 0.07, N = 312.389.999.477.201. (CXX) g++ options: -O3 -lrt

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgGB/s, More Is Bettersimdjson 3.10Throughput Test: DistinctUserIDAVX-512 OnAVX-512 Off3691215SE +/- 0.07, N = 3SE +/- 0.08, N = 3SE +/- 0.02, N = 3SE +/- 0.05, N = 313.0510.4410.287.591. (CXX) g++ options: -O3 -lrt

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 the Intel oneAPI toolkit. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: Recurrent Neural Network Training - Engine: CPUAVX-512 OnAVX-512 Off30060090012001500SE +/- 1.13, N = 3SE +/- 0.50, N = 3SE +/- 0.97, N = 3SE +/- 0.80, N = 3759.931258.801245.281303.161. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl

simdjson

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgGB/s, More Is Bettersimdjson 3.10Throughput Test: TopTweetAVX-512 OnAVX-512 Off3691215SE +/- 0.14, N = 3SE +/- 0.02, N = 3SE +/- 0.09, N = 3SE +/- 0.01, N = 312.9510.3310.487.641. (CXX) g++ options: -O3 -lrt

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 Model Zoo. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: super-resolution-10 - Device: CPU - Executor: StandardAVX-512 OnAVX-512 Off50100150200250SE +/- 0.18, N = 3SE +/- 0.04, N = 3SE +/- 0.09, N = 3SE +/- 0.52, N = 3227.65151.02144.78137.131. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

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 the Intel oneAPI toolkit. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: Recurrent Neural Network Inference - Engine: CPUAVX-512 OnAVX-512 Off140280420560700SE +/- 0.61, N = 3SE +/- 0.28, N = 3SE +/- 0.67, N = 3SE +/- 0.46, N = 3404.70635.73629.60664.151. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl

OpenVINO

This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Noise Suppression Poconet-Like FP16 - Device: CPUAVX-512 OnAVX-512 Off3691215SE +/- 0.02, N = 3SE +/- 0.03, N = 3SE +/- 0.02, N = 3SE +/- 0.03, N = 38.4313.189.3613.571. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Noise Suppression Poconet-Like FP16 - Device: CPUAVX-512 OnAVX-512 Off400800120016002000SE +/- 3.92, N = 3SE +/- 2.28, N = 3SE +/- 2.92, N = 3SE +/- 2.29, N = 31867.601201.731680.421166.241. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OSPRay Studio

Intel OSPRay Studio is an open-source, interactive visualization and ray-tracing software package. OSPRay Studio makes use of Intel OSPRay, a portable ray-tracing engine for high-performance, high-fidelity visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 1 - Resolution: 4K - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPUAVX-512 OnAVX-512 Off11002200330044005500SE +/- 1.00, N = 3SE +/- 2.03, N = 3SE +/- 2.85, N = 3SE +/- 5.00, N = 33225359440045131

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 1 - Resolution: 4K - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPUAVX-512 OnAVX-512 Off20K40K60K80K100KSE +/- 42.15, N = 3SE +/- 25.03, N = 3SE +/- 142.41, N = 3SE +/- 106.67, N = 354022599586761185277

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 1 - Resolution: 4K - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPUAVX-512 OnAVX-512 Off40K80K120K160K200KSE +/- 68.30, N = 3SE +/- 147.11, N = 3SE +/- 209.54, N = 3SE +/- 116.94, N = 3105798117712131605166805

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 3 - Resolution: 4K - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPUAVX-512 OnAVX-512 Off13002600390052006500SE +/- 1.53, N = 3SE +/- 4.33, N = 3SE +/- 10.17, N = 3SE +/- 9.87, N = 33829423547356020

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 3 - Resolution: 4K - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPUAVX-512 OnAVX-512 Off40K80K120K160K200KSE +/- 55.03, N = 3SE +/- 188.82, N = 3SE +/- 26.96, N = 3SE +/- 475.57, N = 3124596138397154740195460

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 3 - Resolution: 4K - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPUAVX-512 OnAVX-512 Off20K40K60K80K100KSE +/- 49.41, N = 3SE +/- 36.23, N = 3SE +/- 199.04, N = 3SE +/- 91.66, N = 363476703647927199475

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 the Intel oneAPI toolkit. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: Deconvolution Batch shapes_1d - Engine: CPUAVX-512 OnAVX-512 Off0.63061.26121.89182.52243.153SE +/- 0.00399, N = 3SE +/- 0.00490, N = 3SE +/- 0.00663, N = 3SE +/- 0.00224, N = 31.788462.411922.753462.802511. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl

simdjson

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgGB/s, More Is Bettersimdjson 3.10Throughput Test: KostyaAVX-512 OnAVX-512 Off246810SE +/- 0.04, N = 3SE +/- 0.01, N = 3SE +/- 0.05, N = 3SE +/- 0.01, N = 37.476.436.014.801. (CXX) g++ options: -O3 -lrt

LeelaChessZero

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgNodes Per Second, More Is BetterLeelaChessZero 0.31.1Backend: EigenAVX-512 OnAVX-512 Off50100150200250SE +/- 1.76, N = 3SE +/- 1.53, N = 3SE +/- 2.03, N = 3SE +/- 1.67, N = 32201621771421. (CXX) g++ options: -flto -pthread

OpenVKL

OpenVKL is the Intel Open Volume Kernel Library that offers high-performance volume computation kernels and part of the Intel oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgItems / Sec, More Is BetterOpenVKL 2.0.0Benchmark: vklBenchmarkCPU ISPCAVX-512 OnAVX-512 Off160320480640800SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.33, N = 3765656647507

Numpy Benchmark

This is a test to obtain the general Numpy performance. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgScore, More Is BetterNumpy BenchmarkAVX-512 OnAVX-512 Off2004006008001000SE +/- 6.93, N = 3SE +/- 2.31, N = 3SE +/- 6.26, N = 3SE +/- 2.88, N = 31054.921010.59794.79731.83

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 Model Zoo. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: CaffeNet 12-int8 - Device: CPU - Executor: StandardAVX-512 OnAVX-512 Off2004006008001000SE +/- 7.80, N = 15SE +/- 0.45, N = 3SE +/- 3.29, N = 3SE +/- 0.54, N = 3872.94803.44695.73605.771. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: fcn-resnet101-11 - Device: CPU - Executor: StandardAVX-512 OnAVX-512 Off1.09232.18463.27694.36925.4615SE +/- 0.01446, N = 3SE +/- 0.00187, N = 3SE +/- 0.00354, N = 3SE +/- 0.00295, N = 34.854883.600323.547753.416361. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

TensorFlow

This is a benchmark of the TensorFlow deep learning framework using the TensorFlow reference benchmarks (tensorflow/benchmarks with tf_cnn_benchmarks.py). Note with the Phoronix Test Suite there is also pts/tensorflow-lite for benchmarking the TensorFlow Lite binaries if desired for complementary metrics. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 1 - Model: ResNet-50AVX-512 OnAVX-512 Off48121620SE +/- 0.04, N = 5SE +/- 0.01, N = 5SE +/- 0.03, N = 5SE +/- 0.02, N = 416.8212.9816.2312.22

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.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 1 - Model: ResNet-50AVX-512 OnAVX-512 Off20406080100SE +/- 0.88, N = 4SE +/- 0.81, N = 4SE +/- 0.54, N = 3SE +/- 0.33, N = 385.1470.1474.4663.23

Embree

Intel Embree is a collection of high-performance ray-tracing kernels for execution on CPUs (and GPUs via SYCL) and supporting instruction sets such as SSE, AVX, AVX2, and AVX-512. Embree also supports making use of the Intel SPMD Program Compiler (ISPC). Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 4.3Binary: Pathtracer ISPC - Model: Asian DragonAVX-512 OnAVX-512 Off918273645SE +/- 0.01, N = 3SE +/- 0.05, N = 3SE +/- 0.03, N = 3SE +/- 0.01, N = 341.3539.5334.7131.04

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 4.3Binary: Pathtracer ISPC - Model: Asian Dragon ObjAVX-512 OnAVX-512 Off816243240SE +/- 0.01, N = 3SE +/- 0.04, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 335.5134.0829.5526.72

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.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.2.1Device: CPU - Batch Size: 64 - Model: ResNet-50AVX-512 OnAVX-512 Off1326395265SE +/- 0.41, N = 3SE +/- 0.36, N = 3SE +/- 0.30, N = 3SE +/- 0.11, N = 357.8947.0550.2343.87

OSPRay

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 3.2Benchmark: particle_volume/pathtracer/real_timeAVX-512 OnAVX-512 Off60120180240300SE +/- 0.40, N = 3SE +/- 0.58, N = 3SE +/- 0.60, N = 3SE +/- 0.33, N = 3277.28261.16239.30210.85

simdjson

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgGB/s, More Is Bettersimdjson 3.10Throughput Test: LargeRandomAVX-512 OnAVX-512 Off0.46580.93161.39741.86322.329SE +/- 0.00, N = 3SE +/- 0.03, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 32.071.901.811.581. (CXX) g++ options: -O3 -lrt

LeelaChessZero

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgNodes Per Second, More Is BetterLeelaChessZero 0.31.1Backend: BLASAVX-512 OnAVX-512 Off50100150200250SE +/- 1.76, N = 3SE +/- 2.40, N = 3SE +/- 1.65, N = 6SE +/- 0.33, N = 32202051711681. (CXX) g++ options: -flto -pthread

Embree

Intel Embree is a collection of high-performance ray-tracing kernels for execution on CPUs (and GPUs via SYCL) and supporting instruction sets such as SSE, AVX, AVX2, and AVX-512. Embree also supports making use of the Intel SPMD Program Compiler (ISPC). Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 4.3Binary: Pathtracer ISPC - Model: CrownAVX-512 OnAVX-512 Off816243240SE +/- 0.04, N = 3SE +/- 0.02, N = 3SE +/- 0.08, N = 3SE +/- 0.07, N = 336.6534.6333.7728.78

Y-Cruncher

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgSeconds, Fewer Is BetterY-Cruncher 0.8.5Pi Digits To Calculate: 1BAVX-512 OnAVX-512 Off510152025SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.02, N = 3SE +/- 0.01, N = 316.4920.2316.0318.06

Mobile Neural Network

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgms, Fewer Is BetterMobile Neural Network 2.9.b11b7037dModel: resnet-v2-50AVX-512 OnAVX-512 Off3691215SE +/- 0.04, N = 3SE +/- 0.06, N = 3SE +/- 0.11, N = 3SE +/- 0.10, N = 310.0511.3210.8812.661. (CXX) g++ options: -std=c++11 -O3 -fvisibility=hidden -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -ffast-math -fno-rtti -fno-exceptions -pthread -ldl

GROMACS

The GROMACS (GROningen MAchine for Chemical Simulations) molecular dynamics package testing with the water_GMX50 data. This test profile allows selecting between CPU and GPU-based GROMACS builds. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgNs Per Day, More Is BetterGROMACS 2024Implementation: MPI CPU - Input: water_GMX50_bareAVX-512 OnAVX-512 Off0.70541.41082.11622.82163.527SE +/- 0.004, N = 3SE +/- 0.004, N = 3SE +/- 0.001, N = 3SE +/- 0.006, N = 33.1352.6692.7402.5921. (CXX) g++ options: -O3 -lm

Y-Cruncher

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgSeconds, Fewer Is BetterY-Cruncher 0.8.5Pi Digits To Calculate: 500MAVX-512 OnAVX-512 Off3691215SE +/- 0.002, N = 5SE +/- 0.008, N = 5SE +/- 0.007, N = 5SE +/- 0.009, N = 57.5709.0737.6968.395

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 Model Zoo. Learn more via the OpenBenchmarking.org test page.

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: ArcFace ResNet-100 - Device: CPU - Executor: StandardAVX-512 OnAVX-512 Off1224364860SE +/- 0.61, N = 3SE +/- 0.22, N = 3SE +/- 0.15, N = 3SE +/- 0.28, N = 355.5149.8147.8547.851. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

XNNPACK

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgus, Fewer Is BetterXNNPACK 2cd86bModel: QU8MobileNetV2AVX-512 OnAVX-512 Off2004006008001000SE +/- 3.93, N = 3SE +/- 0.58, N = 3SE +/- 0.67, N = 3SE +/- 1.45, N = 37858568078701. (CXX) g++ options: -O3 -lrt -lm

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgus, Fewer Is BetterXNNPACK 2cd86bModel: QU8MobileNetV3LargeAVX-512 OnAVX-512 Off30060090012001500SE +/- 1.15, N = 3SE +/- 2.03, N = 3SE +/- 3.38, N = 3SE +/- 3.84, N = 311541177109411301. (CXX) g++ options: -O3 -lrt -lm

CPU Temperature Monitor

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgCelsiusCPU Temperature MonitorPhoronix Test Suite System MonitoringAVX-512 OnAVX-512 Off20406080100Min: 40.5 / Avg: 73.15 / Max: 89.13Min: 39.5 / Avg: 75.07 / Max: 94.5Min: 37.63 / Avg: 80.57 / Max: 96.13Min: 37.25 / Avg: 83.51 / Max: 95.88

CPU Power Consumption Monitor

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgWattsCPU Power Consumption MonitorPhoronix Test Suite System MonitoringAVX-512 OnAVX-512 Off4080120160200Min: 22.78 / Avg: 148.64 / Max: 205.19Min: 13.82 / Avg: 152.26 / Max: 203.94Min: 10.68 / Avg: 169.29 / Max: 237.09Min: 11.34 / Avg: 172.33 / Max: 237.26

CPU Peak Freq (Highest CPU Core Frequency) Monitor

Ryzen 9 9950XRyzen 9 7950XOpenBenchmarking.orgMegahertzCPU Peak Freq (Highest CPU Core Frequency) MonitorPhoronix Test Suite System MonitoringAVX-512 OnAVX-512 Off10002000300040005000Min: 2477 / Avg: 5097.18 / Max: 5799Min: 600 / Avg: 4986.91 / Max: 5728Min: 545 / Avg: 5001.33 / Max: 5725Min: 545 / Avg: 4954.31 / Max: 5725

74 Results Shown

OpenVINO:
  Weld Porosity Detection FP16 - CPU:
    ms
    FPS
  Face Detection FP16 - CPU:
    FPS
    ms
  Age Gender Recognition Retail 0013 FP16 - CPU:
    ms
    FPS
  Person Re-Identification Retail FP16 - CPU:
    ms
  Handwritten English Recognition FP16 - CPU:
    ms
    FPS
  Person Re-Identification Retail FP16 - CPU:
    FPS
  Person Vehicle Bike Detection FP16 - CPU:
    ms
    FPS
TensorFlow
OpenVINO:
  Face Detection Retail FP16 - CPU:
    ms
    FPS
NAMD:
  ATPase with 327,506 Atoms
  STMV with 1,066,628 Atoms
OpenVINO:
  Machine Translation EN To DE FP16 - CPU:
    ms
    FPS
OSPRay:
  gravity_spheres_volume/dim_512/scivis/real_time
  gravity_spheres_volume/dim_512/ao/real_time
oneDNN
miniBUDE:
  OpenMP - BM1:
    Billion Interactions/s
    GFInst/s
oneDNN
ONNX Runtime
OpenVINO:
  Vehicle Detection FP16 - CPU:
    ms
    FPS
oneDNN
OpenVINO:
  Person Detection FP16 - CPU:
    FPS
    ms
  Road Segmentation ADAS FP16-INT8 - CPU:
    FPS
    ms
OSPRay
simdjson:
  PartialTweets
  DistinctUserID
oneDNN
simdjson
ONNX Runtime
oneDNN
OpenVINO:
  Noise Suppression Poconet-Like FP16 - CPU:
    ms
    FPS
OSPRay Studio:
  1 - 4K - 1 - Path Tracer - CPU
  1 - 4K - 16 - Path Tracer - CPU
  1 - 4K - 32 - Path Tracer - CPU
  3 - 4K - 1 - Path Tracer - CPU
  3 - 4K - 32 - Path Tracer - CPU
  3 - 4K - 16 - Path Tracer - CPU
oneDNN
simdjson
LeelaChessZero
OpenVKL
Numpy Benchmark
ONNX Runtime:
  CaffeNet 12-int8 - CPU - Standard
  fcn-resnet101-11 - CPU - Standard
TensorFlow
PyTorch
Embree:
  Pathtracer ISPC - Asian Dragon
  Pathtracer ISPC - Asian Dragon Obj
PyTorch
OSPRay
simdjson
LeelaChessZero
Embree
Y-Cruncher
Mobile Neural Network
GROMACS
Y-Cruncher
ONNX Runtime
XNNPACK:
  QU8MobileNetV2
  QU8MobileNetV3Large
CPU Temperature Monitor:
  Phoronix Test Suite System Monitoring:
    Celsius
    Watts
    Megahertz