Xeon Max Linux 6.9 Kernel Benchmarks

2 x Intel Xeon Max 9480 benchmarks 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 2404035-NE-XEONMAXLI15
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:

Timed Code Compilation 3 Tests
C/C++ Compiler Tests 4 Tests
CPU Massive 13 Tests
Creator Workloads 17 Tests
Database Test Suite 2 Tests
Encoding 4 Tests
Game Development 3 Tests
HPC - High Performance Computing 7 Tests
Imaging 3 Tests
Common Kernel Benchmarks 3 Tests
Large Language Models 2 Tests
Machine Learning 6 Tests
Multi-Core 18 Tests
NVIDIA GPU Compute 3 Tests
Intel oneAPI 6 Tests
Programmer / Developer System Benchmarks 3 Tests
Python Tests 5 Tests
Raytracing 2 Tests
Renderers 4 Tests
Server 3 Tests
Server CPU Tests 8 Tests
Video Encoding 3 Tests
Common Workstation Benchmarks 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
Linux 6.8
March 29
  1 Day, 17 Hours, 1 Minute
Linux 6.9-rc2
April 01
  1 Day, 15 Hours, 4 Minutes
Invert Hiding All Results Option
  1 Day, 16 Hours, 2 Minutes
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):

Linux 6.8

Processor: 2 x Intel Xeon Max 9480 @ 3.50GHz (112 Cores / 224 Threads), Motherboard: Supermicro SYS-221H-TNR X13DEM v1.10 (1.3 BIOS), Chipset: Intel Device 1bce, Memory: 512GB, Disk: 7682GB INTEL SSDPF2KX076TZ, Graphics: ASPEED, Network: 2 x Broadcom BCM57508 NetXtreme-E 10Gb/25Gb/40Gb/50Gb/100Gb/200Gb

OS: Ubuntu 23.10, Kernel: 6.8.0-060800-generic (x86_64), Desktop: GNOME Shell 45.0, Display Server: X Server 1.21.1.7, Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1024x768

Kernel Notes: Transparent Huge Pages: madvise
Compiler Notes: --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,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-defaulted --enable-offload-targets=nvptx-none=/build/gcc-13-XYspKM/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-XYspKM/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-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
Processor Notes: Scaling Governor: intel_cpufreq performance - CPU Microcode: 0x2c000290
Python Notes: Python 3.11.6
Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + 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 RSB filling PBRSB-eIBRS: SW sequence + srbds: Not affected + tsx_async_abort: Not affected

Linux 6.9-rc2

OS: Ubuntu 23.10, Kernel: 6.9.0-060900rc2-generic (x86_64), Desktop: GNOME Shell 45.0, Display Server: X Server 1.21.1.7, Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1024x768

Kernel Notes: Transparent Huge Pages: madvise
Compiler Notes: --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,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-defaulted --enable-offload-targets=nvptx-none=/build/gcc-13-XYspKM/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-XYspKM/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-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
Processor Notes: Scaling Governor: intel_cpufreq performance - CPU Microcode: 0x2c000290
Python Notes: Python 3.11.6
Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + reg_file_data_sampling: Not affected + retbleed: Not affected + spec_rstack_overflow: 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 RSB filling PBRSB-eIBRS: SW sequence + srbds: Not affected + tsx_async_abort: Not affected

Xeon Max Linux 6.9 Kernel BenchmarksOpenBenchmarking.orgPhoronix Test Suite2 x Intel Xeon Max 9480 @ 3.50GHz (112 Cores / 224 Threads)Supermicro SYS-221H-TNR X13DEM v1.10 (1.3 BIOS)Intel Device 1bce512GB7682GB INTEL SSDPF2KX076TZASPEED2 x Broadcom BCM57508 NetXtreme-E 10Gb/25Gb/40Gb/50Gb/100Gb/200GbUbuntu 23.106.8.0-060800-generic (x86_64)6.9.0-060900rc2-generic (x86_64)GNOME Shell 45.0X Server 1.21.1.7GCC 13.2.0ext41024x768ProcessorMotherboardChipsetMemoryDiskGraphicsNetworkOSKernelsDesktopDisplay ServerCompilerFile-SystemScreen ResolutionXeon Max Linux 6.9 Kernel Benchmarks PerformanceSystem 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,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-defaulted --enable-offload-targets=nvptx-none=/build/gcc-13-XYspKM/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-XYspKM/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-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_cpufreq performance - CPU Microcode: 0x2c000290- Python 3.11.6- Linux 6.8: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: 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 RSB filling PBRSB-eIBRS: SW sequence + srbds: Not affected + tsx_async_abort: Not affected - Linux 6.9-rc2: 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 RSB filling PBRSB-eIBRS: SW sequence + srbds: Not affected + tsx_async_abort: Not affected

Linux 6.8 vs. Linux 6.9-rc2 ComparisonPhoronix Test SuiteBaseline+27.9%+27.9%+55.8%+55.8%+83.7%+83.7%22%13.9%10.5%9.9%9.9%8.8%8.3%8%6.6%6.1%5.9%5.9%5.8%5.7%5.4%5.4%5%5%4.9%4.9%4.9%4.6%4.5%4.2%4.1%4.1%4.1%4%3.9%3.9%3.8%3.7%3.4%3.4%3.4%3.3%3.2%3.2%3.1%3%2.8%2.8%2.8%2.8%2.7%2.7%2.5%2.2%2.2%2.2%2.1%2.1%2%2%2%8.7%11.7%7.5%3.1%9.5%6%3.6%16.6%5.6%4.4%25.2%5.9%4.7%R.N.N.T - CPU111.5%fcn-resnet101-11 - CPU - StandardT5 Encoder - CPU - Standard32 - Process13.1%fcn-resnet101-11 - CPU - ParallelPthread10.2%R.v.1.i - CPU - StandardA.G.R.R.0.F - CPUPathtracer - Asian Dragon Obj9.1%A.G.R.R.0.F - CPUPathtracer ISPC - Crown8.7%yolov4 - CPU - Standardbertsquad-12 - CPU - StandardD.B.s - CPUlibx265 - Platform6.3%CPU - 1 - AlexNet1024F.R.C.R.5.F.i - CPU - StandardT5 Encoder - CPU - ParallelSENDFILEG.C.S.FCaffeNet 12-int8 - CPU - StandardCPU - 64 - AlexNet5.1%CPU - 64 - GoogLeNet5%V.D.F - CPUV.D.F - CPUS.V.M.PQ.7.C.E.74.9%Context Switching4.9%P.D.F - CPUP.D.F - CPUArcFace ResNet-100 - CPU - Parallel500MRT.ldr_alb_nrm.3840x2160 - CPU-Only4.5%Default4.4%Pathtracer - Crown4.3%MMAPparticle_volume/pathtracer/real_time4.1%W.P.D.F.I - CPUJPEG - 90super-resolution-10 - CPU - StandardW.P.D.F.I - CPUChess BenchmarkSocket Activity3.9%Pathtracer ISPC - Asian Dragon2 - 1080p - 32 - Path Tracer - CPU3.9%GPT-2 - CPU - ParallelSHA2563.7%gravity_spheres_volume/dim_512/pathtracer/real_timeIO_uring3.7%P.D.F - CPUBosphorus 4K - Faster3.4%llama-2-13b.Q4_0.ggufP.D.F - CPUFishy Cat - CPU-Only3.4%ArcFace ResNet-100 - CPU - StandardR.S.A.F - CPUR.S.A.F - CPU3 - 1080p - 32 - Path Tracer - CPUSemaphores3.1%gravity_spheres_volume/dim_512/scivis/real_timeQ.9.C.E.73%particle_volume/ao/real_time3%Barbershop - CPU-Only2.8%CPU - 16 - ResNet-50RSA4096PNG - 902.8%Cryptollama-2-70b-chat.Q5_0.ggufPathtracer ISPC - Asian Dragon ObjMPI CPU - water_GMX50_bare2.7%Bosphorus 1080p - Faster2.7%16 - Thread3 - 4K - 1 - Path Tracer - CPU2.7%wizardcoder-python-34b-v1.0.Q6_K - CPU2.6%1 - 1080p - 1 - Path Tracer - CPU2.6%R.R.W.R2.5%2 - 1080p - 1 - Path Tracer - CPUZlib2.4%Memory Copying2.3%RT.hdr_alb_nrm.3840x2160 - CPU-Only2.2%C.B.S.A - CPUV.P.M3 - 4K - 16 - Path Tracer - CPUJunkshop - CPU-Only2.1%3 - 1080p - 16 - Path Tracer - CPUMatrix MathdefconfigMonero - 1MH.E.R.F.I - CPU2%CPU - 64 - ResNet-50H.E.R.F.I - CPU2%yolov4 - CPU - Standardfcn-resnet101-11 - CPU - Parallelbertsquad-12 - CPU - StandardArcFace ResNet-100 - CPU - StandardR.v.1.i - CPU - StandardT5 Encoder - CPU - ParallelGPT-2 - CPU - ParallelT5 Encoder - CPU - StandardCaffeNet 12-int8 - CPU - Standardsuper-resolution-10 - CPU - Standardfcn-resnet101-11 - CPU - StandardF.R.C.R.5.F.i - CPU - StandardArcFace ResNet-100 - CPU - ParalleloneDNNONNX RuntimeONNX RuntimeHackbenchONNX RuntimeStress-NGONNX RuntimeOpenVINOEmbreeOpenVINOEmbreeONNX RuntimeONNX RuntimeoneDNNFFmpegTensorFlowMariaDB mariadb-slapONNX RuntimeONNX RuntimeStress-NGStress-NGONNX RuntimeTensorFlowTensorFlowOpenVINOOpenVINOStress-NGWebP2 Image EncodeStress-NGOpenVINOOpenVINOONNX RuntimeY-CruncherIntel Open Image DenoiseWebP2 Image EncodeEmbreeStress-NGOSPRayOpenVINOJPEG-XL libjxlONNX RuntimeOpenVINOStockfishStress-NGEmbreeOSPRay StudioONNX RuntimeOpenSSLOSPRayStress-NGOpenVINOVVenCLlama.cppOpenVINOBlenderONNX RuntimeOpenVINOOpenVINOOSPRay StudioStress-NGOSPRayWebP2 Image EncodeOSPRayBlenderTensorFlowOpenSSLJPEG-XL libjxlStress-NGLlama.cppEmbreeGROMACSVVenCHackbenchOSPRay StudioLlamafileOSPRay StudioRocksDBOSPRay StudioStress-NGStress-NGIntel Open Image DenoiseoneDNNBRL-CADOSPRay StudioBlenderOSPRay StudioStress-NGTimed Linux Kernel CompilationXmrigOpenVINOTensorFlowOpenVINOONNX RuntimeONNX RuntimeONNX RuntimeONNX RuntimeONNX RuntimeONNX RuntimeONNX RuntimeONNX RuntimeONNX RuntimeONNX RuntimeONNX RuntimeONNX RuntimeONNX RuntimeLinux 6.8Linux 6.9-rc2

Xeon Max Linux 6.9 Kernel Benchmarksmysqlslap: 1024mysqlslap: 512quicksilver: CORAL2 P2stockfish: Chess Benchmarkllama-cpp: llama-2-70b-chat.Q5_0.ggufffmpeg: libx265 - Platformblender: Barbershop - CPU-Onlyspeedb: Seq Fillffmpeg: libx265 - Video On Demandbrl-cad: VGR Performance Metricospray: particle_volume/pathtracer/real_timetensorflow: CPU - 32 - ResNet-50vvenc: Bosphorus 4K - Fastopenssl: SHA256tensorflow: CPU - 256 - ResNet-50ospray-studio: 2 - 1080p - 32 - Path Tracer - CPUtensorflow: CPU - 64 - ResNet-50tensorflow: CPU - 256 - GoogLeNetwebp2: Quality 100, Lossless Compressionllama-cpp: llama-2-13b.Q4_0.ggufospray-studio: 1 - 4K - 1 - Path Tracer - CPUffmpeg: libx265 - Livetensorflow: CPU - 1 - ResNet-50llamafile: wizardcoder-python-34b-v1.0.Q6_K - CPUbuild-linux-kernel: allmodconfigonednn: Recurrent Neural Network Training - CPUopenvino: Age Gender Recognition Retail 0013 FP16-INT8 - CPUopenvino: Age Gender Recognition Retail 0013 FP16-INT8 - CPUvvenc: Bosphorus 4K - Fasterffmpeg: libx264 - Uploadospray-studio: 3 - 1080p - 1 - Path Tracer - CPUonnx: yolov4 - CPU - Standardonnx: yolov4 - CPU - Standardonnx: fcn-resnet101-11 - CPU - Parallelonnx: fcn-resnet101-11 - CPU - Parallelonnx: bertsquad-12 - CPU - Standardonnx: bertsquad-12 - CPU - Standardonnx: ArcFace ResNet-100 - CPU - Standardonnx: ArcFace ResNet-100 - CPU - Standardospray-studio: 2 - 1080p - 16 - Path Tracer - CPUonnx: ResNet50 v1-12-int8 - CPU - Standardonnx: ResNet50 v1-12-int8 - CPU - Standardonnx: T5 Encoder - CPU - Parallelonnx: T5 Encoder - CPU - Parallelospray-studio: 1 - 4K - 32 - Path Tracer - CPUblender: Pabellon Barcelona - CPU-Onlyjpegxl: JPEG - 80ospray-studio: 3 - 4K - 32 - Path Tracer - CPUquicksilver: CORAL2 P1v-ray: CPUblender: Junkshop - CPU-Onlytensorflow: CPU - 16 - ResNet-50ffmpeg: libx265 - Uploadospray-studio: 1 - 1080p - 32 - Path Tracer - CPUffmpeg: libx264 - Platformffmpeg: libx264 - Video On Demandospray-studio: 3 - 4K - 1 - Path Tracer - CPUllamafile: mistral-7b-instruct-v0.2.Q8_0 - CPUopenvino: Age Gender Recognition Retail 0013 FP16 - CPUopenvino: Age Gender Recognition Retail 0013 FP16 - CPUonnx: GPT-2 - CPU - Parallelonnx: GPT-2 - CPU - Parallelonnx: T5 Encoder - CPU - Standardonnx: T5 Encoder - CPU - Standardtensorflow: CPU - 32 - GoogLeNetonnx: CaffeNet 12-int8 - CPU - Parallelonnx: CaffeNet 12-int8 - CPU - Parallelonnx: CaffeNet 12-int8 - CPU - Standardonnx: CaffeNet 12-int8 - CPU - Standardonnx: super-resolution-10 - CPU - Standardonnx: super-resolution-10 - CPU - Standardopenssl: SHA512jpegxl: PNG - 80tensorflow: CPU - 64 - GoogLeNetospray: gravity_spheres_volume/dim_512/scivis/real_timeospray: gravity_spheres_volume/dim_512/ao/real_timeospray: particle_volume/scivis/real_timeospray-studio: 2 - 4K - 1 - Path Tracer - CPUembree: Pathtracer ISPC - Asian Dragon Objembree: Pathtracer - Asian Dragon Objstress-ng: Matrix 3D Mathstress-ng: Pipestress-ng: Cryptoonnx: fcn-resnet101-11 - CPU - Standardonnx: fcn-resnet101-11 - CPU - Standardonednn: Recurrent Neural Network Inference - CPUstress-ng: MMAPstress-ng: Glibc C String Functionsspeedb: Read Rand Write Randstress-ng: Matrix Mathstress-ng: Fused Multiply-Addospray-studio: 2 - 4K - 16 - Path Tracer - CPUtensorflow: CPU - 16 - GoogLeNetospray-studio: 3 - 4K - 16 - Path Tracer - CPUsrsran: PDSCH Processor Benchmark, Throughput Totalcachebench: Readcachebench: Read / Modify / Writecachebench: Writeblender: Fishy Cat - CPU-Onlyospray: gravity_spheres_volume/dim_512/pathtracer/real_timeospray-studio: 2 - 4K - 32 - Path Tracer - CPUstress-ng: Vector Floating Pointstress-ng: Floating Pointstress-ng: MEMFDembree: Pathtracer - Crownembree: Pathtracer ISPC - Asian Dragonembree: Pathtracer - Asian Dragonvvenc: Bosphorus 1080p - Fasteroidn: RTLightmap.hdr.4096x4096 - CPU-Onlyjpegxl: JPEG - 90xmrig: GhostRider - 1Mstress-ng: IO_uringospray: particle_volume/ao/real_timeospray-studio: 3 - 1080p - 32 - Path Tracer - CPUbuild-mesa: Time To Compiletensorflow: CPU - 1 - GoogLeNetstress-ng: Wide Vector Mathsvt-av1: Preset 4 - Bosphorus 4Kxmrig: Monero - 1My-cruncher: 1Bstress-ng: Pthreadospray-studio: 3 - 1080p - 16 - Path Tracer - CPUstress-ng: System V Message Passingospray-studio: 1 - 1080p - 16 - Path Tracer - CPUbuild-ffmpeg: Time To Compilerocksdb: Read Rand Write Randospray-studio: 2 - 1080p - 1 - Path Tracer - CPUstress-ng: Mallocbuild-linux-kernel: defconfigopenvino: Weld Porosity Detection FP16-INT8 - CPUopenvino: Weld Porosity Detection FP16-INT8 - CPUwebp2: Quality 95, Compression Effort 7svt-av1: Preset 8 - Bosphorus 4Koidn: RT.hdr_alb_nrm.3840x2160 - CPU-Onlytensorflow: CPU - 256 - AlexNetopenvino: Face Detection FP16-INT8 - CPUopenvino: Face Detection FP16-INT8 - CPUopenvino: Noise Suppression Poconet-Like FP16 - CPUopenvino: Noise Suppression Poconet-Like FP16 - CPUstress-ng: Memory Copyingopenvino: Person Vehicle Bike Detection FP16 - CPUopenvino: Person Vehicle Bike Detection FP16 - CPUopenvino: Face Detection FP16 - CPUopenvino: Face Detection FP16 - CPUopenvino: Road Segmentation ADAS FP16-INT8 - CPUopenvino: Road Segmentation ADAS FP16-INT8 - CPUopenvino: Person Detection FP16 - CPUopenvino: Person Detection FP16 - CPUopenvino: Person Detection FP32 - CPUopenvino: Person Detection FP32 - CPUopenvino: Person Re-Identification Retail FP16 - CPUopenvino: Person Re-Identification Retail FP16 - CPUblender: BMW27 - CPU-Onlyopenvino: Machine Translation EN To DE FP16 - CPUopenvino: Machine Translation EN To DE FP16 - CPUonednn: Deconvolution Batch shapes_1d - CPUopenvino: Vehicle Detection FP16-INT8 - CPUopenvino: Vehicle Detection FP16-INT8 - CPUtensorflow: CPU - 32 - AlexNetopenvino: Road Segmentation ADAS FP16 - CPUopenvino: Road Segmentation ADAS FP16 - CPUonnx: Faster R-CNN R-50-FPN-int8 - CPU - Parallelonnx: Faster R-CNN R-50-FPN-int8 - CPU - Parallelonnx: Faster R-CNN R-50-FPN-int8 - CPU - Standardonnx: Faster R-CNN R-50-FPN-int8 - CPU - Standardopenvino: Face Detection Retail FP16 - CPUopenvino: Face Detection Retail FP16 - CPUembree: Pathtracer ISPC - Crownopenvino: Weld Porosity Detection FP16 - CPUopenvino: Weld Porosity Detection FP16 - CPUopenvino: Face Detection Retail FP16-INT8 - CPUopenvino: Face Detection Retail FP16-INT8 - CPUopenvino: Vehicle Detection FP16 - CPUopenvino: Vehicle Detection FP16 - CPUopenvino: Handwritten English Recognition FP16 - CPUopenvino: Handwritten English Recognition FP16 - CPUonnx: bertsquad-12 - CPU - Parallelonnx: bertsquad-12 - CPU - Parallelonnx: yolov4 - CPU - Parallelonnx: yolov4 - CPU - Parallelopenvino: Handwritten English Recognition FP16-INT8 - CPUopenvino: Handwritten English Recognition FP16-INT8 - CPUonnx: ArcFace ResNet-100 - CPU - Parallelonnx: ArcFace ResNet-100 - CPU - Parallelspeedb: Rand Fillspeedb: Rand Fill Syncspeedb: Update Randrocksdb: Overwriterocksdb: Update Randonnx: ResNet50 v1-12-int8 - CPU - Parallelonnx: ResNet50 v1-12-int8 - CPU - Parallelspeedb: Rand Readonnx: super-resolution-10 - CPU - Parallelonnx: super-resolution-10 - CPU - Parallelopenssl: RSA4096openssl: RSA4096rocksdb: Rand Readstress-ng: Context Switchingospray-studio: 1 - 1080p - 1 - Path Tracer - CPUblender: Classroom - CPU-Onlytensorflow: CPU - 16 - AlexNettensorflow: CPU - 64 - AlexNetffmpeg: libx264 - Livey-cruncher: 500Mospray-studio: 1 - 4K - 16 - Path Tracer - CPUwebp2: Quality 75, Compression Effort 7hackbench: 16 - Threadvvenc: Bosphorus 1080p - Fastjpegxl: PNG - 90primesieve: 1e13svt-av1: Preset 12 - Bosphorus 4Konednn: IP Shapes 1D - CPUstress-ng: NUMAxmrig: CryptoNight-Femto UPX2 - 1Mxmrig: CryptoNight-Heavy - 1Msvt-av1: Preset 13 - Bosphorus 4Kxmrig: KawPow - 1Msvt-av1: Preset 4 - Bosphorus 1080pgromacs: MPI CPU - water_GMX50_barexmrig: Wownero - 1Mstress-ng: Mutexstress-ng: Socket Activitystress-ng: Zlibstress-ng: Mixed Schedulerstress-ng: Glibc Qsort Data Sortingstress-ng: Semaphoresstress-ng: CPU Stressstress-ng: SENDFILEstress-ng: Pollstress-ng: Hashstress-ng: AVL Treestress-ng: Vector Shufflehackbench: 32 - Processcompress-lz4: 9 - Decompression Speedcompress-lz4: 9 - Compression Speedcompress-lz4: 3 - Decompression Speedcompress-lz4: 3 - Compression Speedjpegxl-decode: Allcompress-lz4: 1 - Decompression Speedcompress-lz4: 1 - Compression Speedwebp2: Defaulttensorflow: CPU - 1 - AlexNetjpegxl: PNG - 100jpegxl: JPEG - 100onednn: Convolution Batch Shapes Auto - CPUoidn: RT.ldr_alb_nrm.3840x2160 - CPU-Onlyhackbench: 16 - Processencode-wavpack: WAV To WavPackonednn: IP Shapes 3D - CPUdraco: Church Facadewebp2: Quality 100, Compression Effort 5onednn: Deconvolution Batch shapes_3d - CPUprimesieve: 1e12draco: LionLinux 6.8Linux 6.9-rc234866677333619997170.3642.74271.9932763744.803906839102.950135.304.9028467842066770.101483647.03202.370.071.481801112.431.880.79307.5024821.320.38121702.129.84711.50569122.2368.22351747.0841.3599470.956614.265642.781923.461374029.01538111.57065.41191185.2946136376.2034.6047217268146675985245.6024.1222.311386742.7642.9421084.500.7867789.259.80514101.9012.89501353.57396.313.42357292.1731.84807541.8274.88181205.4142994073574032.477135.1810.5041210.7021229.2815186136.435939.817731442.6930377557.97143967.24231.2844.43649748.4956169.3680643523.241181758384748.54204365411.593531758.504065717006.813344.69962099112.69926093036.30241538.0924.280563514234669.7429600.02688.2636.601131.595735.403333.9861.0632.5629592.33468957.5929.73361720419.5116.055874353.635.86532361.213.41639599.26873217859304.38690917.2452120844501163551396.4038.4624.0526732.880.3448.9432.28723.16337.21331.2323.684621.3221085.1017.526336.78307.74119.9277.341446.4281.90451.1382.87445.909.5911638.5121.8254.98671.6618.337722.065067.40346.1036.281018.4742.683323.427630.820932.45039.9211260.9838.83506.8516147.577.1715532.5414.582530.0430.853626.4192.746710.7817126.3847.9132129.613777.5659.602116.783232649925744629218038801035982510.609894.24873804193896.55011152.6351358616.438204.636049851820343491.5246653.98213.35495.72181.516.483344120.6414.67915.91330.96939.619127.2024.50055482.6233002.232856.3123.24632860.818.66212.08536096.628490133.2847927.759670.3770407.961866.58154589461.19202073.081798203.1210389186.0217071694.52810.19576905.9224.7483537.032.613364.898.43386.2323601.0602.695.9017.2325.66626.0029.006212.3213.1126.3494.7985377478.356.165763.356612936866689667644430990.3740.20279.7433016744.24399180298.869034.864.9368167828417670.531541547.96203.630.071.531774113.991.900.77311.74510195.750.38119735.579.52411.51570112.4288.90936668.6991.5030065.996015.411041.504824.232875068.23121122.6015.10755196.0766116275.5634.1877336769166675899546.5624.8022.241380742.7742.8821644.580.7173754.749.46496105.78272.48234402.66997.533.40972293.0291.74973571.1904.67811213.7372966549905333.045128.7110.823610.509029.0693183937.433536.503031460.4229814480.67147978.33184.7855.41259753.3646427.3385035507.551175562392750.09200468629.323542258.833979416877.613341.46732598711.07694793862.76331139.3725.171263529230503.3429139.60688.7635.078832.829735.650733.0941.0433.8889602.73346567.8028.87541668719.8885.985908316.035.76033015.513.33535919.64855218740350.56680417.2832068588489161652700.0937.6983.8927790.240.3348.5682.23711.74337.78330.7323.974600.9520605.3917.586306.42302.77121.9377.571442.2179.23466.4579.00467.799.6511564.7822.0254.83673.4618.027522.125052.71341.9735.141051.3043.151123.172729.110734.35129.8911295.9235.73486.7616349.057.1415589.5913.892656.4231.013608.2194.333010.6006124.3768.0414230.203704.3856.947817.562332281625960329341439196836150210.511995.14413758374796.51127153.5641396904.038147.436158608319390362.9047853.82210.71471.88182.126.202349530.6114.29816.15230.12839.461126.1874.51166487.6032993.832826.1123.09833072.518.93411.76436180.728934422.1646123.329446.0770036.201858.70150005776.47200671.701900682.0810232390.4117199190.09808.99569593.0027.9943492.032.663349.698.41381.6573628.0602.545.6518.2825.60725.9608.810432.2213.3576.3404.8424577628.455.782023.2996054OpenBenchmarking.org

MariaDB mariadb-slap

OpenBenchmarking.orgQueries Per Second, More Is BetterMariaDB mariadb-slap 11.5Clients: 1024Linux 6.8Linux 6.9-rc2816243240SE +/- 1.23, N = 9SE +/- 1.41, N = 634361. (CXX) g++ options: -fPIC -pie -fstack-protector -O3 -shared -lrt -lpthread -lz -ldl -lm -lstdc++
OpenBenchmarking.orgQueries Per Second, More Is BetterMariaDB mariadb-slap 11.5Clients: 1024Linux 6.8Linux 6.9-rc2816243240Min: 30.57 / Avg: 34.19 / Max: 43.22Min: 29.77 / Avg: 36.25 / Max: 39.811. (CXX) g++ options: -fPIC -pie -fstack-protector -O3 -shared -lrt -lpthread -lz -ldl -lm -lstdc++

OpenBenchmarking.orgQueries Per Second, More Is BetterMariaDB mariadb-slap 11.5Clients: 512Linux 6.8Linux 6.9-rc220406080100SE +/- 0.41, N = 3SE +/- 0.87, N = 986861. (CXX) g++ options: -fPIC -pie -fstack-protector -O3 -shared -lrt -lpthread -lz -ldl -lm -lstdc++
OpenBenchmarking.orgQueries Per Second, More Is BetterMariaDB mariadb-slap 11.5Clients: 512Linux 6.8Linux 6.9-rc21632486480Min: 85.28 / Avg: 86.01 / Max: 86.69Min: 82.76 / Avg: 86.17 / Max: 91.391. (CXX) g++ options: -fPIC -pie -fstack-protector -O3 -shared -lrt -lpthread -lz -ldl -lm -lstdc++

Quicksilver

Quicksilver is a proxy application that represents some elements of the Mercury workload by solving a simplified dynamic Monte Carlo particle transport problem. Quicksilver is developed by Lawrence Livermore National Laboratory (LLNL) and this test profile currently makes use of the OpenMP CPU threaded code path. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFigure Of Merit, More Is BetterQuicksilver 20230818Input: CORAL2 P2Linux 6.8Linux 6.9-rc21.4M2.8M4.2M5.6M7MSE +/- 74194.56, N = 9SE +/- 91964.82, N = 9667733366896671. (CXX) g++ options: -fopenmp -O3 -march=native
OpenBenchmarking.orgFigure Of Merit, More Is BetterQuicksilver 20230818Input: CORAL2 P2Linux 6.8Linux 6.9-rc21.2M2.4M3.6M4.8M6MMin: 6277000 / Avg: 6677333.33 / Max: 6999000Min: 6300000 / Avg: 6689666.67 / Max: 70310001. (CXX) g++ options: -fopenmp -O3 -march=native

Stockfish

OpenBenchmarking.orgNodes Per Second, More Is BetterStockfish 16.1Chess BenchmarkLinux 6.8Linux 6.9-rc214M28M42M56M70MSE +/- 2993041.97, N = 9SE +/- 2705883.43, N = 961999717644430991. (CXX) g++ options: -lgcov -m64 -lpthread -fno-exceptions -std=c++17 -fno-peel-loops -fno-tracer -pedantic -O3 -funroll-loops -msse -msse3 -mpopcnt -mavx2 -mbmi -mavx512f -mavx512bw -mavx512vnni -mavx512dq -mavx512vl -msse4.1 -mssse3 -msse2 -mbmi2 -flto -flto-partition=one -flto=jobserver
OpenBenchmarking.orgNodes Per Second, More Is BetterStockfish 16.1Chess BenchmarkLinux 6.8Linux 6.9-rc211M22M33M44M55MMin: 48014524 / Avg: 61999717.22 / Max: 75476940Min: 49400530 / Avg: 64443099 / Max: 723116671. (CXX) g++ options: -lgcov -m64 -lpthread -fno-exceptions -std=c++17 -fno-peel-loops -fno-tracer -pedantic -O3 -funroll-loops -msse -msse3 -mpopcnt -mavx2 -mbmi -mavx512f -mavx512bw -mavx512vnni -mavx512dq -mavx512vl -msse4.1 -mssse3 -msse2 -mbmi2 -flto -flto-partition=one -flto=jobserver

Llama.cpp

Llama.cpp is a port of Facebook's LLaMA model in C/C++ developed by Georgi Gerganov. Llama.cpp allows the inference of LLaMA and other supported models in C/C++. For CPU inference Llama.cpp supports AVX2/AVX-512, ARM NEON, and other modern ISAs along with features like OpenBLAS usage. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b1808Model: llama-2-70b-chat.Q5_0.ggufLinux 6.8Linux 6.9-rc20.08330.16660.24990.33320.4165SE +/- 0.00, N = 3SE +/- 0.00, N = 30.360.371. (CXX) g++ options: -std=c++11 -fPIC -O3 -pthread -march=native -mtune=native -lopenblas
OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b1808Model: llama-2-70b-chat.Q5_0.ggufLinux 6.8Linux 6.9-rc212345Min: 0.36 / Avg: 0.36 / Max: 0.36Min: 0.37 / Avg: 0.37 / Max: 0.371. (CXX) g++ options: -std=c++11 -fPIC -O3 -pthread -march=native -mtune=native -lopenblas

FFmpeg

This is a benchmark of the FFmpeg multimedia framework. The FFmpeg test profile is making use of a modified version of vbench from Columbia University's Architecture and Design Lab (ARCADE) [http://arcade.cs.columbia.edu/vbench/] that is a benchmark for video-as-a-service workloads. The test profile offers the options of a range of vbench scenarios based on freely distributable video content and offers the options of using the x264 or x265 video encoders for transcoding. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFPS, More Is BetterFFmpeg 6.1Encoder: libx265 - Scenario: PlatformLinux 6.8Linux 6.9-rc21020304050SE +/- 0.71, N = 12SE +/- 0.46, N = 1242.7440.201. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
OpenBenchmarking.orgFPS, More Is BetterFFmpeg 6.1Encoder: libx265 - Scenario: PlatformLinux 6.8Linux 6.9-rc2918273645Min: 39.49 / Avg: 42.74 / Max: 46.96Min: 37.98 / Avg: 40.2 / Max: 42.841. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma

Blender

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.1Blend File: Barbershop - Compute: CPU-OnlyLinux 6.8Linux 6.9-rc260120180240300SE +/- 6.20, N = 8SE +/- 6.67, N = 9271.99279.74
OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.1Blend File: Barbershop - Compute: CPU-OnlyLinux 6.8Linux 6.9-rc250100150200250Min: 249.51 / Avg: 271.99 / Max: 302.23Min: 262.65 / Avg: 279.74 / Max: 312.46

Speedb

Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Sequential FillLinux 6.8Linux 6.9-rc270K140K210K280K350KSE +/- 1183.38, N = 3SE +/- 1955.03, N = 33276373301671. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Sequential FillLinux 6.8Linux 6.9-rc260K120K180K240K300KMin: 325425 / Avg: 327637 / Max: 329472Min: 326759 / Avg: 330167 / Max: 3335311. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

FFmpeg

This is a benchmark of the FFmpeg multimedia framework. The FFmpeg test profile is making use of a modified version of vbench from Columbia University's Architecture and Design Lab (ARCADE) [http://arcade.cs.columbia.edu/vbench/] that is a benchmark for video-as-a-service workloads. The test profile offers the options of a range of vbench scenarios based on freely distributable video content and offers the options of using the x264 or x265 video encoders for transcoding. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFPS, More Is BetterFFmpeg 6.1Encoder: libx265 - Scenario: Video On DemandLinux 6.8Linux 6.9-rc21020304050SE +/- 0.50, N = 12SE +/- 0.76, N = 944.8044.241. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
OpenBenchmarking.orgFPS, More Is BetterFFmpeg 6.1Encoder: libx265 - Scenario: Video On DemandLinux 6.8Linux 6.9-rc2918273645Min: 42.52 / Avg: 44.8 / Max: 47.48Min: 41.46 / Avg: 44.24 / Max: 47.61. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma

BRL-CAD

OpenBenchmarking.orgVGR Performance Metric, More Is BetterBRL-CAD 7.38.2VGR Performance MetricLinux 6.8Linux 6.9-rc2900K1800K2700K3600K4500K390683939918021. (CXX) g++ options: -std=c++17 -pipe -fvisibility=hidden -fno-strict-aliasing -fno-common -fexceptions -ftemplate-depth-128 -m64 -ggdb3 -O3 -fipa-pta -fstrength-reduce -finline-functions -flto -ltcl8.6 -lnetpbm -lregex_brl -lz_brl -lassimp -ldl -lm -ltk8.6

OSPRay

OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 3.1Benchmark: particle_volume/pathtracer/real_timeLinux 6.8Linux 6.9-rc220406080100SE +/- 1.85, N = 12SE +/- 1.98, N = 9102.9598.87
OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 3.1Benchmark: particle_volume/pathtracer/real_timeLinux 6.8Linux 6.9-rc220406080100Min: 90.18 / Avg: 102.95 / Max: 112.87Min: 90.47 / Avg: 98.87 / Max: 107.19

TensorFlow

OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 32 - Model: ResNet-50Linux 6.8Linux 6.9-rc2816243240SE +/- 0.48, N = 15SE +/- 0.42, N = 1535.3034.86
OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 32 - Model: ResNet-50Linux 6.8Linux 6.9-rc2816243240Min: 31.45 / Avg: 35.3 / Max: 37.99Min: 31.08 / Avg: 34.86 / Max: 37.24

VVenC

OpenBenchmarking.orgFrames Per Second, More Is BetterVVenC 1.11Video Input: Bosphorus 4K - Video Preset: FastLinux 6.8Linux 6.9-rc21.11062.22123.33184.44245.553SE +/- 0.075, N = 10SE +/- 0.077, N = 124.9024.9361. (CXX) g++ options: -O3 -flto=auto -fno-fat-lto-objects
OpenBenchmarking.orgFrames Per Second, More Is BetterVVenC 1.11Video Input: Bosphorus 4K - Video Preset: FastLinux 6.8Linux 6.9-rc2246810Min: 4.5 / Avg: 4.9 / Max: 5.22Min: 4.45 / Avg: 4.94 / Max: 5.331. (CXX) g++ options: -O3 -flto=auto -fno-fat-lto-objects

OpenSSL

OpenSSL is an open-source toolkit that implements SSL (Secure Sockets Layer) and TLS (Transport Layer Security) protocols. The system/openssl test profiles relies on benchmarking the system/OS-supplied openssl binary rather than the pts/openssl test profile that uses the locally-built OpenSSL for benchmarking. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgbyte/s, More Is BetterOpenSSLAlgorithm: SHA256Linux 6.8Linux 6.9-rc220000M40000M60000M80000M100000MSE +/- 767626108.54, N = 3SE +/- 717052715.91, N = 1284678420667816782841761. OpenSSL 3.0.10 1 Aug 2023 (Library: OpenSSL 3.0.10 1 Aug 2023)
OpenBenchmarking.orgbyte/s, More Is BetterOpenSSLAlgorithm: SHA256Linux 6.8Linux 6.9-rc215000M30000M45000M60000M75000MMin: 83194731580 / Avg: 84678420666.67 / Max: 85761951540Min: 78428217190 / Avg: 81678284175.83 / Max: 857650217201. OpenSSL 3.0.10 1 Aug 2023 (Library: OpenSSL 3.0.10 1 Aug 2023)

TensorFlow

OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 256 - Model: ResNet-50Linux 6.8Linux 6.9-rc21632486480SE +/- 0.38, N = 3SE +/- 0.47, N = 370.1070.53
OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 256 - Model: ResNet-50Linux 6.8Linux 6.9-rc21428425670Min: 69.38 / Avg: 70.1 / Max: 70.68Min: 69.89 / Avg: 70.53 / Max: 71.45

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 2 - Resolution: 1080p - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc23K6K9K12K15KSE +/- 268.32, N = 12SE +/- 235.77, N = 151483615415
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 2 - Resolution: 1080p - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc23K6K9K12K15KMin: 13942 / Avg: 14835.67 / Max: 17063Min: 14120 / Avg: 15414.73 / Max: 17087

TensorFlow

OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 64 - Model: ResNet-50Linux 6.8Linux 6.9-rc21122334455SE +/- 0.54, N = 12SE +/- 0.42, N = 347.0347.96
OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 64 - Model: ResNet-50Linux 6.8Linux 6.9-rc21020304050Min: 43.05 / Avg: 47.03 / Max: 49.27Min: 47.25 / Avg: 47.96 / Max: 48.71

OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 256 - Model: GoogLeNetLinux 6.8Linux 6.9-rc24080120160200SE +/- 2.20, N = 12SE +/- 1.95, N = 3202.37203.63
OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 256 - Model: GoogLeNetLinux 6.8Linux 6.9-rc24080120160200Min: 192.36 / Avg: 202.37 / Max: 212.28Min: 200.52 / Avg: 203.63 / Max: 207.21

WebP2 Image Encode

This is a test of Google's libwebp2 library with the WebP2 image encode utility and using a sample 6000x4000 pixel JPEG image as the input, similar to the WebP/libwebp test profile. WebP2 is currently experimental and under heavy development as ultimately the successor to WebP. WebP2 supports 10-bit HDR, more efficienct lossy compression, improved lossless compression, animation support, and full multi-threading support compared to WebP. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMP/s, More Is BetterWebP2 Image Encode 20220823Encode Settings: Quality 100, Lossless CompressionLinux 6.8Linux 6.9-rc20.01580.03160.04740.06320.079SE +/- 0.00, N = 3SE +/- 0.00, N = 30.070.071. (CXX) g++ options: -msse4.2 -fno-rtti -O3 -ldl
OpenBenchmarking.orgMP/s, More Is BetterWebP2 Image Encode 20220823Encode Settings: Quality 100, Lossless CompressionLinux 6.8Linux 6.9-rc212345Min: 0.07 / Avg: 0.07 / Max: 0.07Min: 0.07 / Avg: 0.07 / Max: 0.071. (CXX) g++ options: -msse4.2 -fno-rtti -O3 -ldl

Llama.cpp

Llama.cpp is a port of Facebook's LLaMA model in C/C++ developed by Georgi Gerganov. Llama.cpp allows the inference of LLaMA and other supported models in C/C++. For CPU inference Llama.cpp supports AVX2/AVX-512, ARM NEON, and other modern ISAs along with features like OpenBLAS usage. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b1808Model: llama-2-13b.Q4_0.ggufLinux 6.8Linux 6.9-rc20.34430.68861.03291.37721.7215SE +/- 0.01, N = 3SE +/- 0.02, N = 31.481.531. (CXX) g++ options: -std=c++11 -fPIC -O3 -pthread -march=native -mtune=native -lopenblas
OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b1808Model: llama-2-13b.Q4_0.ggufLinux 6.8Linux 6.9-rc2246810Min: 1.46 / Avg: 1.48 / Max: 1.5Min: 1.49 / Avg: 1.53 / Max: 1.561. (CXX) g++ options: -std=c++11 -fPIC -O3 -pthread -march=native -mtune=native -lopenblas

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 1 - Resolution: 4K - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc2400800120016002000SE +/- 40.26, N = 15SE +/- 22.91, N = 1518011774
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 1 - Resolution: 4K - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc230060090012001500Min: 1660 / Avg: 1800.87 / Max: 2211Min: 1679 / Avg: 1773.8 / Max: 2040

FFmpeg

This is a benchmark of the FFmpeg multimedia framework. The FFmpeg test profile is making use of a modified version of vbench from Columbia University's Architecture and Design Lab (ARCADE) [http://arcade.cs.columbia.edu/vbench/] that is a benchmark for video-as-a-service workloads. The test profile offers the options of a range of vbench scenarios based on freely distributable video content and offers the options of using the x264 or x265 video encoders for transcoding. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFPS, More Is BetterFFmpeg 6.1Encoder: libx265 - Scenario: LiveLinux 6.8Linux 6.9-rc2306090120150SE +/- 1.36, N = 15SE +/- 1.80, N = 15112.43113.991. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
OpenBenchmarking.orgFPS, More Is BetterFFmpeg 6.1Encoder: libx265 - Scenario: LiveLinux 6.8Linux 6.9-rc220406080100Min: 105.9 / Avg: 112.43 / Max: 124.31Min: 100.03 / Avg: 113.99 / Max: 125.221. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma

TensorFlow

OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 1 - Model: ResNet-50Linux 6.8Linux 6.9-rc20.42750.8551.28251.712.1375SE +/- 0.03, N = 15SE +/- 0.03, N = 151.881.90
OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 1 - Model: ResNet-50Linux 6.8Linux 6.9-rc2246810Min: 1.65 / Avg: 1.88 / Max: 2.13Min: 1.66 / Avg: 1.9 / Max: 2.07

Llamafile

Mozilla's Llamafile allows distributing and running large language models (LLMs) as a single file. Llamafile aims to make open-source LLMs more accessible to developers and users. Llamafile supports a variety of models, CPUs and GPUs, and other options. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTokens Per Second, More Is BetterLlamafile 0.6Test: wizardcoder-python-34b-v1.0.Q6_K - Acceleration: CPULinux 6.8Linux 6.9-rc20.17780.35560.53340.71120.889SE +/- 0.01, N = 4SE +/- 0.00, N = 30.790.77
OpenBenchmarking.orgTokens Per Second, More Is BetterLlamafile 0.6Test: wizardcoder-python-34b-v1.0.Q6_K - Acceleration: CPULinux 6.8Linux 6.9-rc2246810Min: 0.77 / Avg: 0.79 / Max: 0.81Min: 0.76 / Avg: 0.77 / Max: 0.77

Timed Linux Kernel Compilation

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Linux Kernel Compilation 6.8Build: allmodconfigLinux 6.8Linux 6.9-rc270140210280350SE +/- 3.84, N = 3SE +/- 1.26, N = 3307.50311.75
OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Linux Kernel Compilation 6.8Build: allmodconfigLinux 6.8Linux 6.9-rc260120180240300Min: 301.03 / Avg: 307.5 / Max: 314.3Min: 309.92 / Avg: 311.74 / Max: 314.15

oneDNN

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: Recurrent Neural Network Training - Engine: CPULinux 6.8Linux 6.9-rc22K4K6K8K10KSE +/- 22.69, N = 3SE +/- 498.89, N = 134821.3210195.75MIN: 4659.99MIN: 4568.31. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: Recurrent Neural Network Training - Engine: CPULinux 6.8Linux 6.9-rc22K4K6K8K10KMin: 4790.9 / Avg: 4821.32 / Max: 4865.7Min: 7116.6 / Avg: 10195.75 / Max: 12117.81. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

OpenVINO

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Age Gender Recognition Retail 0013 FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc20.08550.1710.25650.3420.4275SE +/- 0.01, N = 15SE +/- 0.00, N = 150.380.38MIN: 0.23 / MAX: 46.5MIN: 0.24 / MAX: 50.651. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Age Gender Recognition Retail 0013 FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc212345Min: 0.36 / Avg: 0.38 / Max: 0.43Min: 0.36 / Avg: 0.38 / Max: 0.41. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Age Gender Recognition Retail 0013 FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc230K60K90K120K150KSE +/- 1455.69, N = 15SE +/- 1149.56, N = 15121702.12119735.571. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Age Gender Recognition Retail 0013 FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc220K40K60K80K100KMin: 111011.67 / Avg: 121702.12 / Max: 128851.87Min: 110782.85 / Avg: 119735.57 / Max: 127032.621. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

VVenC

OpenBenchmarking.orgFrames Per Second, More Is BetterVVenC 1.11Video Input: Bosphorus 4K - Video Preset: FasterLinux 6.8Linux 6.9-rc23691215SE +/- 0.200, N = 12SE +/- 0.172, N = 159.8479.5241. (CXX) g++ options: -O3 -flto=auto -fno-fat-lto-objects
OpenBenchmarking.orgFrames Per Second, More Is BetterVVenC 1.11Video Input: Bosphorus 4K - Video Preset: FasterLinux 6.8Linux 6.9-rc23691215Min: 8.73 / Avg: 9.85 / Max: 10.84Min: 8.19 / Avg: 9.52 / Max: 10.691. (CXX) g++ options: -O3 -flto=auto -fno-fat-lto-objects

FFmpeg

This is a benchmark of the FFmpeg multimedia framework. The FFmpeg test profile is making use of a modified version of vbench from Columbia University's Architecture and Design Lab (ARCADE) [http://arcade.cs.columbia.edu/vbench/] that is a benchmark for video-as-a-service workloads. The test profile offers the options of a range of vbench scenarios based on freely distributable video content and offers the options of using the x264 or x265 video encoders for transcoding. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFPS, More Is BetterFFmpeg 6.1Encoder: libx264 - Scenario: UploadLinux 6.8Linux 6.9-rc23691215SE +/- 0.01, N = 3SE +/- 0.01, N = 311.5011.511. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
OpenBenchmarking.orgFPS, More Is BetterFFmpeg 6.1Encoder: libx264 - Scenario: UploadLinux 6.8Linux 6.9-rc23691215Min: 11.49 / Avg: 11.5 / Max: 11.52Min: 11.5 / Avg: 11.51 / Max: 11.521. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 3 - Resolution: 1080p - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc2120240360480600SE +/- 4.36, N = 15SE +/- 4.85, N = 15569570
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 3 - Resolution: 1080p - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc2100200300400500Min: 541 / Avg: 568.93 / Max: 611Min: 540 / Avg: 569.6 / Max: 609

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.

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: yolov4 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2306090120150SE +/- 2.31, N = 15SE +/- 1.33, N = 13122.24112.431. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: yolov4 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc220406080100Min: 106.47 / Avg: 122.24 / Max: 135.26Min: 104.1 / Avg: 112.43 / Max: 120.051. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: yolov4 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2246810SE +/- 0.16273, N = 15SE +/- 0.10643, N = 138.223518.909361. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: yolov4 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2306090120150SE +/- 2.31, N = 15SE +/- 1.33, N = 13122.24112.431. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: yolov4 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc23691215Min: 7.39 / Avg: 8.22 / Max: 9.39Min: 8.33 / Avg: 8.91 / Max: 9.611. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: fcn-resnet101-11 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc2160320480640800SE +/- 25.74, N = 15SE +/- 14.65, N = 12747.08668.701. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: fcn-resnet101-11 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc2130260390520650Min: 610.95 / Avg: 747.08 / Max: 916.61Min: 615.36 / Avg: 668.7 / Max: 762.921. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: fcn-resnet101-11 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc20.33820.67641.01461.35281.691SE +/- 0.04472, N = 15SE +/- 0.03151, N = 121.359941.503001. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: fcn-resnet101-11 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc2160320480640800SE +/- 25.74, N = 15SE +/- 14.65, N = 12747.08668.701. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: fcn-resnet101-11 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc2246810Min: 1.09 / Avg: 1.36 / Max: 1.64Min: 1.31 / Avg: 1.5 / Max: 1.631. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: bertsquad-12 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc21632486480SE +/- 2.24, N = 12SE +/- 2.28, N = 1570.9666.001. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: bertsquad-12 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc21428425670Min: 54.79 / Avg: 70.96 / Max: 82.06Min: 55.02 / Avg: 66 / Max: 79.781. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: bertsquad-12 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc248121620SE +/- 0.50, N = 12SE +/- 0.54, N = 1514.2715.411. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: bertsquad-12 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc21632486480SE +/- 2.24, N = 12SE +/- 2.28, N = 1570.9666.001. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: bertsquad-12 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc248121620Min: 12.19 / Avg: 14.27 / Max: 18.25Min: 12.53 / Avg: 15.41 / Max: 18.171. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: ArcFace ResNet-100 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc21020304050SE +/- 0.81, N = 12SE +/- 0.86, N = 1542.7841.501. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: ArcFace ResNet-100 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2918273645Min: 39.16 / Avg: 42.78 / Max: 47.49Min: 37.7 / Avg: 41.5 / Max: 47.91. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: ArcFace ResNet-100 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2612182430SE +/- 0.43, N = 12SE +/- 0.48, N = 1523.4624.231. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: ArcFace ResNet-100 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc21020304050SE +/- 0.81, N = 12SE +/- 0.86, N = 1542.7841.501. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: ArcFace ResNet-100 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2612182430Min: 21.06 / Avg: 23.46 / Max: 25.54Min: 20.88 / Avg: 24.23 / Max: 26.521. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 2 - Resolution: 1080p - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc216003200480064008000SE +/- 105.66, N = 3SE +/- 118.69, N = 1574027506
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 2 - Resolution: 1080p - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc213002600390052006500Min: 7201 / Avg: 7402 / Max: 7559Min: 6963 / Avg: 7505.73 / Max: 8565

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.

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: ResNet50 v1-12-int8 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc23691215SE +/- 0.20919, N = 12SE +/- 0.21896, N = 159.015388.231211. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: ResNet50 v1-12-int8 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc23691215Min: 8.1 / Avg: 9.02 / Max: 10.07Min: 7.46 / Avg: 8.23 / Max: 9.971. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: ResNet50 v1-12-int8 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2306090120150SE +/- 2.58, N = 12SE +/- 3.02, N = 15111.57122.601. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: ResNet50 v1-12-int8 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc23691215SE +/- 0.20919, N = 12SE +/- 0.21896, N = 159.015388.231211. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: ResNet50 v1-12-int8 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc220406080100Min: 99.34 / Avg: 111.57 / Max: 123.41Min: 100.28 / Avg: 122.6 / Max: 134.091. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: T5 Encoder - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc21.21772.43543.65314.87086.0885SE +/- 0.09595, N = 12SE +/- 0.06160, N = 145.411915.107551. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: T5 Encoder - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc2246810Min: 4.97 / Avg: 5.41 / Max: 6.2Min: 4.66 / Avg: 5.11 / Max: 5.511. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: T5 Encoder - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc24080120160200SE +/- 3.10, N = 12SE +/- 2.39, N = 14185.29196.081. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: T5 Encoder - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc21.21772.43543.65314.87086.0885SE +/- 0.09595, N = 12SE +/- 0.06160, N = 145.411915.107551. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: T5 Encoder - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc24080120160200Min: 161.21 / Avg: 185.29 / Max: 201.13Min: 181.45 / Avg: 196.08 / Max: 214.31. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 1 - Resolution: 4K - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc213K26K39K52K65KSE +/- 443.96, N = 15SE +/- 567.54, N = 76136361162
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 1 - Resolution: 4K - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc211K22K33K44K55KMin: 58023 / Avg: 61363 / Max: 64305Min: 60047 / Avg: 61161.86 / Max: 64493

Blender

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.1Blend File: Pabellon Barcelona - Compute: CPU-OnlyLinux 6.8Linux 6.9-rc220406080100SE +/- 0.83, N = 5SE +/- 0.61, N = 1576.2075.56
OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.1Blend File: Pabellon Barcelona - Compute: CPU-OnlyLinux 6.8Linux 6.9-rc21530456075Min: 74.42 / Avg: 76.2 / Max: 79.26Min: 72.66 / Avg: 75.56 / Max: 80.08

JPEG-XL libjxl

OpenBenchmarking.orgMP/s, More Is BetterJPEG-XL libjxl 0.10.1Input: JPEG - Quality: 80Linux 6.8Linux 6.9-rc2816243240SE +/- 0.51, N = 12SE +/- 0.53, N = 1534.6034.191. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm
OpenBenchmarking.orgMP/s, More Is BetterJPEG-XL libjxl 0.10.1Input: JPEG - Quality: 80Linux 6.8Linux 6.9-rc2714212835Min: 32.28 / Avg: 34.6 / Max: 38.1Min: 31.32 / Avg: 34.19 / Max: 37.851. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 3 - Resolution: 4K - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc216K32K48K64K80KSE +/- 506.90, N = 3SE +/- 855.02, N = 157217273367
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 3 - Resolution: 4K - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc213K26K39K52K65KMin: 71200 / Avg: 72171.67 / Max: 72908Min: 70780 / Avg: 73366.67 / Max: 81395

Quicksilver

Quicksilver is a proxy application that represents some elements of the Mercury workload by solving a simplified dynamic Monte Carlo particle transport problem. Quicksilver is developed by Lawrence Livermore National Laboratory (LLNL) and this test profile currently makes use of the OpenMP CPU threaded code path. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFigure Of Merit, More Is BetterQuicksilver 20230818Input: CORAL2 P1Linux 6.8Linux 6.9-rc21.5M3M4.5M6M7.5MSE +/- 67589.78, N = 6SE +/- 46491.34, N = 3681466769166671. (CXX) g++ options: -fopenmp -O3 -march=native
OpenBenchmarking.orgFigure Of Merit, More Is BetterQuicksilver 20230818Input: CORAL2 P1Linux 6.8Linux 6.9-rc21.2M2.4M3.6M4.8M6MMin: 6672000 / Avg: 6814666.67 / Max: 7128000Min: 6861000 / Avg: 6916666.67 / Max: 70090001. (CXX) g++ options: -fopenmp -O3 -march=native

Chaos Group V-RAY

OpenBenchmarking.orgvsamples, More Is BetterChaos Group V-RAY 6.0Mode: CPULinux 6.8Linux 6.9-rc213K26K39K52K65KSE +/- 530.58, N = 15SE +/- 557.85, N = 35985258995
OpenBenchmarking.orgvsamples, More Is BetterChaos Group V-RAY 6.0Mode: CPULinux 6.8Linux 6.9-rc210K20K30K40K50KMin: 55464 / Avg: 59851.8 / Max: 62789Min: 57945 / Avg: 58994.67 / Max: 59847

Blender

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.1Blend File: Junkshop - Compute: CPU-OnlyLinux 6.8Linux 6.9-rc21122334455SE +/- 0.47, N = 15SE +/- 0.40, N = 1545.6046.56
OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.1Blend File: Junkshop - Compute: CPU-OnlyLinux 6.8Linux 6.9-rc2918273645Min: 42.49 / Avg: 45.6 / Max: 49.86Min: 44.93 / Avg: 46.56 / Max: 51.27

TensorFlow

OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 16 - Model: ResNet-50Linux 6.8Linux 6.9-rc2612182430SE +/- 0.30, N = 3SE +/- 0.28, N = 1524.1224.80
OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 16 - Model: ResNet-50Linux 6.8Linux 6.9-rc2612182430Min: 23.76 / Avg: 24.12 / Max: 24.72Min: 22.27 / Avg: 24.8 / Max: 26.2

FFmpeg

This is a benchmark of the FFmpeg multimedia framework. The FFmpeg test profile is making use of a modified version of vbench from Columbia University's Architecture and Design Lab (ARCADE) [http://arcade.cs.columbia.edu/vbench/] that is a benchmark for video-as-a-service workloads. The test profile offers the options of a range of vbench scenarios based on freely distributable video content and offers the options of using the x264 or x265 video encoders for transcoding. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFPS, More Is BetterFFmpeg 6.1Encoder: libx265 - Scenario: UploadLinux 6.8Linux 6.9-rc2510152025SE +/- 0.31, N = 3SE +/- 0.26, N = 422.3122.241. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
OpenBenchmarking.orgFPS, More Is BetterFFmpeg 6.1Encoder: libx265 - Scenario: UploadLinux 6.8Linux 6.9-rc2510152025Min: 21.72 / Avg: 22.31 / Max: 22.78Min: 21.62 / Avg: 22.24 / Max: 22.841. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 1 - Resolution: 1080p - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc23K6K9K12K15KSE +/- 112.81, N = 9SE +/- 180.26, N = 31386713807
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 1 - Resolution: 1080p - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc22K4K6K8K10KMin: 13341 / Avg: 13867.33 / Max: 14375Min: 13502 / Avg: 13807.33 / Max: 14126

FFmpeg

This is a benchmark of the FFmpeg multimedia framework. The FFmpeg test profile is making use of a modified version of vbench from Columbia University's Architecture and Design Lab (ARCADE) [http://arcade.cs.columbia.edu/vbench/] that is a benchmark for video-as-a-service workloads. The test profile offers the options of a range of vbench scenarios based on freely distributable video content and offers the options of using the x264 or x265 video encoders for transcoding. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFPS, More Is BetterFFmpeg 6.1Encoder: libx264 - Scenario: PlatformLinux 6.8Linux 6.9-rc21020304050SE +/- 0.17, N = 3SE +/- 0.10, N = 342.7642.771. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
OpenBenchmarking.orgFPS, More Is BetterFFmpeg 6.1Encoder: libx264 - Scenario: PlatformLinux 6.8Linux 6.9-rc2918273645Min: 42.42 / Avg: 42.76 / Max: 43Min: 42.66 / Avg: 42.77 / Max: 42.971. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma

OpenBenchmarking.orgFPS, More Is BetterFFmpeg 6.1Encoder: libx264 - Scenario: Video On DemandLinux 6.8Linux 6.9-rc21020304050SE +/- 0.06, N = 3SE +/- 0.10, N = 342.9442.881. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
OpenBenchmarking.orgFPS, More Is BetterFFmpeg 6.1Encoder: libx264 - Scenario: Video On DemandLinux 6.8Linux 6.9-rc2918273645Min: 42.81 / Avg: 42.94 / Max: 43.01Min: 42.69 / Avg: 42.88 / Max: 43.051. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 3 - Resolution: 4K - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc25001000150020002500SE +/- 21.99, N = 4SE +/- 26.16, N = 1321082164
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 3 - Resolution: 4K - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc2400800120016002000Min: 2071 / Avg: 2108.25 / Max: 2172Min: 2071 / Avg: 2164.38 / Max: 2437

Llamafile

Mozilla's Llamafile allows distributing and running large language models (LLMs) as a single file. Llamafile aims to make open-source LLMs more accessible to developers and users. Llamafile supports a variety of models, CPUs and GPUs, and other options. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTokens Per Second, More Is BetterLlamafile 0.6Test: mistral-7b-instruct-v0.2.Q8_0 - Acceleration: CPULinux 6.8Linux 6.9-rc21.03052.0613.09154.1225.1525SE +/- 0.04, N = 3SE +/- 0.03, N = 34.504.58
OpenBenchmarking.orgTokens Per Second, More Is BetterLlamafile 0.6Test: mistral-7b-instruct-v0.2.Q8_0 - Acceleration: CPULinux 6.8Linux 6.9-rc2246810Min: 4.41 / Avg: 4.5 / Max: 4.55Min: 4.53 / Avg: 4.58 / Max: 4.64

OpenVINO

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Age Gender Recognition Retail 0013 FP16 - Device: CPULinux 6.8Linux 6.9-rc20.17550.3510.52650.7020.8775SE +/- 0.01, N = 15SE +/- 0.03, N = 30.780.71MIN: 0.3 / MAX: 138.08MIN: 0.32 / MAX: 74.711. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Age Gender Recognition Retail 0013 FP16 - Device: CPULinux 6.8Linux 6.9-rc2246810Min: 0.72 / Avg: 0.78 / Max: 0.84Min: 0.66 / Avg: 0.71 / Max: 0.751. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Age Gender Recognition Retail 0013 FP16 - Device: CPULinux 6.8Linux 6.9-rc216K32K48K64K80KSE +/- 711.41, N = 15SE +/- 1055.18, N = 367789.2573754.741. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Age Gender Recognition Retail 0013 FP16 - Device: CPULinux 6.8Linux 6.9-rc213K26K39K52K65KMin: 63890.84 / Avg: 67789.25 / Max: 73660.57Min: 71774.24 / Avg: 73754.74 / Max: 75376.191. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -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.

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: GPT-2 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc23691215SE +/- 0.11034, N = 3SE +/- 0.12308, N = 159.805149.464961. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: GPT-2 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc23691215Min: 9.59 / Avg: 9.81 / Max: 9.93Min: 8.88 / Avg: 9.46 / Max: 10.51. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: GPT-2 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc220406080100SE +/- 1.16, N = 3SE +/- 1.34, N = 15101.90105.781. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: GPT-2 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc23691215SE +/- 0.11034, N = 3SE +/- 0.12308, N = 159.805149.464961. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: GPT-2 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc220406080100Min: 100.57 / Avg: 101.9 / Max: 104.21Min: 95.15 / Avg: 105.78 / Max: 112.551. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: T5 Encoder - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc20.65141.30281.95422.60563.257SE +/- 0.12453, N = 15SE +/- 0.01855, N = 32.895012.482341. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: T5 Encoder - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2246810Min: 2.46 / Avg: 2.9 / Max: 3.77Min: 2.46 / Avg: 2.48 / Max: 2.521. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: T5 Encoder - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc290180270360450SE +/- 14.02, N = 15SE +/- 3.01, N = 3353.57402.671. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: T5 Encoder - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc20.65141.30281.95422.60563.257SE +/- 0.12453, N = 15SE +/- 0.01855, N = 32.895012.482341. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: T5 Encoder - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc270140210280350Min: 265.29 / Avg: 353.57 / Max: 406.48Min: 396.72 / Avg: 402.67 / Max: 406.451. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

TensorFlow

OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 32 - Model: GoogLeNetLinux 6.8Linux 6.9-rc220406080100SE +/- 1.27, N = 12SE +/- 1.10, N = 1596.3197.53
OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 32 - Model: GoogLeNetLinux 6.8Linux 6.9-rc220406080100Min: 85.81 / Avg: 96.31 / Max: 102.25Min: 88.74 / Avg: 97.53 / Max: 104.65

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.

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: CaffeNet 12-int8 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc20.77031.54062.31093.08123.8515SE +/- 0.03367, N = 15SE +/- 0.02459, N = 33.423573.409721. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: CaffeNet 12-int8 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc2246810Min: 3.26 / Avg: 3.42 / Max: 3.7Min: 3.38 / Avg: 3.41 / Max: 3.461. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: CaffeNet 12-int8 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc260120180240300SE +/- 2.79, N = 15SE +/- 2.09, N = 3292.17293.031. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: CaffeNet 12-int8 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc20.77031.54062.31093.08123.8515SE +/- 0.03367, N = 15SE +/- 0.02459, N = 33.423573.409721. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: CaffeNet 12-int8 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc250100150200250Min: 269.79 / Avg: 292.17 / Max: 306.27Min: 288.88 / Avg: 293.03 / Max: 295.511. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: CaffeNet 12-int8 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc20.41580.83161.24741.66322.079SE +/- 0.02432, N = 15SE +/- 0.01721, N = 31.848071.749731. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: CaffeNet 12-int8 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2246810Min: 1.78 / Avg: 1.85 / Max: 2.14Min: 1.72 / Avg: 1.75 / Max: 1.781. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: CaffeNet 12-int8 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2120240360480600SE +/- 6.47, N = 15SE +/- 5.63, N = 3541.83571.191. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: CaffeNet 12-int8 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc20.41580.83161.24741.66322.079SE +/- 0.02432, N = 15SE +/- 0.01721, N = 31.848071.749731. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: CaffeNet 12-int8 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2100200300400500Min: 466.87 / Avg: 541.83 / Max: 562.29Min: 561.47 / Avg: 571.19 / Max: 580.971. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: super-resolution-10 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc21.09842.19683.29524.39365.492SE +/- 0.07270, N = 15SE +/- 0.01745, N = 34.881814.678111. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: super-resolution-10 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2246810Min: 4.64 / Avg: 4.88 / Max: 5.34Min: 4.64 / Avg: 4.68 / Max: 4.71. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: super-resolution-10 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc250100150200250SE +/- 2.99, N = 15SE +/- 0.80, N = 3205.41213.741. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: super-resolution-10 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc21.09842.19683.29524.39365.492SE +/- 0.07270, N = 15SE +/- 0.01745, N = 34.881814.678111. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: super-resolution-10 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc24080120160200Min: 187.3 / Avg: 205.41 / Max: 215.32Min: 212.71 / Avg: 213.74 / Max: 215.321. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenSSL

OpenSSL is an open-source toolkit that implements SSL (Secure Sockets Layer) and TLS (Transport Layer Security) protocols. The system/openssl test profiles relies on benchmarking the system/OS-supplied openssl binary rather than the pts/openssl test profile that uses the locally-built OpenSSL for benchmarking. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgbyte/s, More Is BetterOpenSSLAlgorithm: SHA512Linux 6.8Linux 6.9-rc26000M12000M18000M24000M30000MSE +/- 92272355.83, N = 3SE +/- 252323355.66, N = 329940735740296654990531. OpenSSL 3.0.10 1 Aug 2023 (Library: OpenSSL 3.0.10 1 Aug 2023)
OpenBenchmarking.orgbyte/s, More Is BetterOpenSSLAlgorithm: SHA512Linux 6.8Linux 6.9-rc25000M10000M15000M20000M25000MMin: 29758092770 / Avg: 29940735740 / Max: 30054942200Min: 29221344810 / Avg: 29665499053.33 / Max: 300950526401. OpenSSL 3.0.10 1 Aug 2023 (Library: OpenSSL 3.0.10 1 Aug 2023)

JPEG-XL libjxl

OpenBenchmarking.orgMP/s, More Is BetterJPEG-XL libjxl 0.10.1Input: PNG - Quality: 80Linux 6.8Linux 6.9-rc2816243240SE +/- 0.38, N = 3SE +/- 0.36, N = 1532.4833.051. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm
OpenBenchmarking.orgMP/s, More Is BetterJPEG-XL libjxl 0.10.1Input: PNG - Quality: 80Linux 6.8Linux 6.9-rc2714212835Min: 31.76 / Avg: 32.48 / Max: 33.06Min: 30.9 / Avg: 33.05 / Max: 35.141. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm

TensorFlow

OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 64 - Model: GoogLeNetLinux 6.8Linux 6.9-rc2306090120150SE +/- 0.94, N = 3SE +/- 1.70, N = 15135.18128.71
OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 64 - Model: GoogLeNetLinux 6.8Linux 6.9-rc2306090120150Min: 133.71 / Avg: 135.18 / Max: 136.92Min: 119.1 / Avg: 128.71 / Max: 141.51

OSPRay

OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 3.1Benchmark: gravity_spheres_volume/dim_512/scivis/real_timeLinux 6.8Linux 6.9-rc23691215SE +/- 0.08, N = 15SE +/- 0.07, N = 310.5010.82
OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 3.1Benchmark: gravity_spheres_volume/dim_512/scivis/real_timeLinux 6.8Linux 6.9-rc23691215Min: 9.91 / Avg: 10.5 / Max: 11.03Min: 10.69 / Avg: 10.82 / Max: 10.9

OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 3.1Benchmark: gravity_spheres_volume/dim_512/ao/real_timeLinux 6.8Linux 6.9-rc23691215SE +/- 0.14, N = 15SE +/- 0.05, N = 310.7010.51
OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 3.1Benchmark: gravity_spheres_volume/dim_512/ao/real_timeLinux 6.8Linux 6.9-rc23691215Min: 9.61 / Avg: 10.7 / Max: 11.69Min: 10.42 / Avg: 10.51 / Max: 10.57

OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 3.1Benchmark: particle_volume/scivis/real_timeLinux 6.8Linux 6.9-rc2714212835SE +/- 0.26, N = 3SE +/- 0.33, N = 329.2829.07
OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 3.1Benchmark: particle_volume/scivis/real_timeLinux 6.8Linux 6.9-rc2612182430Min: 28.76 / Avg: 29.28 / Max: 29.6Min: 28.57 / Avg: 29.07 / Max: 29.68

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 2 - Resolution: 4K - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc2400800120016002000SE +/- 13.93, N = 11SE +/- 26.21, N = 318611839
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 2 - Resolution: 4K - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc230060090012001500Min: 1789 / Avg: 1861.36 / Max: 1969Min: 1789 / Avg: 1838.67 / Max: 1878

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.

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 4.3Binary: Pathtracer ISPC - Model: Asian Dragon ObjLinux 6.8Linux 6.9-rc2918273645SE +/- 0.94, N = 15SE +/- 0.88, N = 1536.4437.43MIN: 22.56 / MAX: 49.05MIN: 27.23 / MAX: 49.01
OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 4.3Binary: Pathtracer ISPC - Model: Asian Dragon ObjLinux 6.8Linux 6.9-rc2816243240Min: 31.27 / Avg: 36.44 / Max: 45.43Min: 32.08 / Avg: 37.43 / Max: 44.2

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 4.3Binary: Pathtracer - Model: Asian Dragon ObjLinux 6.8Linux 6.9-rc2918273645SE +/- 0.92, N = 15SE +/- 0.60, N = 1539.8236.50MIN: 26.64 / MAX: 50.91MIN: 22.66 / MAX: 46.51
OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 4.3Binary: Pathtracer - Model: Asian Dragon ObjLinux 6.8Linux 6.9-rc2816243240Min: 33.25 / Avg: 39.82 / Max: 45.68Min: 32.19 / Avg: 36.5 / Max: 39.45

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Matrix 3D MathLinux 6.8Linux 6.9-rc27K14K21K28K35KSE +/- 669.33, N = 15SE +/- 437.41, N = 1531442.6931460.421. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Matrix 3D MathLinux 6.8Linux 6.9-rc25K10K15K20K25KMin: 29829.5 / Avg: 31442.69 / Max: 38571.65Min: 30346.56 / Avg: 31460.42 / Max: 37089.941. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: PipeLinux 6.8Linux 6.9-rc27M14M21M28M35MSE +/- 749025.19, N = 15SE +/- 2106194.07, N = 1530377557.9729814480.671. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: PipeLinux 6.8Linux 6.9-rc25M10M15M20M25MMin: 25118542.17 / Avg: 30377557.97 / Max: 35408658.56Min: 24507990.51 / Avg: 29814480.67 / Max: 58382980.491. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: CryptoLinux 6.8Linux 6.9-rc230K60K90K120K150KSE +/- 1720.03, N = 15SE +/- 2039.96, N = 15143967.24147978.331. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: CryptoLinux 6.8Linux 6.9-rc230K60K90K120K150KMin: 129124.76 / Avg: 143967.24 / Max: 162008.05Min: 142636.41 / Avg: 147978.33 / Max: 172381.361. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

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.

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: fcn-resnet101-11 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc250100150200250SE +/- 11.46, N = 12SE +/- 1.80, N = 3231.28184.791. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: fcn-resnet101-11 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc24080120160200Min: 189.13 / Avg: 231.28 / Max: 288.64Min: 182.44 / Avg: 184.79 / Max: 188.331. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: fcn-resnet101-11 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc21.21782.43563.65344.87126.089SE +/- 0.20811, N = 12SE +/- 0.05231, N = 34.436495.412591. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: fcn-resnet101-11 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc250100150200250SE +/- 11.46, N = 12SE +/- 1.80, N = 3231.28184.791. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: fcn-resnet101-11 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2246810Min: 3.46 / Avg: 4.44 / Max: 5.29Min: 5.31 / Avg: 5.41 / Max: 5.481. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

oneDNN

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: Recurrent Neural Network Inference - Engine: CPULinux 6.8Linux 6.9-rc2160320480640800SE +/- 7.46, N = 3SE +/- 6.16, N = 9748.50753.36MIN: 677.68MIN: 652.831. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: Recurrent Neural Network Inference - Engine: CPULinux 6.8Linux 6.9-rc2130260390520650Min: 734.18 / Avg: 748.49 / Max: 759.31Min: 733.88 / Avg: 753.36 / Max: 795.11. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: MMAPLinux 6.8Linux 6.9-rc214002800420056007000SE +/- 70.88, N = 14SE +/- 66.09, N = 156169.366427.331. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: MMAPLinux 6.8Linux 6.9-rc211002200330044005500Min: 6033.64 / Avg: 6169.36 / Max: 7082.81Min: 6237.85 / Avg: 6427.33 / Max: 7162.331. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Glibc C String FunctionsLinux 6.8Linux 6.9-rc220M40M60M80M100MSE +/- 2138024.53, N = 13SE +/- 1504028.14, N = 1580643523.2485035507.551. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Glibc C String FunctionsLinux 6.8Linux 6.9-rc215M30M45M60M75MMin: 56423773.21 / Avg: 80643523.24 / Max: 86903860.46Min: 78762589.99 / Avg: 85035507.55 / Max: 101856587.951. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

Speedb

Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Read Random Write RandomLinux 6.8Linux 6.9-rc2300K600K900K1200K1500KSE +/- 8674.90, N = 11SE +/- 13609.90, N = 3118175811755621. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Read Random Write RandomLinux 6.8Linux 6.9-rc2200K400K600K800K1000KMin: 1133258 / Avg: 1181758.45 / Max: 1238775Min: 1150128 / Avg: 1175561.67 / Max: 11966771. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Matrix MathLinux 6.8Linux 6.9-rc280K160K240K320K400KSE +/- 4925.11, N = 15SE +/- 9278.38, N = 12384748.54392750.091. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Matrix MathLinux 6.8Linux 6.9-rc270K140K210K280K350KMin: 364151.44 / Avg: 384748.54 / Max: 424991.22Min: 362090.91 / Avg: 392750.09 / Max: 4714591. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Fused Multiply-AddLinux 6.8Linux 6.9-rc240M80M120M160M200MSE +/- 6187723.80, N = 12SE +/- 3824188.11, N = 15204365411.59200468629.321. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Fused Multiply-AddLinux 6.8Linux 6.9-rc240M80M120M160M200MMin: 189121670.73 / Avg: 204365411.59 / Max: 249896559.51Min: 166011504.98 / Avg: 200468629.32 / Max: 2294149111. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 2 - Resolution: 4K - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc28K16K24K32K40KSE +/- 302.37, N = 15SE +/- 483.46, N = 33531735422
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 2 - Resolution: 4K - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc26K12K18K24K30KMin: 33676 / Avg: 35316.73 / Max: 37909Min: 34540 / Avg: 35422.33 / Max: 36206

TensorFlow

OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 16 - Model: GoogLeNetLinux 6.8Linux 6.9-rc21326395265SE +/- 1.31, N = 12SE +/- 1.24, N = 1258.5058.83
OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 16 - Model: GoogLeNetLinux 6.8Linux 6.9-rc21224364860Min: 53.07 / Avg: 58.5 / Max: 69.05Min: 51.38 / Avg: 58.83 / Max: 64.46

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 3 - Resolution: 4K - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc29K18K27K36K45KSE +/- 626.71, N = 12SE +/- 252.89, N = 34065739794
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 3 - Resolution: 4K - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc27K14K21K28K35KMin: 38816 / Avg: 40657.08 / Max: 45903Min: 39453 / Avg: 39794 / Max: 40288

srsRAN Project

OpenBenchmarking.orgMbps, More Is BettersrsRAN Project 23.10.1-20240219Test: PDSCH Processor Benchmark, Throughput TotalLinux 6.8Linux 6.9-rc24K8K12K16K20KSE +/- 263.34, N = 15SE +/- 377.19, N = 1217006.816877.61. (CXX) g++ options: -march=native -mavx2 -mavx -msse4.1 -mfma -mavx512f -mavx512cd -mavx512bw -mavx512dq -O3 -fno-trapping-math -fno-math-errno -ldl
OpenBenchmarking.orgMbps, More Is BettersrsRAN Project 23.10.1-20240219Test: PDSCH Processor Benchmark, Throughput TotalLinux 6.8Linux 6.9-rc23K6K9K12K15KMin: 14046 / Avg: 17006.77 / Max: 18404Min: 12950.8 / Avg: 16877.58 / Max: 17847.51. (CXX) g++ options: -march=native -mavx2 -mavx -msse4.1 -mfma -mavx512f -mavx512cd -mavx512bw -mavx512dq -O3 -fno-trapping-math -fno-math-errno -ldl

CacheBench

This is a performance test of CacheBench, which is part of LLCbench. CacheBench is designed to test the memory and cache bandwidth performance Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterCacheBenchTest: ReadLinux 6.8Linux 6.9-rc23K6K9K12K15KSE +/- 0.52, N = 3SE +/- 0.88, N = 313344.7013341.47MIN: 13338.06 / MAX: 13346.23MIN: 13335.31 / MAX: 13343.851. (CC) gcc options: -O3 -lrt
OpenBenchmarking.orgMB/s, More Is BetterCacheBenchTest: ReadLinux 6.8Linux 6.9-rc22K4K6K8K10KMin: 13343.66 / Avg: 13344.7 / Max: 13345.3Min: 13339.77 / Avg: 13341.47 / Max: 13342.71. (CC) gcc options: -O3 -lrt

OpenBenchmarking.orgMB/s, More Is BetterCacheBenchTest: Read / Modify / WriteLinux 6.8Linux 6.9-rc220K40K60K80K100KSE +/- 2.71, N = 3SE +/- 71.07, N = 399112.7098711.08MIN: 88982.6 / MAX: 105154.55MIN: 85707.72 / MAX: 105115.441. (CC) gcc options: -O3 -lrt
OpenBenchmarking.orgMB/s, More Is BetterCacheBenchTest: Read / Modify / WriteLinux 6.8Linux 6.9-rc220K40K60K80K100KMin: 99109.26 / Avg: 99112.7 / Max: 99118.04Min: 98624.41 / Avg: 98711.08 / Max: 98851.981. (CC) gcc options: -O3 -lrt

OpenBenchmarking.orgMB/s, More Is BetterCacheBenchTest: WriteLinux 6.8Linux 6.9-rc220K40K60K80K100KSE +/- 36.95, N = 3SE +/- 65.10, N = 393036.3093862.76MIN: 54037.87 / MAX: 103869.49MIN: 54012.69 / MAX: 104121.121. (CC) gcc options: -O3 -lrt
OpenBenchmarking.orgMB/s, More Is BetterCacheBenchTest: WriteLinux 6.8Linux 6.9-rc216K32K48K64K80KMin: 92962.42 / Avg: 93036.3 / Max: 93074.36Min: 93781.39 / Avg: 93862.76 / Max: 93991.471. (CC) gcc options: -O3 -lrt

Blender

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.1Blend File: Fishy Cat - Compute: CPU-OnlyLinux 6.8Linux 6.9-rc2918273645SE +/- 0.38, N = 15SE +/- 0.44, N = 438.0939.37
OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.1Blend File: Fishy Cat - Compute: CPU-OnlyLinux 6.8Linux 6.9-rc2816243240Min: 35.37 / Avg: 38.09 / Max: 40.9Min: 38.58 / Avg: 39.37 / Max: 40.64

OSPRay

OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 3.1Benchmark: gravity_spheres_volume/dim_512/pathtracer/real_timeLinux 6.8Linux 6.9-rc2612182430SE +/- 0.41, N = 12SE +/- 0.20, N = 324.2825.17
OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 3.1Benchmark: gravity_spheres_volume/dim_512/pathtracer/real_timeLinux 6.8Linux 6.9-rc2612182430Min: 22.24 / Avg: 24.28 / Max: 26.57Min: 24.77 / Avg: 25.17 / Max: 25.44

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 2 - Resolution: 4K - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc214K28K42K56K70KSE +/- 591.53, N = 7SE +/- 885.61, N = 36351463529
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 2 - Resolution: 4K - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc211K22K33K44K55KMin: 61606 / Avg: 63513.71 / Max: 66228Min: 62027 / Avg: 63529.33 / Max: 65093

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Vector Floating PointLinux 6.8Linux 6.9-rc250K100K150K200K250KSE +/- 5616.14, N = 12SE +/- 8579.53, N = 12234669.74230503.341. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Vector Floating PointLinux 6.8Linux 6.9-rc240K80K120K160K200KMin: 173368 / Avg: 234669.74 / Max: 244620.56Min: 138822.38 / Avg: 230503.34 / Max: 253860.891. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Floating PointLinux 6.8Linux 6.9-rc26K12K18K24K30KSE +/- 632.70, N = 12SE +/- 214.66, N = 1129600.0229139.601. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Floating PointLinux 6.8Linux 6.9-rc25K10K15K20K25KMin: 23607.4 / Avg: 29600.02 / Max: 31592.19Min: 28569.69 / Avg: 29139.6 / Max: 30703.731. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: MEMFDLinux 6.8Linux 6.9-rc2150300450600750SE +/- 14.68, N = 15SE +/- 5.94, N = 8688.26688.761. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: MEMFDLinux 6.8Linux 6.9-rc2120240360480600Min: 670.63 / Avg: 688.26 / Max: 893.49Min: 678.8 / Avg: 688.76 / Max: 728.491. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

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.

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 4.3Binary: Pathtracer - Model: CrownLinux 6.8Linux 6.9-rc2816243240SE +/- 0.59, N = 15SE +/- 1.07, N = 1536.6035.08MIN: 25.71 / MAX: 54.49MIN: 22.8 / MAX: 59.06
OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 4.3Binary: Pathtracer - Model: CrownLinux 6.8Linux 6.9-rc2816243240Min: 31.09 / Avg: 36.6 / Max: 39.94Min: 28.14 / Avg: 35.08 / Max: 41.32

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 4.3Binary: Pathtracer ISPC - Model: Asian DragonLinux 6.8Linux 6.9-rc2816243240SE +/- 0.77, N = 13SE +/- 0.82, N = 1531.6032.83MIN: 20.8 / MAX: 52.83MIN: 22.56 / MAX: 48.16
OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 4.3Binary: Pathtracer ISPC - Model: Asian DragonLinux 6.8Linux 6.9-rc2714212835Min: 26.15 / Avg: 31.6 / Max: 35.39Min: 25.26 / Avg: 32.83 / Max: 36.5

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 4.3Binary: Pathtracer - Model: Asian DragonLinux 6.8Linux 6.9-rc2816243240SE +/- 0.86, N = 15SE +/- 0.54, N = 1535.4035.65MIN: 23.5 / MAX: 52.61MIN: 24.54 / MAX: 51.88
OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 4.3Binary: Pathtracer - Model: Asian DragonLinux 6.8Linux 6.9-rc2816243240Min: 26.95 / Avg: 35.4 / Max: 39.75Min: 30.66 / Avg: 35.65 / Max: 39.93

VVenC

OpenBenchmarking.orgFrames Per Second, More Is BetterVVenC 1.11Video Input: Bosphorus 1080p - Video Preset: FasterLinux 6.8Linux 6.9-rc2816243240SE +/- 0.29, N = 15SE +/- 0.36, N = 1533.9933.091. (CXX) g++ options: -O3 -flto=auto -fno-fat-lto-objects
OpenBenchmarking.orgFrames Per Second, More Is BetterVVenC 1.11Video Input: Bosphorus 1080p - Video Preset: FasterLinux 6.8Linux 6.9-rc2714212835Min: 31.35 / Avg: 33.99 / Max: 35.48Min: 29.81 / Avg: 33.09 / Max: 35.111. (CXX) g++ options: -O3 -flto=auto -fno-fat-lto-objects

Intel Open Image Denoise

Open Image Denoise is a denoising library for ray-tracing and part of the Intel oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgImages / Sec, More Is BetterIntel Open Image Denoise 2.2Run: RTLightmap.hdr.4096x4096 - Device: CPU-OnlyLinux 6.8Linux 6.9-rc20.23850.4770.71550.9541.1925SE +/- 0.01, N = 7SE +/- 0.01, N = 151.061.04
OpenBenchmarking.orgImages / Sec, More Is BetterIntel Open Image Denoise 2.2Run: RTLightmap.hdr.4096x4096 - Device: CPU-OnlyLinux 6.8Linux 6.9-rc2246810Min: 1.02 / Avg: 1.06 / Max: 1.1Min: 1 / Avg: 1.04 / Max: 1.09

JPEG-XL libjxl

OpenBenchmarking.orgMP/s, More Is BetterJPEG-XL libjxl 0.10.1Input: JPEG - Quality: 90Linux 6.8Linux 6.9-rc2816243240SE +/- 0.37, N = 3SE +/- 0.38, N = 1532.5633.891. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm
OpenBenchmarking.orgMP/s, More Is BetterJPEG-XL libjxl 0.10.1Input: JPEG - Quality: 90Linux 6.8Linux 6.9-rc2714212835Min: 31.96 / Avg: 32.56 / Max: 33.23Min: 32.05 / Avg: 33.89 / Max: 37.151. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm

Xmrig

Xmrig is an open-source cross-platform CPU/GPU miner for RandomX, KawPow, CryptoNight and AstroBWT. This test profile is setup to measure the Xmrig CPU mining performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgH/s, More Is BetterXmrig 6.21Variant: GhostRider - Hash Count: 1MLinux 6.8Linux 6.9-rc22K4K6K8K10KSE +/- 31.05, N = 3SE +/- 42.79, N = 39592.39602.71. (CXX) g++ options: -fexceptions -fno-rtti -maes -O3 -Ofast -static-libgcc -static-libstdc++ -rdynamic -lssl -lcrypto -luv -lpthread -lrt -ldl -lhwloc
OpenBenchmarking.orgH/s, More Is BetterXmrig 6.21Variant: GhostRider - Hash Count: 1MLinux 6.8Linux 6.9-rc217003400510068008500Min: 9536.1 / Avg: 9592.27 / Max: 9643.3Min: 9537.5 / Avg: 9602.7 / Max: 9683.31. (CXX) g++ options: -fexceptions -fno-rtti -maes -O3 -Ofast -static-libgcc -static-libstdc++ -rdynamic -lssl -lcrypto -luv -lpthread -lrt -ldl -lhwloc

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: IO_uringLinux 6.8Linux 6.9-rc2700K1400K2100K2800K3500KSE +/- 31595.27, N = 3SE +/- 43339.28, N = 153468957.593346567.801. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: IO_uringLinux 6.8Linux 6.9-rc2600K1200K1800K2400K3000KMin: 3411726.95 / Avg: 3468957.59 / Max: 3520773.62Min: 2951579.79 / Avg: 3346567.8 / Max: 3574079.351. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OSPRay

OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 3.1Benchmark: particle_volume/ao/real_timeLinux 6.8Linux 6.9-rc2714212835SE +/- 0.11, N = 3SE +/- 0.19, N = 329.7328.88
OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 3.1Benchmark: particle_volume/ao/real_timeLinux 6.8Linux 6.9-rc2714212835Min: 29.53 / Avg: 29.73 / Max: 29.9Min: 28.55 / Avg: 28.88 / Max: 29.19

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 3 - Resolution: 1080p - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc24K8K12K16K20KSE +/- 101.18, N = 3SE +/- 156.23, N = 61720416687
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 3 - Resolution: 1080p - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc23K6K9K12K15KMin: 17006 / Avg: 17203.67 / Max: 17340Min: 16283 / Avg: 16686.5 / Max: 17176

Timed Mesa Compilation

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Mesa Compilation 24.0Time To CompileLinux 6.8Linux 6.9-rc2510152025SE +/- 0.51, N = 15SE +/- 0.39, N = 1519.5119.89
OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Mesa Compilation 24.0Time To CompileLinux 6.8Linux 6.9-rc2510152025Min: 17.55 / Avg: 19.51 / Max: 22.26Min: 17.56 / Avg: 19.89 / Max: 21.33

TensorFlow

OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 1 - Model: GoogLeNetLinux 6.8Linux 6.9-rc2246810SE +/- 0.13, N = 12SE +/- 0.14, N = 156.055.98
OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 1 - Model: GoogLeNetLinux 6.8Linux 6.9-rc2246810Min: 5.46 / Avg: 6.05 / Max: 7.02Min: 5.32 / Avg: 5.98 / Max: 6.95

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Wide Vector MathLinux 6.8Linux 6.9-rc21.3M2.6M3.9M5.2M6.5MSE +/- 72950.92, N = 4SE +/- 55746.31, N = 155874353.635908316.031. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Wide Vector MathLinux 6.8Linux 6.9-rc21000K2000K3000K4000K5000KMin: 5734796.52 / Avg: 5874353.63 / Max: 6078708.21Min: 5527618.1 / Avg: 5908316.03 / Max: 6416763.811. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

SVT-AV1

OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-AV1 2.0Encoder Mode: Preset 4 - Input: Bosphorus 4KLinux 6.8Linux 6.9-rc21.31962.63923.95885.27846.598SE +/- 0.035, N = 3SE +/- 0.051, N = 155.8655.7601. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-AV1 2.0Encoder Mode: Preset 4 - Input: Bosphorus 4KLinux 6.8Linux 6.9-rc2246810Min: 5.8 / Avg: 5.87 / Max: 5.92Min: 5.47 / Avg: 5.76 / Max: 6.061. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq

Xmrig

Xmrig is an open-source cross-platform CPU/GPU miner for RandomX, KawPow, CryptoNight and AstroBWT. This test profile is setup to measure the Xmrig CPU mining performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgH/s, More Is BetterXmrig 6.21Variant: Monero - Hash Count: 1MLinux 6.8Linux 6.9-rc27K14K21K28K35KSE +/- 320.50, N = 13SE +/- 181.84, N = 332361.233015.51. (CXX) g++ options: -fexceptions -fno-rtti -maes -O3 -Ofast -static-libgcc -static-libstdc++ -rdynamic -lssl -lcrypto -luv -lpthread -lrt -ldl -lhwloc
OpenBenchmarking.orgH/s, More Is BetterXmrig 6.21Variant: Monero - Hash Count: 1MLinux 6.8Linux 6.9-rc26K12K18K24K30KMin: 28745.5 / Avg: 32361.16 / Max: 33220.4Min: 32652 / Avg: 33015.47 / Max: 33208.21. (CXX) g++ options: -fexceptions -fno-rtti -maes -O3 -Ofast -static-libgcc -static-libstdc++ -rdynamic -lssl -lcrypto -luv -lpthread -lrt -ldl -lhwloc

Y-Cruncher

Y-Cruncher is a multi-threaded Pi benchmark capable of computing Pi to trillions of digits. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterY-Cruncher 0.8.3Pi Digits To Calculate: 1BLinux 6.8Linux 6.9-rc23691215SE +/- 0.30, N = 15SE +/- 0.24, N = 1513.4213.34
OpenBenchmarking.orgSeconds, Fewer Is BetterY-Cruncher 0.8.3Pi Digits To Calculate: 1BLinux 6.8Linux 6.9-rc248121620Min: 12.19 / Avg: 13.42 / Max: 16.18Min: 12.47 / Avg: 13.33 / Max: 15.31

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: PthreadLinux 6.8Linux 6.9-rc28K16K24K32K40KSE +/- 49.70, N = 3SE +/- 450.35, N = 1539599.2635919.641. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: PthreadLinux 6.8Linux 6.9-rc27K14K21K28K35KMin: 39501.61 / Avg: 39599.26 / Max: 39664.14Min: 34802.3 / Avg: 35919.64 / Max: 41939.151. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 3 - Resolution: 1080p - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc22K4K6K8K10KSE +/- 70.21, N = 3SE +/- 26.21, N = 387328552
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 3 - Resolution: 1080p - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc215003000450060007500Min: 8657 / Avg: 8731.67 / Max: 8872Min: 8500 / Avg: 8552.33 / Max: 8581

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: System V Message PassingLinux 6.8Linux 6.9-rc24M8M12M16M20MSE +/- 189741.33, N = 15SE +/- 35075.39, N = 317859304.3818740350.561. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: System V Message PassingLinux 6.8Linux 6.9-rc23M6M9M12M15MMin: 17359798.67 / Avg: 17859304.38 / Max: 19573809.48Min: 18670511.29 / Avg: 18740350.56 / Max: 18780989.211. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 1 - Resolution: 1080p - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc215003000450060007500SE +/- 83.68, N = 3SE +/- 15.17, N = 369096804
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 1 - Resolution: 1080p - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc212002400360048006000Min: 6742 / Avg: 6909 / Max: 7002Min: 6788 / Avg: 6803.67 / Max: 6834

Timed FFmpeg Compilation

This test times how long it takes to build the FFmpeg multimedia library. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed FFmpeg Compilation 6.1Time To CompileLinux 6.8Linux 6.9-rc248121620SE +/- 0.39, N = 15SE +/- 0.39, N = 1517.2517.28
OpenBenchmarking.orgSeconds, Fewer Is BetterTimed FFmpeg Compilation 6.1Time To CompileLinux 6.8Linux 6.9-rc248121620Min: 16.03 / Avg: 17.24 / Max: 21.87Min: 15.9 / Avg: 17.28 / Max: 22.2

RocksDB

OpenBenchmarking.orgOp/s, More Is BetterRocksDB 9.0Test: Read Random Write RandomLinux 6.8Linux 6.9-rc2500K1000K1500K2000K2500KSE +/- 17412.72, N = 3SE +/- 22543.45, N = 5212084420685881. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.orgOp/s, More Is BetterRocksDB 9.0Test: Read Random Write RandomLinux 6.8Linux 6.9-rc2400K800K1200K1600K2000KMin: 2093440 / Avg: 2120843.67 / Max: 2153157Min: 2027056 / Avg: 2068588.2 / Max: 21517311. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 2 - Resolution: 1080p - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc2110220330440550SE +/- 5.26, N = 5SE +/- 2.33, N = 3501489
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 2 - Resolution: 1080p - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc290180270360450Min: 487 / Avg: 500.8 / Max: 516Min: 485 / Avg: 488.67 / Max: 493

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: MallocLinux 6.8Linux 6.9-rc240M80M120M160M200MSE +/- 201545.00, N = 3SE +/- 3675496.40, N = 12163551396.40161652700.091. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: MallocLinux 6.8Linux 6.9-rc230M60M90M120M150MMin: 163156317.95 / Avg: 163551396.4 / Max: 163818188.26Min: 152920426.85 / Avg: 161652700.09 / Max: 197690253.651. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

Timed Linux Kernel Compilation

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Linux Kernel Compilation 6.8Build: defconfigLinux 6.8Linux 6.9-rc2918273645SE +/- 0.36, N = 6SE +/- 0.37, N = 638.4637.70
OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Linux Kernel Compilation 6.8Build: defconfigLinux 6.8Linux 6.9-rc2816243240Min: 37.62 / Avg: 38.46 / Max: 40.02Min: 37.1 / Avg: 37.7 / Max: 39.49

OpenVINO

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Weld Porosity Detection FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc20.91131.82262.73393.64524.5565SE +/- 0.05, N = 4SE +/- 0.01, N = 34.053.89MIN: 2.67 / MAX: 52.43MIN: 2.64 / MAX: 47.091. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Weld Porosity Detection FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc2246810Min: 3.95 / Avg: 4.05 / Max: 4.17Min: 3.86 / Avg: 3.89 / Max: 3.911. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Weld Porosity Detection FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc26K12K18K24K30KSE +/- 285.82, N = 4SE +/- 155.42, N = 326732.8827790.241. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Weld Porosity Detection FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc25K10K15K20K25KMin: 25987.7 / Avg: 26732.88 / Max: 27373.74Min: 27500.36 / Avg: 27790.24 / Max: 28032.381. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

WebP2 Image Encode

This is a test of Google's libwebp2 library with the WebP2 image encode utility and using a sample 6000x4000 pixel JPEG image as the input, similar to the WebP/libwebp test profile. WebP2 is currently experimental and under heavy development as ultimately the successor to WebP. WebP2 supports 10-bit HDR, more efficienct lossy compression, improved lossless compression, animation support, and full multi-threading support compared to WebP. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMP/s, More Is BetterWebP2 Image Encode 20220823Encode Settings: Quality 95, Compression Effort 7Linux 6.8Linux 6.9-rc20.07650.1530.22950.3060.3825SE +/- 0.00, N = 3SE +/- 0.00, N = 30.340.331. (CXX) g++ options: -msse4.2 -fno-rtti -O3 -ldl
OpenBenchmarking.orgMP/s, More Is BetterWebP2 Image Encode 20220823Encode Settings: Quality 95, Compression Effort 7Linux 6.8Linux 6.9-rc212345Min: 0.34 / Avg: 0.34 / Max: 0.34Min: 0.33 / Avg: 0.33 / Max: 0.331. (CXX) g++ options: -msse4.2 -fno-rtti -O3 -ldl

SVT-AV1

OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-AV1 2.0Encoder Mode: Preset 8 - Input: Bosphorus 4KLinux 6.8Linux 6.9-rc21122334455SE +/- 1.10, N = 12SE +/- 1.18, N = 1548.9448.571. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-AV1 2.0Encoder Mode: Preset 8 - Input: Bosphorus 4KLinux 6.8Linux 6.9-rc21020304050Min: 43.12 / Avg: 48.94 / Max: 56.45Min: 41.67 / Avg: 48.57 / Max: 56.661. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq

Intel Open Image Denoise

Open Image Denoise is a denoising library for ray-tracing and part of the Intel oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgImages / Sec, More Is BetterIntel Open Image Denoise 2.2Run: RT.hdr_alb_nrm.3840x2160 - Device: CPU-OnlyLinux 6.8Linux 6.9-rc20.5131.0261.5392.0522.565SE +/- 0.02, N = 15SE +/- 0.02, N = 152.282.23
OpenBenchmarking.orgImages / Sec, More Is BetterIntel Open Image Denoise 2.2Run: RT.hdr_alb_nrm.3840x2160 - Device: CPU-OnlyLinux 6.8Linux 6.9-rc2246810Min: 2.13 / Avg: 2.28 / Max: 2.36Min: 2.11 / Avg: 2.23 / Max: 2.4

TensorFlow

OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 256 - Model: AlexNetLinux 6.8Linux 6.9-rc2160320480640800SE +/- 9.01, N = 3SE +/- 6.62, N = 7723.16711.74
OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 256 - Model: AlexNetLinux 6.8Linux 6.9-rc2130260390520650Min: 705.61 / Avg: 723.16 / Max: 735.49Min: 691.56 / Avg: 711.74 / Max: 738.71

OpenVINO

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Face Detection FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc270140210280350SE +/- 0.74, N = 3SE +/- 0.81, N = 3337.21337.78MIN: 261.24 / MAX: 615.09MIN: 253.21 / MAX: 781.061. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Face Detection FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc260120180240300Min: 336.1 / Avg: 337.21 / Max: 338.62Min: 336.92 / Avg: 337.78 / Max: 339.411. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Face Detection FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc270140210280350SE +/- 0.75, N = 3SE +/- 0.82, N = 3331.23330.731. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Face Detection FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc260120180240300Min: 329.82 / Avg: 331.23 / Max: 332.39Min: 329.1 / Avg: 330.73 / Max: 331.651. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Noise Suppression Poconet-Like FP16 - Device: CPULinux 6.8Linux 6.9-rc2612182430SE +/- 0.17, N = 3SE +/- 0.20, N = 323.6823.97MIN: 11.64 / MAX: 175.69MIN: 11.24 / MAX: 358.11. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Noise Suppression Poconet-Like FP16 - Device: CPULinux 6.8Linux 6.9-rc2612182430Min: 23.44 / Avg: 23.68 / Max: 24.01Min: 23.56 / Avg: 23.97 / Max: 24.211. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Noise Suppression Poconet-Like FP16 - Device: CPULinux 6.8Linux 6.9-rc210002000300040005000SE +/- 54.03, N = 3SE +/- 29.05, N = 34621.324600.951. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Noise Suppression Poconet-Like FP16 - Device: CPULinux 6.8Linux 6.9-rc28001600240032004000Min: 4538.42 / Avg: 4621.32 / Max: 4722.8Min: 4562.77 / Avg: 4600.95 / Max: 4657.961. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Memory CopyingLinux 6.8Linux 6.9-rc25K10K15K20K25KSE +/- 178.25, N = 3SE +/- 155.03, N = 1021085.1020605.391. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Memory CopyingLinux 6.8Linux 6.9-rc24K8K12K16K20KMin: 20788.72 / Avg: 21085.1 / Max: 21404.88Min: 19594.52 / Avg: 20605.39 / Max: 21203.311. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenVINO

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Person Vehicle Bike Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc248121620SE +/- 0.06, N = 3SE +/- 0.07, N = 317.5217.58MIN: 11.97 / MAX: 108.98MIN: 12.03 / MAX: 66.231. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Person Vehicle Bike Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc248121620Min: 17.41 / Avg: 17.52 / Max: 17.61Min: 17.45 / Avg: 17.58 / Max: 17.71. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Person Vehicle Bike Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc214002800420056007000SE +/- 18.88, N = 3SE +/- 23.77, N = 36336.786306.421. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Person Vehicle Bike Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc211002200330044005500Min: 6301.25 / Avg: 6336.78 / Max: 6365.6Min: 6270.98 / Avg: 6306.42 / Max: 6351.581. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Face Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc270140210280350SE +/- 0.84, N = 3SE +/- 1.30, N = 3307.74302.77MIN: 183.58 / MAX: 728.14MIN: 181.55 / MAX: 894.731. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Face Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc260120180240300Min: 306.2 / Avg: 307.74 / Max: 309.1Min: 300.19 / Avg: 302.77 / Max: 304.311. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Face Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc2306090120150SE +/- 0.33, N = 3SE +/- 0.55, N = 3119.92121.931. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Face Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc220406080100Min: 119.39 / Avg: 119.92 / Max: 120.53Min: 121.32 / Avg: 121.93 / Max: 123.021. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Road Segmentation ADAS FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc220406080100SE +/- 0.09, N = 3SE +/- 0.28, N = 377.3477.57MIN: 56.48 / MAX: 158.67MIN: 57.16 / MAX: 148.671. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Road Segmentation ADAS FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc21530456075Min: 77.2 / Avg: 77.34 / Max: 77.5Min: 77.07 / Avg: 77.57 / Max: 78.031. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Road Segmentation ADAS FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc230060090012001500SE +/- 1.59, N = 3SE +/- 5.16, N = 31446.421442.211. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Road Segmentation ADAS FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc230060090012001500Min: 1443.54 / Avg: 1446.42 / Max: 1449.02Min: 1433.83 / Avg: 1442.21 / Max: 1451.631. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Person Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc220406080100SE +/- 0.43, N = 3SE +/- 0.20, N = 381.9079.23MIN: 49.43 / MAX: 345.35MIN: 50.1 / MAX: 290.061. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Person Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc21632486480Min: 81.24 / Avg: 81.9 / Max: 82.7Min: 78.88 / Avg: 79.23 / Max: 79.581. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Person Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc2100200300400500SE +/- 2.33, N = 3SE +/- 1.20, N = 3451.13466.451. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Person Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc280160240320400Min: 446.81 / Avg: 451.13 / Max: 454.78Min: 464.39 / Avg: 466.45 / Max: 468.541. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Person Detection FP32 - Device: CPULinux 6.8Linux 6.9-rc220406080100SE +/- 0.57, N = 3SE +/- 0.62, N = 382.8779.00MIN: 51.69 / MAX: 529.56MIN: 49.24 / MAX: 319.951. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Person Detection FP32 - Device: CPULinux 6.8Linux 6.9-rc21632486480Min: 81.86 / Avg: 82.87 / Max: 83.85Min: 77.78 / Avg: 79 / Max: 79.831. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Person Detection FP32 - Device: CPULinux 6.8Linux 6.9-rc2100200300400500SE +/- 3.09, N = 3SE +/- 3.72, N = 3445.90467.791. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Person Detection FP32 - Device: CPULinux 6.8Linux 6.9-rc280160240320400Min: 440.63 / Avg: 445.9 / Max: 451.34Min: 462.78 / Avg: 467.79 / Max: 475.051. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Person Re-Identification Retail FP16 - Device: CPULinux 6.8Linux 6.9-rc23691215SE +/- 0.02, N = 3SE +/- 0.03, N = 39.599.65MIN: 7.62 / MAX: 27.98MIN: 7.62 / MAX: 34.181. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Person Re-Identification Retail FP16 - Device: CPULinux 6.8Linux 6.9-rc23691215Min: 9.55 / Avg: 9.59 / Max: 9.62Min: 9.61 / Avg: 9.65 / Max: 9.71. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Person Re-Identification Retail FP16 - Device: CPULinux 6.8Linux 6.9-rc22K4K6K8K10KSE +/- 20.55, N = 3SE +/- 28.95, N = 311638.5111564.781. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Person Re-Identification Retail FP16 - Device: CPULinux 6.8Linux 6.9-rc22K4K6K8K10KMin: 11605.42 / Avg: 11638.51 / Max: 11676.18Min: 11510.92 / Avg: 11564.78 / Max: 11610.131. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

Blender

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.1Blend File: BMW27 - Compute: CPU-OnlyLinux 6.8Linux 6.9-rc2510152025SE +/- 0.33, N = 14SE +/- 0.08, N = 321.8222.02
OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.1Blend File: BMW27 - Compute: CPU-OnlyLinux 6.8Linux 6.9-rc2510152025Min: 21.07 / Avg: 21.82 / Max: 26.04Min: 21.89 / Avg: 22.02 / Max: 22.16

OpenVINO

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Machine Translation EN To DE FP16 - Device: CPULinux 6.8Linux 6.9-rc21224364860SE +/- 0.47, N = 3SE +/- 0.25, N = 354.9854.83MIN: 33.9 / MAX: 423.36MIN: 34.71 / MAX: 375.041. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Machine Translation EN To DE FP16 - Device: CPULinux 6.8Linux 6.9-rc21122334455Min: 54.49 / Avg: 54.98 / Max: 55.93Min: 54.34 / Avg: 54.83 / Max: 55.191. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Machine Translation EN To DE FP16 - Device: CPULinux 6.8Linux 6.9-rc2150300450600750SE +/- 5.75, N = 3SE +/- 3.16, N = 3671.66673.461. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Machine Translation EN To DE FP16 - Device: CPULinux 6.8Linux 6.9-rc2120240360480600Min: 660.17 / Avg: 671.66 / Max: 677.85Min: 669.03 / Avg: 673.46 / Max: 679.571. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

oneDNN

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: Deconvolution Batch shapes_1d - Engine: CPULinux 6.8Linux 6.9-rc2510152025SE +/- 0.23, N = 15SE +/- 0.14, N = 318.3418.03MIN: 10.09MIN: 13.291. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: Deconvolution Batch shapes_1d - Engine: CPULinux 6.8Linux 6.9-rc2510152025Min: 17.33 / Avg: 18.34 / Max: 20.5Min: 17.79 / Avg: 18.03 / Max: 18.271. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

OpenVINO

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Vehicle Detection FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc2510152025SE +/- 0.04, N = 3SE +/- 0.08, N = 322.0622.12MIN: 15.51 / MAX: 86.25MIN: 15.73 / MAX: 71.51. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Vehicle Detection FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc2510152025Min: 21.98 / Avg: 22.06 / Max: 22.11Min: 21.96 / Avg: 22.12 / Max: 22.211. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Vehicle Detection FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc211002200330044005500SE +/- 9.46, N = 3SE +/- 18.34, N = 35067.405052.711. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Vehicle Detection FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc29001800270036004500Min: 5055.58 / Avg: 5067.4 / Max: 5086.11Min: 5032.9 / Avg: 5052.71 / Max: 5089.341. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

TensorFlow

OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 32 - Model: AlexNetLinux 6.8Linux 6.9-rc280160240320400SE +/- 2.86, N = 15SE +/- 5.16, N = 15346.10341.97
OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 32 - Model: AlexNetLinux 6.8Linux 6.9-rc260120180240300Min: 321.22 / Avg: 346.1 / Max: 359.55Min: 305.73 / Avg: 341.97 / Max: 369.36

OpenVINO

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Road Segmentation ADAS FP16 - Device: CPULinux 6.8Linux 6.9-rc2816243240SE +/- 0.35, N = 3SE +/- 0.23, N = 336.2835.14MIN: 24.78 / MAX: 205.05MIN: 25.25 / MAX: 437.271. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Road Segmentation ADAS FP16 - Device: CPULinux 6.8Linux 6.9-rc2816243240Min: 35.86 / Avg: 36.28 / Max: 36.98Min: 34.77 / Avg: 35.14 / Max: 35.551. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Road Segmentation ADAS FP16 - Device: CPULinux 6.8Linux 6.9-rc22004006008001000SE +/- 9.82, N = 3SE +/- 6.69, N = 31018.471051.301. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Road Segmentation ADAS FP16 - Device: CPULinux 6.8Linux 6.9-rc22004006008001000Min: 998.97 / Avg: 1018.47 / Max: 1030.3Min: 1039.16 / Avg: 1051.3 / Max: 1062.241. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -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.

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc21020304050SE +/- 0.23, N = 3SE +/- 0.13, N = 342.6843.151. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc2918273645Min: 42.34 / Avg: 42.68 / Max: 43.12Min: 42.91 / Avg: 43.15 / Max: 43.361. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc2612182430SE +/- 0.13, N = 3SE +/- 0.07, N = 323.4323.171. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc21020304050SE +/- 0.23, N = 3SE +/- 0.13, N = 342.6843.151. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc2510152025Min: 23.19 / Avg: 23.43 / Max: 23.62Min: 23.06 / Avg: 23.17 / Max: 23.31. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2714212835SE +/- 0.38, N = 3SE +/- 0.22, N = 330.8229.111. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2714212835Min: 30.19 / Avg: 30.82 / Max: 31.51Min: 28.8 / Avg: 29.11 / Max: 29.531. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2816243240SE +/- 0.40, N = 3SE +/- 0.25, N = 332.4534.351. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2714212835SE +/- 0.38, N = 3SE +/- 0.22, N = 330.8229.111. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: StandardLinux 6.8Linux 6.9-rc2714212835Min: 31.74 / Avg: 32.45 / Max: 33.12Min: 33.86 / Avg: 34.35 / Max: 34.721. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenVINO

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Face Detection Retail FP16 - Device: CPULinux 6.8Linux 6.9-rc23691215SE +/- 0.02, N = 3SE +/- 0.02, N = 39.929.89MIN: 7.45 / MAX: 33.82MIN: 7.63 / MAX: 36.571. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Face Detection Retail FP16 - Device: CPULinux 6.8Linux 6.9-rc23691215Min: 9.88 / Avg: 9.92 / Max: 9.94Min: 9.85 / Avg: 9.89 / Max: 9.921. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Face Detection Retail FP16 - Device: CPULinux 6.8Linux 6.9-rc22K4K6K8K10KSE +/- 23.83, N = 3SE +/- 26.72, N = 311260.9811295.921. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Face Detection Retail FP16 - Device: CPULinux 6.8Linux 6.9-rc22K4K6K8K10KMin: 11233.61 / Avg: 11260.98 / Max: 11308.46Min: 11262.68 / Avg: 11295.92 / Max: 11348.781. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

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.

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 4.3Binary: Pathtracer ISPC - Model: CrownLinux 6.8Linux 6.9-rc2918273645SE +/- 0.47, N = 4SE +/- 1.44, N = 1238.8435.73MIN: 30.82 / MAX: 52.25MIN: 20.32 / MAX: 57.51
OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 4.3Binary: Pathtracer ISPC - Model: CrownLinux 6.8Linux 6.9-rc2816243240Min: 37.73 / Avg: 38.84 / Max: 39.7Min: 24.09 / Avg: 35.73 / Max: 41.67

OpenVINO

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Weld Porosity Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc2246810SE +/- 0.08, N = 3SE +/- 0.05, N = 36.856.76MIN: 4.47 / MAX: 51.04MIN: 4.36 / MAX: 70.911. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Weld Porosity Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc23691215Min: 6.73 / Avg: 6.85 / Max: 7Min: 6.7 / Avg: 6.76 / Max: 6.861. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Weld Porosity Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc24K8K12K16K20KSE +/- 194.69, N = 3SE +/- 107.94, N = 316147.5716349.051. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Weld Porosity Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc23K6K9K12K15KMin: 15782.51 / Avg: 16147.57 / Max: 16447.41Min: 16137.98 / Avg: 16349.05 / Max: 16493.791. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Face Detection Retail FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc2246810SE +/- 0.07, N = 3SE +/- 0.00, N = 37.177.14MIN: 5.78 / MAX: 54.3MIN: 5.79 / MAX: 44.661. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Face Detection Retail FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc23691215Min: 7.09 / Avg: 7.17 / Max: 7.3Min: 7.13 / Avg: 7.14 / Max: 7.141. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Face Detection Retail FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc23K6K9K12K15KSE +/- 132.85, N = 3SE +/- 25.09, N = 315532.5415589.591. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Face Detection Retail FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc23K6K9K12K15KMin: 15289.78 / Avg: 15532.54 / Max: 15747.45Min: 15541.44 / Avg: 15589.59 / Max: 15625.911. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Vehicle Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc248121620SE +/- 0.10, N = 3SE +/- 0.14, N = 314.5813.89MIN: 9.5 / MAX: 109.83MIN: 9.12 / MAX: 90.571. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Vehicle Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc248121620Min: 14.4 / Avg: 14.58 / Max: 14.72Min: 13.72 / Avg: 13.89 / Max: 14.161. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Vehicle Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc26001200180024003000SE +/- 16.72, N = 3SE +/- 25.96, N = 32530.042656.421. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Vehicle Detection FP16 - Device: CPULinux 6.8Linux 6.9-rc25001000150020002500Min: 2506.18 / Avg: 2530.04 / Max: 2562.26Min: 2605.38 / Avg: 2656.42 / Max: 2690.171. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Handwritten English Recognition FP16 - Device: CPULinux 6.8Linux 6.9-rc2714212835SE +/- 0.03, N = 3SE +/- 0.09, N = 330.8531.01MIN: 24.82 / MAX: 82.03MIN: 24.42 / MAX: 90.861. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Handwritten English Recognition FP16 - Device: CPULinux 6.8Linux 6.9-rc2714212835Min: 30.82 / Avg: 30.85 / Max: 30.9Min: 30.82 / Avg: 31.01 / Max: 31.121. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Handwritten English Recognition FP16 - Device: CPULinux 6.8Linux 6.9-rc28001600240032004000SE +/- 3.14, N = 3SE +/- 10.89, N = 33626.413608.211. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Handwritten English Recognition FP16 - Device: CPULinux 6.8Linux 6.9-rc26001200180024003000Min: 3620.16 / Avg: 3626.41 / Max: 3629.97Min: 3595.29 / Avg: 3608.21 / Max: 3629.851. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -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.

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: bertsquad-12 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc220406080100SE +/- 0.33, N = 3SE +/- 0.44, N = 392.7594.331. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: bertsquad-12 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc220406080100Min: 92.15 / Avg: 92.75 / Max: 93.29Min: 93.71 / Avg: 94.33 / Max: 95.191. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: bertsquad-12 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc23691215SE +/- 0.04, N = 3SE +/- 0.05, N = 310.7810.601. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: bertsquad-12 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc220406080100SE +/- 0.33, N = 3SE +/- 0.44, N = 392.7594.331. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: bertsquad-12 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc23691215Min: 10.72 / Avg: 10.78 / Max: 10.85Min: 10.5 / Avg: 10.6 / Max: 10.671. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: yolov4 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc2306090120150SE +/- 1.18, N = 3SE +/- 1.32, N = 3126.38124.381. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: yolov4 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc220406080100Min: 125 / Avg: 126.38 / Max: 128.74Min: 121.92 / Avg: 124.38 / Max: 126.421. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: yolov4 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc2246810SE +/- 0.07338, N = 3SE +/- 0.08546, N = 37.913218.041421. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: yolov4 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc2306090120150SE +/- 1.18, N = 3SE +/- 1.32, N = 3126.38124.381. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: yolov4 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc23691215Min: 7.77 / Avg: 7.91 / Max: 8Min: 7.91 / Avg: 8.04 / Max: 8.21. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenVINO

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Handwritten English Recognition FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc2714212835SE +/- 0.08, N = 3SE +/- 0.07, N = 329.6130.20MIN: 25.51 / MAX: 92.13MIN: 22.08 / MAX: 107.871. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Handwritten English Recognition FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc2714212835Min: 29.47 / Avg: 29.61 / Max: 29.76Min: 30.07 / Avg: 30.2 / Max: 30.311. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Handwritten English Recognition FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc28001600240032004000SE +/- 10.67, N = 3SE +/- 8.36, N = 33777.563704.381. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Handwritten English Recognition FP16-INT8 - Device: CPULinux 6.8Linux 6.9-rc27001400210028003500Min: 3759.23 / Avg: 3777.56 / Max: 3796.18Min: 3690.91 / Avg: 3704.38 / Max: 3719.71. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -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.

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: ArcFace ResNet-100 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc21326395265SE +/- 0.81, N = 3SE +/- 0.56, N = 359.6056.951. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: ArcFace ResNet-100 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc21224364860Min: 58.08 / Avg: 59.6 / Max: 60.85Min: 55.87 / Avg: 56.95 / Max: 57.751. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: ArcFace ResNet-100 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc248121620SE +/- 0.23, N = 3SE +/- 0.17, N = 316.7817.561. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: ArcFace ResNet-100 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc21326395265SE +/- 0.81, N = 3SE +/- 0.56, N = 359.6056.951. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: ArcFace ResNet-100 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc248121620Min: 16.43 / Avg: 16.78 / Max: 17.22Min: 17.31 / Avg: 17.56 / Max: 17.91. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

Speedb

Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Random FillLinux 6.8Linux 6.9-rc270K140K210K280K350KSE +/- 2148.02, N = 3SE +/- 3492.28, N = 33264993228161. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Random FillLinux 6.8Linux 6.9-rc260K120K180K240K300KMin: 322244 / Avg: 326499.33 / Max: 329138Min: 318867 / Avg: 322816.33 / Max: 3297801. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Random Fill SyncLinux 6.8Linux 6.9-rc260K120K180K240K300KSE +/- 1490.88, N = 3SE +/- 3603.61, N = 32574462596031. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Random Fill SyncLinux 6.8Linux 6.9-rc240K80K120K160K200KMin: 255625 / Avg: 257445.67 / Max: 260401Min: 253153 / Avg: 259603 / Max: 2656131. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Update RandomLinux 6.8Linux 6.9-rc260K120K180K240K300KSE +/- 3488.80, N = 3SE +/- 3235.16, N = 32921802934141. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Update RandomLinux 6.8Linux 6.9-rc250K100K150K200K250KMin: 285224 / Avg: 292180.33 / Max: 296130Min: 286944 / Avg: 293413.67 / Max: 2967281. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

RocksDB

OpenBenchmarking.orgOp/s, More Is BetterRocksDB 9.0Test: OverwriteLinux 6.8Linux 6.9-rc280K160K240K320K400KSE +/- 881.98, N = 3SE +/- 3053.81, N = 33880103919681. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.orgOp/s, More Is BetterRocksDB 9.0Test: OverwriteLinux 6.8Linux 6.9-rc270K140K210K280K350KMin: 386260 / Avg: 388010.33 / Max: 389075Min: 388352 / Avg: 391968.33 / Max: 3980391. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

OpenBenchmarking.orgOp/s, More Is BetterRocksDB 9.0Test: Update RandomLinux 6.8Linux 6.9-rc280K160K240K320K400KSE +/- 1470.46, N = 3SE +/- 1522.72, N = 33598253615021. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.orgOp/s, More Is BetterRocksDB 9.0Test: Update RandomLinux 6.8Linux 6.9-rc260K120K180K240K300KMin: 357432 / Avg: 359825 / Max: 362502Min: 359507 / Avg: 361501.67 / Max: 3644921. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

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.

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: ResNet50 v1-12-int8 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc23691215SE +/- 0.10, N = 3SE +/- 0.14, N = 310.6110.511. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: ResNet50 v1-12-int8 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc23691215Min: 10.49 / Avg: 10.61 / Max: 10.82Min: 10.27 / Avg: 10.51 / Max: 10.741. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: ResNet50 v1-12-int8 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc220406080100SE +/- 0.92, N = 3SE +/- 1.23, N = 394.2595.141. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: ResNet50 v1-12-int8 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc23691215SE +/- 0.10, N = 3SE +/- 0.14, N = 310.6110.511. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: ResNet50 v1-12-int8 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc220406080100Min: 92.41 / Avg: 94.25 / Max: 95.31Min: 93.11 / Avg: 95.14 / Max: 97.351. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

Speedb

Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Random ReadLinux 6.8Linux 6.9-rc280M160M240M320M400MSE +/- 1002006.26, N = 3SE +/- 4146957.10, N = 33804193893758374791. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Random ReadLinux 6.8Linux 6.9-rc270M140M210M280M350MMin: 379212895 / Avg: 380419388.67 / Max: 382408395Min: 368286871 / Avg: 375837479.33 / Max: 3825848191. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

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.

OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: super-resolution-10 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc2246810SE +/- 0.04024, N = 3SE +/- 0.06370, N = 36.550116.511271. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: super-resolution-10 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc23691215Min: 6.49 / Avg: 6.55 / Max: 6.63Min: 6.41 / Avg: 6.51 / Max: 6.631. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: super-resolution-10 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc2306090120150SE +/- 0.93, N = 3SE +/- 1.50, N = 3152.64153.561. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: super-resolution-10 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc2246810SE +/- 0.04024, N = 3SE +/- 0.06370, N = 36.550116.511271. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: super-resolution-10 - Device: CPU - Executor: ParallelLinux 6.8Linux 6.9-rc2306090120150Min: 150.85 / Avg: 152.64 / Max: 154Min: 150.83 / Avg: 153.56 / Max: 155.981. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenSSL

OpenSSL is an open-source toolkit that implements SSL (Secure Sockets Layer) and TLS (Transport Layer Security) protocols. The system/openssl test profiles relies on benchmarking the system/OS-supplied openssl binary rather than the pts/openssl test profile that uses the locally-built OpenSSL for benchmarking. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgverify/s, More Is BetterOpenSSLAlgorithm: RSA4096Linux 6.8Linux 6.9-rc2300K600K900K1200K1500KSE +/- 19064.62, N = 3SE +/- 833.27, N = 31358616.41396904.01. OpenSSL 3.0.10 1 Aug 2023 (Library: OpenSSL 3.0.10 1 Aug 2023)
OpenBenchmarking.orgverify/s, More Is BetterOpenSSLAlgorithm: RSA4096Linux 6.8Linux 6.9-rc2200K400K600K800K1000KMin: 1334905.6 / Avg: 1358616.43 / Max: 1396331.6Min: 1395519.5 / Avg: 1396904 / Max: 1398399.61. OpenSSL 3.0.10 1 Aug 2023 (Library: OpenSSL 3.0.10 1 Aug 2023)

OpenBenchmarking.orgsign/s, More Is BetterOpenSSLAlgorithm: RSA4096Linux 6.8Linux 6.9-rc28K16K24K32K40KSE +/- 46.27, N = 3SE +/- 90.23, N = 338204.638147.41. OpenSSL 3.0.10 1 Aug 2023 (Library: OpenSSL 3.0.10 1 Aug 2023)
OpenBenchmarking.orgsign/s, More Is BetterOpenSSLAlgorithm: RSA4096Linux 6.8Linux 6.9-rc27K14K21K28K35KMin: 38132.2 / Avg: 38204.57 / Max: 38290.7Min: 38041.4 / Avg: 38147.43 / Max: 38326.91. OpenSSL 3.0.10 1 Aug 2023 (Library: OpenSSL 3.0.10 1 Aug 2023)

RocksDB

OpenBenchmarking.orgOp/s, More Is BetterRocksDB 9.0Test: Random ReadLinux 6.8Linux 6.9-rc280M160M240M320M400MSE +/- 1573786.50, N = 3SE +/- 1447290.34, N = 33604985183615860831. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.orgOp/s, More Is BetterRocksDB 9.0Test: Random ReadLinux 6.8Linux 6.9-rc260M120M180M240M300MMin: 357380021 / Avg: 360498518.33 / Max: 362427419Min: 358698290 / Avg: 361586083.33 / Max: 3632015471. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Context SwitchingLinux 6.8Linux 6.9-rc24M8M12M16M20MSE +/- 178479.09, N = 8SE +/- 265846.30, N = 320343491.5219390362.901. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Context SwitchingLinux 6.8Linux 6.9-rc24M8M12M16M20MMin: 19183596.73 / Avg: 20343491.52 / Max: 20714486.65Min: 18914689.1 / Avg: 19390362.9 / Max: 19833926.781. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 1 - Resolution: 1080p - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc2100200300400500SE +/- 3.71, N = 3SE +/- 2.33, N = 3466478
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 1 - Resolution: 1080p - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc280160240320400Min: 459 / Avg: 466.33 / Max: 471Min: 476 / Avg: 478.33 / Max: 483

Blender

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.1Blend File: Classroom - Compute: CPU-OnlyLinux 6.8Linux 6.9-rc21224364860SE +/- 0.10, N = 3SE +/- 0.09, N = 353.9853.82
OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.1Blend File: Classroom - Compute: CPU-OnlyLinux 6.8Linux 6.9-rc21122334455Min: 53.85 / Avg: 53.98 / Max: 54.18Min: 53.66 / Avg: 53.82 / Max: 53.97

TensorFlow

OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 16 - Model: AlexNetLinux 6.8Linux 6.9-rc250100150200250SE +/- 1.58, N = 15SE +/- 1.89, N = 15213.35210.71
OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 16 - Model: AlexNetLinux 6.8Linux 6.9-rc24080120160200Min: 200.69 / Avg: 213.35 / Max: 224.18Min: 196.43 / Avg: 210.71 / Max: 223.24

OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 64 - Model: AlexNetLinux 6.8Linux 6.9-rc2110220330440550SE +/- 6.05, N = 3SE +/- 5.17, N = 15495.72471.88
OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 64 - Model: AlexNetLinux 6.8Linux 6.9-rc290180270360450Min: 489.56 / Avg: 495.72 / Max: 507.82Min: 442.85 / Avg: 471.88 / Max: 513.73

FFmpeg

This is a benchmark of the FFmpeg multimedia framework. The FFmpeg test profile is making use of a modified version of vbench from Columbia University's Architecture and Design Lab (ARCADE) [http://arcade.cs.columbia.edu/vbench/] that is a benchmark for video-as-a-service workloads. The test profile offers the options of a range of vbench scenarios based on freely distributable video content and offers the options of using the x264 or x265 video encoders for transcoding. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFPS, More Is BetterFFmpeg 6.1Encoder: libx264 - Scenario: LiveLinux 6.8Linux 6.9-rc24080120160200SE +/- 1.29, N = 3SE +/- 0.70, N = 3181.51182.121. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
OpenBenchmarking.orgFPS, More Is BetterFFmpeg 6.1Encoder: libx264 - Scenario: LiveLinux 6.8Linux 6.9-rc2306090120150Min: 179.04 / Avg: 181.51 / Max: 183.42Min: 181.34 / Avg: 182.12 / Max: 183.521. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma

Y-Cruncher

Y-Cruncher is a multi-threaded Pi benchmark capable of computing Pi to trillions of digits. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterY-Cruncher 0.8.3Pi Digits To Calculate: 500MLinux 6.8Linux 6.9-rc2246810SE +/- 0.143, N = 12SE +/- 0.083, N = 156.4836.202
OpenBenchmarking.orgSeconds, Fewer Is BetterY-Cruncher 0.8.3Pi Digits To Calculate: 500MLinux 6.8Linux 6.9-rc23691215Min: 5.82 / Avg: 6.48 / Max: 7.55Min: 5.66 / Avg: 6.2 / Max: 6.78

OSPRay Studio

OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 1 - Resolution: 4K - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc27K14K21K28K35KSE +/- 294.16, N = 3SE +/- 454.99, N = 33441234953
OpenBenchmarking.orgms, Fewer Is BetterOSPRay Studio 1.0Camera: 1 - Resolution: 4K - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPULinux 6.8Linux 6.9-rc26K12K18K24K30KMin: 33902 / Avg: 34412.33 / Max: 34921Min: 34360 / Avg: 34952.67 / Max: 35847

WebP2 Image Encode

This is a test of Google's libwebp2 library with the WebP2 image encode utility and using a sample 6000x4000 pixel JPEG image as the input, similar to the WebP/libwebp test profile. WebP2 is currently experimental and under heavy development as ultimately the successor to WebP. WebP2 supports 10-bit HDR, more efficienct lossy compression, improved lossless compression, animation support, and full multi-threading support compared to WebP. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMP/s, More Is BetterWebP2 Image Encode 20220823Encode Settings: Quality 75, Compression Effort 7Linux 6.8Linux 6.9-rc20.1440.2880.4320.5760.72SE +/- 0.01, N = 3SE +/- 0.01, N = 40.640.611. (CXX) g++ options: -msse4.2 -fno-rtti -O3 -ldl
OpenBenchmarking.orgMP/s, More Is BetterWebP2 Image Encode 20220823Encode Settings: Quality 75, Compression Effort 7Linux 6.8Linux 6.9-rc2246810Min: 0.63 / Avg: 0.64 / Max: 0.66Min: 0.59 / Avg: 0.61 / Max: 0.621. (CXX) g++ options: -msse4.2 -fno-rtti -O3 -ldl

Hackbench

This is a benchmark of Hackbench, a test of the Linux kernel scheduler. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterHackbenchCount: 16 - Type: ThreadLinux 6.8Linux 6.9-rc248121620SE +/- 0.00, N = 3SE +/- 0.14, N = 1514.6814.301. (CC) gcc options: -lpthread
OpenBenchmarking.orgSeconds, Fewer Is BetterHackbenchCount: 16 - Type: ThreadLinux 6.8Linux 6.9-rc248121620Min: 14.68 / Avg: 14.68 / Max: 14.68Min: 13.56 / Avg: 14.3 / Max: 15.061. (CC) gcc options: -lpthread

VVenC

OpenBenchmarking.orgFrames Per Second, More Is BetterVVenC 1.11Video Input: Bosphorus 1080p - Video Preset: FastLinux 6.8Linux 6.9-rc248121620SE +/- 0.11, N = 3SE +/- 0.07, N = 315.9116.151. (CXX) g++ options: -O3 -flto=auto -fno-fat-lto-objects
OpenBenchmarking.orgFrames Per Second, More Is BetterVVenC 1.11Video Input: Bosphorus 1080p - Video Preset: FastLinux 6.8Linux 6.9-rc248121620Min: 15.74 / Avg: 15.91 / Max: 16.13Min: 16.01 / Avg: 16.15 / Max: 16.231. (CXX) g++ options: -O3 -flto=auto -fno-fat-lto-objects

JPEG-XL libjxl

OpenBenchmarking.orgMP/s, More Is BetterJPEG-XL libjxl 0.10.1Input: PNG - Quality: 90Linux 6.8Linux 6.9-rc2714212835SE +/- 0.28, N = 3SE +/- 0.18, N = 330.9730.131. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm
OpenBenchmarking.orgMP/s, More Is BetterJPEG-XL libjxl 0.10.1Input: PNG - Quality: 90Linux 6.8Linux 6.9-rc2714212835Min: 30.44 / Avg: 30.97 / Max: 31.39Min: 29.95 / Avg: 30.13 / Max: 30.481. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm

Primesieve

OpenBenchmarking.orgSeconds, Fewer Is BetterPrimesieve 12.1Length: 1e13Linux 6.8Linux 6.9-rc2918273645SE +/- 0.06, N = 3SE +/- 0.09, N = 339.6239.461. (CXX) g++ options: -O3
OpenBenchmarking.orgSeconds, Fewer Is BetterPrimesieve 12.1Length: 1e13Linux 6.8Linux 6.9-rc2816243240Min: 39.51 / Avg: 39.62 / Max: 39.73Min: 39.28 / Avg: 39.46 / Max: 39.581. (CXX) g++ options: -O3

SVT-AV1

OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-AV1 2.0Encoder Mode: Preset 12 - Input: Bosphorus 4KLinux 6.8Linux 6.9-rc2306090120150SE +/- 5.13, N = 15SE +/- 5.21, N = 15127.20126.191. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-AV1 2.0Encoder Mode: Preset 12 - Input: Bosphorus 4KLinux 6.8Linux 6.9-rc220406080100Min: 94.71 / Avg: 127.2 / Max: 150.63Min: 90.51 / Avg: 126.19 / Max: 150.651. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq

oneDNN

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: IP Shapes 1D - Engine: CPULinux 6.8Linux 6.9-rc21.01512.03023.04534.06045.0755SE +/- 0.04580, N = 3SE +/- 0.05180, N = 124.500554.51166MIN: 3.52MIN: 3.371. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: IP Shapes 1D - Engine: CPULinux 6.8Linux 6.9-rc2246810Min: 4.41 / Avg: 4.5 / Max: 4.57Min: 4 / Avg: 4.51 / Max: 4.721. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: NUMALinux 6.8Linux 6.9-rc2110220330440550SE +/- 4.27, N = 3SE +/- 5.89, N = 4482.62487.601. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: NUMALinux 6.8Linux 6.9-rc290180270360450Min: 474.29 / Avg: 482.62 / Max: 488.43Min: 474.08 / Avg: 487.6 / Max: 502.231. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

Xmrig

Xmrig is an open-source cross-platform CPU/GPU miner for RandomX, KawPow, CryptoNight and AstroBWT. This test profile is setup to measure the Xmrig CPU mining performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgH/s, More Is BetterXmrig 6.21Variant: CryptoNight-Femto UPX2 - Hash Count: 1MLinux 6.8Linux 6.9-rc27K14K21K28K35KSE +/- 110.64, N = 3SE +/- 170.37, N = 333002.232993.81. (CXX) g++ options: -fexceptions -fno-rtti -maes -O3 -Ofast -static-libgcc -static-libstdc++ -rdynamic -lssl -lcrypto -luv -lpthread -lrt -ldl -lhwloc
OpenBenchmarking.orgH/s, More Is BetterXmrig 6.21Variant: CryptoNight-Femto UPX2 - Hash Count: 1MLinux 6.8Linux 6.9-rc26K12K18K24K30KMin: 32822.4 / Avg: 33002.2 / Max: 33203.8Min: 32742.9 / Avg: 32993.8 / Max: 33318.91. (CXX) g++ options: -fexceptions -fno-rtti -maes -O3 -Ofast -static-libgcc -static-libstdc++ -rdynamic -lssl -lcrypto -luv -lpthread -lrt -ldl -lhwloc

OpenBenchmarking.orgH/s, More Is BetterXmrig 6.21Variant: CryptoNight-Heavy - Hash Count: 1MLinux 6.8Linux 6.9-rc27K14K21K28K35KSE +/- 49.86, N = 3SE +/- 160.32, N = 332856.332826.11. (CXX) g++ options: -fexceptions -fno-rtti -maes -O3 -Ofast -static-libgcc -static-libstdc++ -rdynamic -lssl -lcrypto -luv -lpthread -lrt -ldl -lhwloc
OpenBenchmarking.orgH/s, More Is BetterXmrig 6.21Variant: CryptoNight-Heavy - Hash Count: 1MLinux 6.8Linux 6.9-rc26K12K18K24K30KMin: 32770.8 / Avg: 32856.33 / Max: 32943.5Min: 32598.8 / Avg: 32826.1 / Max: 33135.61. (CXX) g++ options: -fexceptions -fno-rtti -maes -O3 -Ofast -static-libgcc -static-libstdc++ -rdynamic -lssl -lcrypto -luv -lpthread -lrt -ldl -lhwloc

SVT-AV1

OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-AV1 2.0Encoder Mode: Preset 13 - Input: Bosphorus 4KLinux 6.8Linux 6.9-rc2306090120150SE +/- 3.82, N = 12SE +/- 5.42, N = 15123.25123.101. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-AV1 2.0Encoder Mode: Preset 13 - Input: Bosphorus 4KLinux 6.8Linux 6.9-rc220406080100Min: 104.97 / Avg: 123.25 / Max: 143.14Min: 89.32 / Avg: 123.1 / Max: 147.881. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq

Xmrig

Xmrig is an open-source cross-platform CPU/GPU miner for RandomX, KawPow, CryptoNight and AstroBWT. This test profile is setup to measure the Xmrig CPU mining performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgH/s, More Is BetterXmrig 6.21Variant: KawPow - Hash Count: 1MLinux 6.8Linux 6.9-rc27K14K21K28K35KSE +/- 68.36, N = 3SE +/- 137.96, N = 332860.833072.51. (CXX) g++ options: -fexceptions -fno-rtti -maes -O3 -Ofast -static-libgcc -static-libstdc++ -rdynamic -lssl -lcrypto -luv -lpthread -lrt -ldl -lhwloc
OpenBenchmarking.orgH/s, More Is BetterXmrig 6.21Variant: KawPow - Hash Count: 1MLinux 6.8Linux 6.9-rc26K12K18K24K30KMin: 32757.9 / Avg: 32860.8 / Max: 32990.2Min: 32908.8 / Avg: 33072.5 / Max: 33346.71. (CXX) g++ options: -fexceptions -fno-rtti -maes -O3 -Ofast -static-libgcc -static-libstdc++ -rdynamic -lssl -lcrypto -luv -lpthread -lrt -ldl -lhwloc

SVT-AV1

OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-AV1 2.0Encoder Mode: Preset 4 - Input: Bosphorus 1080pLinux 6.8Linux 6.9-rc2510152025SE +/- 0.13, N = 15SE +/- 0.21, N = 318.6618.931. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-AV1 2.0Encoder Mode: Preset 4 - Input: Bosphorus 1080pLinux 6.8Linux 6.9-rc2510152025Min: 17.32 / Avg: 18.66 / Max: 19.18Min: 18.52 / Avg: 18.93 / Max: 19.161. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq

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.

OpenBenchmarking.orgNs Per Day, More Is BetterGROMACS 2024Implementation: MPI CPU - Input: water_GMX50_bareLinux 6.8Linux 6.9-rc23691215SE +/- 0.06, N = 3SE +/- 0.08, N = 312.0911.761. (CXX) g++ options: -O3 -lm
OpenBenchmarking.orgNs Per Day, More Is BetterGROMACS 2024Implementation: MPI CPU - Input: water_GMX50_bareLinux 6.8Linux 6.9-rc248121620Min: 11.98 / Avg: 12.09 / Max: 12.17Min: 11.62 / Avg: 11.76 / Max: 11.911. (CXX) g++ options: -O3 -lm

Xmrig

Xmrig is an open-source cross-platform CPU/GPU miner for RandomX, KawPow, CryptoNight and AstroBWT. This test profile is setup to measure the Xmrig CPU mining performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgH/s, More Is BetterXmrig 6.21Variant: Wownero - Hash Count: 1MLinux 6.8Linux 6.9-rc28K16K24K32K40KSE +/- 245.88, N = 3SE +/- 490.33, N = 336096.636180.71. (CXX) g++ options: -fexceptions -fno-rtti -maes -O3 -Ofast -static-libgcc -static-libstdc++ -rdynamic -lssl -lcrypto -luv -lpthread -lrt -ldl -lhwloc
OpenBenchmarking.orgH/s, More Is BetterXmrig 6.21Variant: Wownero - Hash Count: 1MLinux 6.8Linux 6.9-rc26K12K18K24K30KMin: 35747.5 / Avg: 36096.6 / Max: 36571.1Min: 35226.2 / Avg: 36180.7 / Max: 36852.81. (CXX) g++ options: -fexceptions -fno-rtti -maes -O3 -Ofast -static-libgcc -static-libstdc++ -rdynamic -lssl -lcrypto -luv -lpthread -lrt -ldl -lhwloc

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: MutexLinux 6.8Linux 6.9-rc26M12M18M24M30MSE +/- 281178.86, N = 3SE +/- 294297.65, N = 328490133.2828934422.161. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: MutexLinux 6.8Linux 6.9-rc25M10M15M20M25MMin: 27968463.5 / Avg: 28490133.28 / Max: 28932847.6Min: 28349730.23 / Avg: 28934422.16 / Max: 29285375.461. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Socket ActivityLinux 6.8Linux 6.9-rc210K20K30K40K50KSE +/- 79.93, N = 3SE +/- 21.40, N = 347927.7546123.321. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Socket ActivityLinux 6.8Linux 6.9-rc28K16K24K32K40KMin: 47826.25 / Avg: 47927.75 / Max: 48085.46Min: 46086.1 / Avg: 46123.32 / Max: 46160.231. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: ZlibLinux 6.8Linux 6.9-rc22K4K6K8K10KSE +/- 25.13, N = 3SE +/- 124.93, N = 39670.379446.071. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: ZlibLinux 6.8Linux 6.9-rc22K4K6K8K10KMin: 9621.44 / Avg: 9670.37 / Max: 9704.78Min: 9196.27 / Avg: 9446.07 / Max: 9575.241. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Mixed SchedulerLinux 6.8Linux 6.9-rc215K30K45K60K75KSE +/- 153.71, N = 3SE +/- 281.29, N = 370407.9670036.201. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Mixed SchedulerLinux 6.8Linux 6.9-rc212K24K36K48K60KMin: 70188.63 / Avg: 70407.96 / Max: 70704.18Min: 69610.61 / Avg: 70036.2 / Max: 70567.641. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Glibc Qsort Data SortingLinux 6.8Linux 6.9-rc2400800120016002000SE +/- 4.40, N = 3SE +/- 4.16, N = 31866.581858.701. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Glibc Qsort Data SortingLinux 6.8Linux 6.9-rc230060090012001500Min: 1858.16 / Avg: 1866.58 / Max: 1872.99Min: 1851.48 / Avg: 1858.7 / Max: 1865.881. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: SemaphoresLinux 6.8Linux 6.9-rc230M60M90M120M150MSE +/- 1498916.25, N = 3SE +/- 1330349.25, N = 3154589461.19150005776.471. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: SemaphoresLinux 6.8Linux 6.9-rc230M60M90M120M150MMin: 152969984.38 / Avg: 154589461.19 / Max: 157583969.81Min: 147805489.97 / Avg: 150005776.47 / Max: 152401505.251. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: CPU StressLinux 6.8Linux 6.9-rc240K80K120K160K200KSE +/- 569.84, N = 3SE +/- 335.84, N = 3202073.08200671.701. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: CPU StressLinux 6.8Linux 6.9-rc240K80K120K160K200KMin: 201472.61 / Avg: 202073.08 / Max: 203212.21Min: 200186.6 / Avg: 200671.7 / Max: 201316.581. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: SENDFILELinux 6.8Linux 6.9-rc2400K800K1200K1600K2000KSE +/- 609.53, N = 3SE +/- 5873.31, N = 31798203.121900682.081. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: SENDFILELinux 6.8Linux 6.9-rc2300K600K900K1200K1500KMin: 1796994 / Avg: 1798203.12 / Max: 1798942.24Min: 1891226.4 / Avg: 1900682.08 / Max: 1911445.641. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: PollLinux 6.8Linux 6.9-rc22M4M6M8M10MSE +/- 37074.52, N = 3SE +/- 30731.99, N = 310389186.0210232390.411. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: PollLinux 6.8Linux 6.9-rc22M4M6M8M10MMin: 10315486.51 / Avg: 10389186.02 / Max: 10433096.08Min: 10172379.63 / Avg: 10232390.41 / Max: 10273902.141. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: HashLinux 6.8Linux 6.9-rc24M8M12M16M20MSE +/- 82496.53, N = 3SE +/- 192774.29, N = 317071694.5217199190.091. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: HashLinux 6.8Linux 6.9-rc23M6M9M12M15MMin: 16923497.51 / Avg: 17071694.52 / Max: 17208604.39Min: 16837658.17 / Avg: 17199190.09 / Max: 17495959.441. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: AVL TreeLinux 6.8Linux 6.9-rc22004006008001000SE +/- 6.53, N = 3SE +/- 2.13, N = 3810.19808.991. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: AVL TreeLinux 6.8Linux 6.9-rc2140280420560700Min: 803.64 / Avg: 810.19 / Max: 823.25Min: 806 / Avg: 808.99 / Max: 813.11. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Vector ShuffleLinux 6.8Linux 6.9-rc2120K240K360K480K600KSE +/- 1416.85, N = 3SE +/- 1897.90, N = 3576905.92569593.001. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz
OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Vector ShuffleLinux 6.8Linux 6.9-rc2100K200K300K400K500KMin: 575427.57 / Avg: 576905.92 / Max: 579738.72Min: 566389.15 / Avg: 569593 / Max: 572957.771. (CXX) g++ options: -lm -latomic -lc -lcrypt -ldl -lEGL -lGLESv2 -ljpeg -lpthread -lrt -lsctp -lz

Hackbench

This is a benchmark of Hackbench, a test of the Linux kernel scheduler. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterHackbenchCount: 32 - Type: ProcessLinux 6.8Linux 6.9-rc2714212835SE +/- 0.27, N = 3SE +/- 0.39, N = 324.7527.991. (CC) gcc options: -lpthread
OpenBenchmarking.orgSeconds, Fewer Is BetterHackbenchCount: 32 - Type: ProcessLinux 6.8Linux 6.9-rc2612182430Min: 24.23 / Avg: 24.75 / Max: 25.11Min: 27.3 / Avg: 27.99 / Max: 28.641. (CC) gcc options: -lpthread

LZ4 Compression

This test measures the time needed to compress/decompress a sample file (silesia archive) using LZ4 compression. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 9 - Decompression SpeedLinux 6.8Linux 6.9-rc28001600240032004000SE +/- 16.01, N = 3SE +/- 57.81, N = 33537.03492.01. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 9 - Decompression SpeedLinux 6.8Linux 6.9-rc26001200180024003000Min: 3506.3 / Avg: 3537.03 / Max: 3560.2Min: 3405.5 / Avg: 3492 / Max: 3601.71. (CC) gcc options: -O3

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 9 - Compression SpeedLinux 6.8Linux 6.9-rc2816243240SE +/- 0.08, N = 3SE +/- 0.07, N = 332.6132.661. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 9 - Compression SpeedLinux 6.8Linux 6.9-rc2714212835Min: 32.51 / Avg: 32.61 / Max: 32.77Min: 32.54 / Avg: 32.66 / Max: 32.771. (CC) gcc options: -O3

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 3 - Decompression SpeedLinux 6.8Linux 6.9-rc27001400210028003500SE +/- 33.92, N = 3SE +/- 34.17, N = 33364.83349.61. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 3 - Decompression SpeedLinux 6.8Linux 6.9-rc26001200180024003000Min: 3298.2 / Avg: 3364.8 / Max: 3409.3Min: 3296.1 / Avg: 3349.63 / Max: 3413.21. (CC) gcc options: -O3

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 3 - Compression SpeedLinux 6.8Linux 6.9-rc220406080100SE +/- 0.06, N = 3SE +/- 0.06, N = 398.4398.411. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 3 - Compression SpeedLinux 6.8Linux 6.9-rc220406080100Min: 98.31 / Avg: 98.43 / Max: 98.52Min: 98.29 / Avg: 98.41 / Max: 98.491. (CC) gcc options: -O3

JPEG-XL Decoding libjxl

OpenBenchmarking.orgMP/s, More Is BetterJPEG-XL Decoding libjxl 0.10.1CPU Threads: AllLinux 6.8Linux 6.9-rc280160240320400SE +/- 4.23, N = 3SE +/- 2.26, N = 3386.23381.66
OpenBenchmarking.orgMP/s, More Is BetterJPEG-XL Decoding libjxl 0.10.1CPU Threads: AllLinux 6.8Linux 6.9-rc270140210280350Min: 377.96 / Avg: 386.23 / Max: 391.88Min: 378.26 / Avg: 381.66 / Max: 385.94

LZ4 Compression

This test measures the time needed to compress/decompress a sample file (silesia archive) using LZ4 compression. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 1 - Decompression SpeedLinux 6.8Linux 6.9-rc28001600240032004000SE +/- 44.31, N = 3SE +/- 18.57, N = 33601.03628.01. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 1 - Decompression SpeedLinux 6.8Linux 6.9-rc26001200180024003000Min: 3512.4 / Avg: 3600.97 / Max: 3648.1Min: 3590.9 / Avg: 3627.97 / Max: 3648.51. (CC) gcc options: -O3

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 1 - Compression SpeedLinux 6.8Linux 6.9-rc2130260390520650SE +/- 1.15, N = 3SE +/- 0.16, N = 3602.69602.541. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 1 - Compression SpeedLinux 6.8Linux 6.9-rc2110220330440550Min: 600.42 / Avg: 602.69 / Max: 604.13Min: 602.26 / Avg: 602.54 / Max: 602.831. (CC) gcc options: -O3

WebP2 Image Encode

This is a test of Google's libwebp2 library with the WebP2 image encode utility and using a sample 6000x4000 pixel JPEG image as the input, similar to the WebP/libwebp test profile. WebP2 is currently experimental and under heavy development as ultimately the successor to WebP. WebP2 supports 10-bit HDR, more efficienct lossy compression, improved lossless compression, animation support, and full multi-threading support compared to WebP. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMP/s, More Is BetterWebP2 Image Encode 20220823Encode Settings: DefaultLinux 6.8Linux 6.9-rc21.32752.6553.98255.316.6375SE +/- 0.11, N = 15SE +/- 0.09, N = 155.905.651. (CXX) g++ options: -msse4.2 -fno-rtti -O3 -ldl
OpenBenchmarking.orgMP/s, More Is BetterWebP2 Image Encode 20220823Encode Settings: DefaultLinux 6.8Linux 6.9-rc2246810Min: 4.96 / Avg: 5.9 / Max: 6.43Min: 4.94 / Avg: 5.65 / Max: 6.131. (CXX) g++ options: -msse4.2 -fno-rtti -O3 -ldl

TensorFlow

OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 1 - Model: AlexNetLinux 6.8Linux 6.9-rc248121620SE +/- 0.25, N = 12SE +/- 0.25, N = 317.2318.28
OpenBenchmarking.orgimages/sec, More Is BetterTensorFlow 2.16.1Device: CPU - Batch Size: 1 - Model: AlexNetLinux 6.8Linux 6.9-rc2510152025Min: 16.05 / Avg: 17.23 / Max: 18.33Min: 17.81 / Avg: 18.28 / Max: 18.63

JPEG-XL libjxl

OpenBenchmarking.orgMP/s, More Is BetterJPEG-XL libjxl 0.10.1Input: PNG - Quality: 100Linux 6.8Linux 6.9-rc2612182430SE +/- 0.15, N = 3SE +/- 0.21, N = 325.6725.611. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm
OpenBenchmarking.orgMP/s, More Is BetterJPEG-XL libjxl 0.10.1Input: PNG - Quality: 100Linux 6.8Linux 6.9-rc2612182430Min: 25.38 / Avg: 25.67 / Max: 25.89Min: 25.31 / Avg: 25.61 / Max: 26.021. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm

OpenBenchmarking.orgMP/s, More Is BetterJPEG-XL libjxl 0.10.1Input: JPEG - Quality: 100Linux 6.8Linux 6.9-rc2612182430SE +/- 0.11, N = 3SE +/- 0.08, N = 326.0025.961. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm
OpenBenchmarking.orgMP/s, More Is BetterJPEG-XL libjxl 0.10.1Input: JPEG - Quality: 100Linux 6.8Linux 6.9-rc2612182430Min: 25.89 / Avg: 26 / Max: 26.22Min: 25.8 / Avg: 25.96 / Max: 26.071. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm

oneDNN

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: Convolution Batch Shapes Auto - Engine: CPULinux 6.8Linux 6.9-rc23691215SE +/- 0.03328, N = 3SE +/- 0.09863, N = 159.006218.81043MIN: 8.12MIN: 6.251. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: Convolution Batch Shapes Auto - Engine: CPULinux 6.8Linux 6.9-rc23691215Min: 8.97 / Avg: 9.01 / Max: 9.07Min: 8.1 / Avg: 8.81 / Max: 9.431. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

Intel Open Image Denoise

Open Image Denoise is a denoising library for ray-tracing and part of the Intel oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgImages / Sec, More Is BetterIntel Open Image Denoise 2.2Run: RT.ldr_alb_nrm.3840x2160 - Device: CPU-OnlyLinux 6.8Linux 6.9-rc20.5221.0441.5662.0882.61SE +/- 0.03, N = 3SE +/- 0.03, N = 32.322.22
OpenBenchmarking.orgImages / Sec, More Is BetterIntel Open Image Denoise 2.2Run: RT.ldr_alb_nrm.3840x2160 - Device: CPU-OnlyLinux 6.8Linux 6.9-rc2246810Min: 2.29 / Avg: 2.32 / Max: 2.38Min: 2.18 / Avg: 2.22 / Max: 2.27

Hackbench

This is a benchmark of Hackbench, a test of the Linux kernel scheduler. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterHackbenchCount: 16 - Type: ProcessLinux 6.8Linux 6.9-rc23691215SE +/- 0.11, N = 3SE +/- 0.19, N = 313.1113.361. (CC) gcc options: -lpthread
OpenBenchmarking.orgSeconds, Fewer Is BetterHackbenchCount: 16 - Type: ProcessLinux 6.8Linux 6.9-rc248121620Min: 12.98 / Avg: 13.11 / Max: 13.32Min: 13.04 / Avg: 13.36 / Max: 13.681. (CC) gcc options: -lpthread

WavPack Audio Encoding

OpenBenchmarking.orgSeconds, Fewer Is BetterWavPack Audio Encoding 5.7WAV To WavPackLinux 6.8Linux 6.9-rc2246810SE +/- 0.023, N = 5SE +/- 0.019, N = 56.3496.340
OpenBenchmarking.orgSeconds, Fewer Is BetterWavPack Audio Encoding 5.7WAV To WavPackLinux 6.8Linux 6.9-rc23691215Min: 6.3 / Avg: 6.35 / Max: 6.42Min: 6.3 / Avg: 6.34 / Max: 6.41

oneDNN

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: IP Shapes 3D - Engine: CPULinux 6.8Linux 6.9-rc21.08962.17923.26884.35845.448SE +/- 0.02352, N = 3SE +/- 0.02357, N = 34.798534.84245MIN: 3.98MIN: 4.011. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: IP Shapes 3D - Engine: CPULinux 6.8Linux 6.9-rc2246810Min: 4.75 / Avg: 4.8 / Max: 4.82Min: 4.8 / Avg: 4.84 / Max: 4.871. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

Google Draco

OpenBenchmarking.orgms, Fewer Is BetterGoogle Draco 1.5.6Model: Church FacadeLinux 6.8Linux 6.9-rc217003400510068008500SE +/- 0.88, N = 3SE +/- 9.56, N = 3774777621. (CXX) g++ options: -O3
OpenBenchmarking.orgms, Fewer Is BetterGoogle Draco 1.5.6Model: Church FacadeLinux 6.8Linux 6.9-rc213002600390052006500Min: 7745 / Avg: 7746.67 / Max: 7748Min: 7745 / Avg: 7762.33 / Max: 77781. (CXX) g++ options: -O3

WebP2 Image Encode

This is a test of Google's libwebp2 library with the WebP2 image encode utility and using a sample 6000x4000 pixel JPEG image as the input, similar to the WebP/libwebp test profile. WebP2 is currently experimental and under heavy development as ultimately the successor to WebP. WebP2 supports 10-bit HDR, more efficienct lossy compression, improved lossless compression, animation support, and full multi-threading support compared to WebP. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMP/s, More Is BetterWebP2 Image Encode 20220823Encode Settings: Quality 100, Compression Effort 5Linux 6.8Linux 6.9-rc2246810SE +/- 0.11, N = 15SE +/- 0.07, N = 38.358.451. (CXX) g++ options: -msse4.2 -fno-rtti -O3 -ldl
OpenBenchmarking.orgMP/s, More Is BetterWebP2 Image Encode 20220823Encode Settings: Quality 100, Compression Effort 5Linux 6.8Linux 6.9-rc23691215Min: 7.59 / Avg: 8.35 / Max: 8.93Min: 8.32 / Avg: 8.45 / Max: 8.541. (CXX) g++ options: -msse4.2 -fno-rtti -O3 -ldl

oneDNN

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: Deconvolution Batch shapes_3d - Engine: CPULinux 6.8Linux 6.9-rc2246810SE +/- 0.10591, N = 15SE +/- 0.04472, N = 36.165765.78202MIN: 3.75MIN: 3.811. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.4Harness: Deconvolution Batch shapes_3d - Engine: CPULinux 6.8Linux 6.9-rc2246810Min: 5.49 / Avg: 6.17 / Max: 6.77Min: 5.71 / Avg: 5.78 / Max: 5.861. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

Primesieve

OpenBenchmarking.orgSeconds, Fewer Is BetterPrimesieve 12.1Length: 1e12Linux 6.8Linux 6.9-rc20.75511.51022.26533.02043.7755SE +/- 0.024, N = 11SE +/- 0.044, N = 33.3563.2991. (CXX) g++ options: -O3
OpenBenchmarking.orgSeconds, Fewer Is BetterPrimesieve 12.1Length: 1e12Linux 6.8Linux 6.9-rc2246810Min: 3.21 / Avg: 3.36 / Max: 3.48Min: 3.23 / Avg: 3.3 / Max: 3.381. (CXX) g++ options: -O3

Google Draco

OpenBenchmarking.orgms, Fewer Is BetterGoogle Draco 1.5.6Model: LionLinux 6.8Linux 6.9-rc213002600390052006500SE +/- 40.84, N = 3SE +/- 29.28, N = 3612960541. (CXX) g++ options: -O3
OpenBenchmarking.orgms, Fewer Is BetterGoogle Draco 1.5.6Model: LionLinux 6.8Linux 6.9-rc211002200330044005500Min: 6087 / Avg: 6129.33 / Max: 6211Min: 5996 / Avg: 6054 / Max: 60901. (CXX) g++ options: -O3

254 Results Shown

MariaDB mariadb-slap:
  1024
  512
Quicksilver
Stockfish
Llama.cpp
FFmpeg
Blender
Speedb
FFmpeg
BRL-CAD
OSPRay
TensorFlow
VVenC
OpenSSL
TensorFlow
OSPRay Studio
TensorFlow:
  CPU - 64 - ResNet-50
  CPU - 256 - GoogLeNet
WebP2 Image Encode
Llama.cpp
OSPRay Studio
FFmpeg
TensorFlow
Llamafile
Timed Linux Kernel Compilation
oneDNN
OpenVINO:
  Age Gender Recognition Retail 0013 FP16-INT8 - CPU:
    ms
    FPS
VVenC
FFmpeg
OSPRay Studio
ONNX Runtime:
  yolov4 - CPU - Standard:
    Inference Time Cost (ms)
    Inferences Per Second
  fcn-resnet101-11 - CPU - Parallel:
    Inference Time Cost (ms)
    Inferences Per Second
  bertsquad-12 - CPU - Standard:
    Inference Time Cost (ms)
    Inferences Per Second
  ArcFace ResNet-100 - CPU - Standard:
    Inference Time Cost (ms)
    Inferences Per Second
OSPRay Studio
ONNX Runtime:
  ResNet50 v1-12-int8 - CPU - Standard:
    Inference Time Cost (ms)
    Inferences Per Second
  T5 Encoder - CPU - Parallel:
    Inference Time Cost (ms)
    Inferences Per Second
OSPRay Studio
Blender
JPEG-XL libjxl
OSPRay Studio
Quicksilver
Chaos Group V-RAY
Blender
TensorFlow
FFmpeg
OSPRay Studio
FFmpeg:
  libx264 - Platform
  libx264 - Video On Demand
OSPRay Studio
Llamafile
OpenVINO:
  Age Gender Recognition Retail 0013 FP16 - CPU:
    ms
    FPS
ONNX Runtime:
  GPT-2 - CPU - Parallel:
    Inference Time Cost (ms)
    Inferences Per Second
  T5 Encoder - CPU - Standard:
    Inference Time Cost (ms)
    Inferences Per Second
TensorFlow
ONNX Runtime:
  CaffeNet 12-int8 - CPU - Parallel:
    Inference Time Cost (ms)
    Inferences Per Second
  CaffeNet 12-int8 - CPU - Standard:
    Inference Time Cost (ms)
    Inferences Per Second
  super-resolution-10 - CPU - Standard:
    Inference Time Cost (ms)
    Inferences Per Second
OpenSSL
JPEG-XL libjxl
TensorFlow
OSPRay:
  gravity_spheres_volume/dim_512/scivis/real_time
  gravity_spheres_volume/dim_512/ao/real_time
  particle_volume/scivis/real_time
OSPRay Studio
Embree:
  Pathtracer ISPC - Asian Dragon Obj
  Pathtracer - Asian Dragon Obj
Stress-NG:
  Matrix 3D Math
  Pipe
  Crypto
ONNX Runtime:
  fcn-resnet101-11 - CPU - Standard:
    Inference Time Cost (ms)
    Inferences Per Second
oneDNN
Stress-NG:
  MMAP
  Glibc C String Functions
Speedb
Stress-NG:
  Matrix Math
  Fused Multiply-Add
OSPRay Studio
TensorFlow
OSPRay Studio
srsRAN Project
CacheBench:
  Read
  Read / Modify / Write
  Write
Blender
OSPRay
OSPRay Studio
Stress-NG:
  Vector Floating Point
  Floating Point
  MEMFD
Embree:
  Pathtracer - Crown
  Pathtracer ISPC - Asian Dragon
  Pathtracer - Asian Dragon
VVenC
Intel Open Image Denoise
JPEG-XL libjxl
Xmrig
Stress-NG
OSPRay
OSPRay Studio
Timed Mesa Compilation
TensorFlow
Stress-NG
SVT-AV1
Xmrig
Y-Cruncher
Stress-NG
OSPRay Studio
Stress-NG
OSPRay Studio
Timed FFmpeg Compilation
RocksDB
OSPRay Studio
Stress-NG
Timed Linux Kernel Compilation
OpenVINO:
  Weld Porosity Detection FP16-INT8 - CPU:
    ms
    FPS
WebP2 Image Encode
SVT-AV1
Intel Open Image Denoise
TensorFlow
OpenVINO:
  Face Detection FP16-INT8 - CPU:
    ms
    FPS
  Noise Suppression Poconet-Like FP16 - CPU:
    ms
    FPS
Stress-NG
OpenVINO:
  Person Vehicle Bike Detection FP16 - CPU:
    ms
    FPS
  Face Detection FP16 - CPU:
    ms
    FPS
  Road Segmentation ADAS FP16-INT8 - CPU:
    ms
    FPS
  Person Detection FP16 - CPU:
    ms
    FPS
  Person Detection FP32 - CPU:
    ms
    FPS
  Person Re-Identification Retail FP16 - CPU:
    ms
    FPS
Blender
OpenVINO:
  Machine Translation EN To DE FP16 - CPU:
    ms
    FPS
oneDNN
OpenVINO:
  Vehicle Detection FP16-INT8 - CPU:
    ms
    FPS
TensorFlow
OpenVINO:
  Road Segmentation ADAS FP16 - CPU:
    ms
    FPS
ONNX Runtime:
  Faster R-CNN R-50-FPN-int8 - CPU - Parallel:
    Inference Time Cost (ms)
    Inferences Per Second
  Faster R-CNN R-50-FPN-int8 - CPU - Standard:
    Inference Time Cost (ms)
    Inferences Per Second
OpenVINO:
  Face Detection Retail FP16 - CPU:
    ms
    FPS
Embree
OpenVINO:
  Weld Porosity Detection FP16 - CPU:
    ms
    FPS
  Face Detection Retail FP16-INT8 - CPU:
    ms
    FPS
  Vehicle Detection FP16 - CPU:
    ms
    FPS
  Handwritten English Recognition FP16 - CPU:
    ms
    FPS
ONNX Runtime:
  bertsquad-12 - CPU - Parallel:
    Inference Time Cost (ms)
    Inferences Per Second
  yolov4 - CPU - Parallel:
    Inference Time Cost (ms)
    Inferences Per Second
OpenVINO:
  Handwritten English Recognition FP16-INT8 - CPU:
    ms
    FPS
ONNX Runtime:
  ArcFace ResNet-100 - CPU - Parallel:
    Inference Time Cost (ms)
    Inferences Per Second
Speedb:
  Rand Fill
  Rand Fill Sync
  Update Rand
RocksDB:
  Overwrite
  Update Rand
ONNX Runtime:
  ResNet50 v1-12-int8 - CPU - Parallel:
    Inference Time Cost (ms)
    Inferences Per Second
Speedb
ONNX Runtime:
  super-resolution-10 - CPU - Parallel:
    Inference Time Cost (ms)
    Inferences Per Second
OpenSSL:
  RSA4096:
    verify/s
    sign/s
RocksDB
Stress-NG
OSPRay Studio
Blender
TensorFlow:
  CPU - 16 - AlexNet
  CPU - 64 - AlexNet
FFmpeg
Y-Cruncher
OSPRay Studio
WebP2 Image Encode
Hackbench
VVenC
JPEG-XL libjxl
Primesieve
SVT-AV1
oneDNN
Stress-NG
Xmrig:
  CryptoNight-Femto UPX2 - 1M
  CryptoNight-Heavy - 1M
SVT-AV1
Xmrig
SVT-AV1
GROMACS
Xmrig
Stress-NG:
  Mutex
  Socket Activity
  Zlib
  Mixed Scheduler
  Glibc Qsort Data Sorting
  Semaphores
  CPU Stress
  SENDFILE
  Poll
  Hash
  AVL Tree
  Vector Shuffle
Hackbench
LZ4 Compression:
  9 - Decompression Speed
  9 - Compression Speed
  3 - Decompression Speed
  3 - Compression Speed
JPEG-XL Decoding libjxl
LZ4 Compression:
  1 - Decompression Speed
  1 - Compression Speed
WebP2 Image Encode
TensorFlow
JPEG-XL libjxl:
  PNG - 100
  JPEG - 100
oneDNN
Intel Open Image Denoise
Hackbench
WavPack Audio Encoding
oneDNN
Google Draco
WebP2 Image Encode
oneDNN
Primesieve
Google Draco

Linux 6.8

Processor: 2 x Intel Xeon Max 9480 @ 3.50GHz (112 Cores / 224 Threads), Motherboard: Supermicro SYS-221H-TNR X13DEM v1.10 (1.3 BIOS), Chipset: Intel Device 1bce, Memory: 512GB, Disk: 7682GB INTEL SSDPF2KX076TZ, Graphics: ASPEED, Network: 2 x Broadcom BCM57508 NetXtreme-E 10Gb/25Gb/40Gb/50Gb/100Gb/200Gb

OS: Ubuntu 23.10, Kernel: 6.8.0-060800-generic (x86_64), Desktop: GNOME Shell 45.0, Display Server: X Server 1.21.1.7, Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1024x768

Kernel Notes: Transparent Huge Pages: madvise
Compiler Notes: --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,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-defaulted --enable-offload-targets=nvptx-none=/build/gcc-13-XYspKM/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-XYspKM/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-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
Processor Notes: Scaling Governor: intel_cpufreq performance - CPU Microcode: 0x2c000290
Python Notes: Python 3.11.6
Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + 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 RSB filling PBRSB-eIBRS: SW sequence + srbds: Not affected + tsx_async_abort: Not affected

Testing initiated at 29 March 2024 19:39 by user phoronix.

Linux 6.9-rc2

Processor: 2 x Intel Xeon Max 9480 @ 3.50GHz (112 Cores / 224 Threads), Motherboard: Supermicro SYS-221H-TNR X13DEM v1.10 (1.3 BIOS), Chipset: Intel Device 1bce, Memory: 512GB, Disk: 7682GB INTEL SSDPF2KX076TZ, Graphics: ASPEED, Network: 2 x Broadcom BCM57508 NetXtreme-E 10Gb/25Gb/40Gb/50Gb/100Gb/200Gb

OS: Ubuntu 23.10, Kernel: 6.9.0-060900rc2-generic (x86_64), Desktop: GNOME Shell 45.0, Display Server: X Server 1.21.1.7, Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1024x768

Kernel Notes: Transparent Huge Pages: madvise
Compiler Notes: --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,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-defaulted --enable-offload-targets=nvptx-none=/build/gcc-13-XYspKM/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-XYspKM/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-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
Processor Notes: Scaling Governor: intel_cpufreq performance - CPU Microcode: 0x2c000290
Python Notes: Python 3.11.6
Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + reg_file_data_sampling: Not affected + retbleed: Not affected + spec_rstack_overflow: 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 RSB filling PBRSB-eIBRS: SW sequence + srbds: Not affected + tsx_async_abort: Not affected

Testing initiated at 1 April 2024 10:41 by user phoronix.