990KS March

Intel Core i9-9900KS testing with a ASUS PRIME Z390-A (1502 BIOS) and ASUS Intel UHD 630 CFL GT2 3GB on Ubuntu 20.04 via the Phoronix Test Suite.

Compare your own system(s) to this result file with the Phoronix Test Suite by running the command: phoronix-test-suite benchmark 2104012-IB-990KSMARC41
Jump To Table - Results

View

Do Not Show Noisy Results
Do Not Show Results With Incomplete Data
Do Not Show Results With Little Change/Spread
List Notable Results
Show Result Confidence Charts

Limit displaying results to tests within:

AV1 3 Tests
Timed Code Compilation 4 Tests
C/C++ Compiler Tests 7 Tests
CPU Massive 9 Tests
Creator Workloads 9 Tests
Encoding 5 Tests
Game Development 2 Tests
HPC - High Performance Computing 4 Tests
Machine Learning 3 Tests
Multi-Core 13 Tests
NVIDIA GPU Compute 2 Tests
OpenCL 2 Tests
Programmer / Developer System Benchmarks 6 Tests
Python Tests 4 Tests
Software Defined Radio 4 Tests
Server CPU Tests 8 Tests
Single-Threaded 2 Tests
Texture Compression 2 Tests
Video Encoding 5 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
1
March 31 2021
  8 Hours, 8 Minutes
2
March 31 2021
  8 Hours, 5 Minutes
3
March 31 2021
  8 Hours, 6 Minutes
Invert Hiding All Results Option
  8 Hours, 6 Minutes

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


990KS MarchProcessorMotherboardChipsetMemoryDiskGraphicsAudioMonitorNetworkOSKernelDesktopDisplay ServerOpenGLOpenCLCompilerFile-SystemScreen Resolution123Intel Core i9-9900KS @ 5.00GHz (8 Cores / 16 Threads)ASUS PRIME Z390-A (1502 BIOS)Intel Cannon Lake PCH32GB240GB Corsair Force MP510ASUS Intel UHD 630 CFL GT2 3GB (1200MHz)Realtek ALC1220G237HLIntel I219-VUbuntu 20.045.9.0-050900rc8daily20201005-generic (x86_64) 20201004GNOME Shell 3.36.2X Server 1.20.84.6 Mesa 20.2.6OpenCL 2.1GCC 9.3.0ext41920x1080OpenBenchmarking.orgKernel Details- Transparent Huge Pages: madviseCompiler Details- --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,brig,d,fortran,objc,obj-c++,gm2 --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-targets=nvptx-none,hsa --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -v Processor Details- Scaling Governor: intel_pstate powersave - CPU Microcode: 0xcc - Thermald 1.9.1 Python Details- Python 2.7.18rc1 + Python 3.8.2Security Details- itlb_multihit: KVM: Mitigation of VMX unsupported + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Mitigation of TSX disabled + tsx_async_abort: Mitigation of TSX disabled

123Result OverviewPhoronix Test Suite100%100%101%101%StockfishsrsLTEAOM AV1Mobile Neural NetworkOpenSCADSVT-VP9Zstd CompressionGNU GMP GMPbenchTimed Erlang/OTP CompilationBasis Universaldav1dSHOC Scalable HeterOgeneous ComputingSVT-HEVCSysbenchLuaRadiooneDNNXcompact3d Incompact3dASTC EncoderTimed Linux Kernel CompilationsimdjsonLiquid-DSPTimed Mesa CompilationGNU RadioViennaCLlibavif avifencBotantoyBrot Fractal GeneratorTimed Node.js CompilationSystemd Total Boot Time

990KS Marchcompress-zstd: 8, Long Mode - Compression Speedaom-av1: Speed 8 Realtime - Bosphorus 4Kaom-av1: Speed 6 Two-Pass - Bosphorus 1080paom-av1: Speed 0 Two-Pass - Bosphorus 1080pshoc: OpenCL - Triadopenscad: Projector Mount Swiveldav1d: Chimera 1080p 10-bitaom-av1: Speed 9 Realtime - Bosphorus 4Kaom-av1: Speed 8 Realtime - Bosphorus 1080pcompress-zstd: 3, Long Mode - Compression Speedcompress-zstd: 8 - Compression Speedsrslte: OFDM_Teststockfish: Total Timeonednn: IP Shapes 3D - f32 - CPUcompress-zstd: 19, Long Mode - Compression Speedmnn: SqueezeNetV1.0viennacl: OpenCL BLAS - sCOPYaom-av1: Speed 9 Realtime - Bosphorus 1080pmnn: inception-v3basis: UASTC Level 0onednn: Deconvolution Batch shapes_3d - u8s8f32 - CPUsrslte: PHY_DL_Testsrslte: PHY_DL_Testmnn: MobileNetV2_224onednn: Recurrent Neural Network Training - bf16bf16bf16 - CPUaom-av1: Speed 4 Two-Pass - Bosphorus 4Kluaradio: Complex Phasecompress-zstd: 8, Long Mode - Decompression Speedmnn: mobilenet-v1-1.0svt-vp9: VMAF Optimized - Bosphorus 1080psvt-vp9: PSNR/SSIM Optimized - Bosphorus 1080pshoc: OpenCL - S3Dopenscad: Leonardo Phone Case Slimavifenc: 0aom-av1: Speed 6 Two-Pass - Bosphorus 4Kmnn: resnet-v2-50openscad: Retro Carcompress-zstd: 19, Long Mode - Decompression Speedviennacl: OpenCL BLAS - dGEMM-NTviennacl: OpenCL BLAS - dCOPYcompress-zstd: 3 - Compression Speedgnuradio: Five Back to Back FIR Filtersluaradio: FM Deemphasis Filtersysbench: RAM / Memorycompress-zstd: 3 - Decompression Speedviennacl: OpenCL BLAS - dGEMV-Ndav1d: Summer Nature 4Kgmpbench: Total Timesvt-vp9: Visual Quality Optimized - Bosphorus 1080pavifenc: 10gnuradio: FIR Filteraom-av1: Speed 6 Realtime - Bosphorus 1080ponednn: IP Shapes 3D - u8s8f32 - CPUliquid-dsp: 2 - 256 - 57luaradio: Five Back to Back FIR Filtersonednn: Deconvolution Batch shapes_3d - f32 - CPUluaradio: Hilbert Transformbuild-erlang: Time To Compileavifenc: 6, Losslessaom-av1: Speed 4 Two-Pass - Bosphorus 1080psimdjson: DistinctUserIDonednn: Deconvolution Batch shapes_1d - f32 - CPUopenscad: Mini-ITX Caseavifenc: 6svt-hevc: 10 - Bosphorus 1080psvt-hevc: 7 - Bosphorus 1080pbasis: ETC1Sastcenc: Mediumshoc: OpenCL - Bus Speed Readbackgnuradio: Hilbert Transformcompress-zstd: 19 - Decompression Speedviennacl: CPU BLAS - sCOPYbotan: Twofishbotan: ChaCha20Poly1305onednn: IP Shapes 1D - f32 - CPUsimdjson: Kostyadav1d: Summer Nature 1080paom-av1: Speed 6 Realtime - Bosphorus 4Kincompact3d: input.i3d 129 Cells Per Directionviennacl: CPU BLAS - dGEMM-NTcompress-zstd: 19 - Compression Speedviennacl: OpenCL BLAS - dGEMV-Tliquid-dsp: 4 - 256 - 57viennacl: OpenCL BLAS - dAXPYavifenc: 10, Losslessviennacl: CPU BLAS - dGEMM-TNgnuradio: Signal Source (Cosine)viennacl: OpenCL BLAS - sDOTbasis: UASTC Level 2dav1d: Chimera 1080ptoybrot: C++ Tasksonednn: Convolution Batch Shapes Auto - u8s8f32 - CPUviennacl: CPU BLAS - sAXPYsimdjson: PartialTweetsliquid-dsp: 8 - 256 - 57viennacl: CPU BLAS - dGEMV-Nbuild-linux-kernel: Time To Compilecompress-zstd: 3, Long Mode - Decompression Speedopenscad: Pistolonednn: Recurrent Neural Network Inference - bf16bf16bf16 - CPUavifenc: 2onednn: Matrix Multiply Batch Shapes Transformer - u8s8f32 - CPUshoc: OpenCL - Reductiongnuradio: IIR Filtertoybrot: C++ Threadstoybrot: TBBshoc: OpenCL - Bus Speed Downloadonednn: Matrix Multiply Batch Shapes Transformer - f32 - CPUsvt-hevc: 1 - Bosphorus 1080ponednn: Convolution Batch Shapes Auto - f32 - CPUbotan: AES-256 - Decryptbotan: Twofish - Decryptshoc: OpenCL - FFT SPgnuradio: FM Deemphasis Filteronednn: IP Shapes 1D - u8s8f32 - CPUincompact3d: input.i3d 193 Cells Per Directioncompress-zstd: 8 - Decompression Speedbuild-mesa: Time To Compilebotan: KASUMI - Decryptastcenc: Exhaustivebotan: Blowfishbuild-nodejs: Time To Compileastcenc: Thoroughonednn: Recurrent Neural Network Training - f32 - CPUsysbench: CPUbasis: UASTC Level 3onednn: Recurrent Neural Network Inference - u8s8f32 - CPUtoybrot: OpenMPliquid-dsp: 16 - 256 - 57botan: Blowfish - Decryptbotan: ChaCha20Poly1305 - Decryptshoc: OpenCL - GEMM SGEMM_Nonednn: Recurrent Neural Network Inference - f32 - CPUonednn: Deconvolution Batch shapes_1d - u8s8f32 - CPUliquid-dsp: 1 - 256 - 57onednn: Recurrent Neural Network Training - u8s8f32 - CPUbotan: CAST-256 - Decryptshoc: OpenCL - Texture Read Bandwidthbotan: KASUMIshoc: OpenCL - Max SP Flopsshoc: OpenCL - MD5 Hashbotan: AES-256botan: CAST-256systemd-boot-total: Userspacesystemd-boot-total: Firmwaresystemd-boot-total: Loadersystemd-boot-total: Kernelsystemd-boot-total: Totalviennacl: OpenCL BLAS - dGEMM-TTviennacl: OpenCL BLAS - dGEMM-TNviennacl: OpenCL BLAS - dGEMM-NNviennacl: OpenCL BLAS - dDOTviennacl: OpenCL BLAS - sAXPYviennacl: CPU BLAS - dGEMM-TTviennacl: CPU BLAS - dGEMM-NNviennacl: CPU BLAS - dGEMV-Tviennacl: CPU BLAS - dDOTviennacl: CPU BLAS - dAXPYviennacl: CPU BLAS - dCOPYviennacl: CPU BLAS - sDOTaom-av1: Speed 0 Two-Pass - Bosphorus 4Ksimdjson: LargeRand123395.138.6819.680.3815.605392.931160.9147.78124.391164.0332.6126700000237778788.2610331.94.89932.9141.4829.4477.1723.97706293.2112.72.6723413.973.42704.24609.82.329173.31176.2920.993116.97572.0746.3425.30817.2593957.516.534.02379.41190.2486.828391.474186.537.9185.136250.3136.763.154822.325.061.832151301833331297.96.6399387.2113.00062.6746.764.538.8389841.90812.952230.27106.1923.4555.794929.0495618.03969.028.2436.078916.1083.577403.1648.2613.1634.337596934.134.534.825085666735.35.77836.53003.237.733.461733.644475614.033542.84.3845271333345.594.7144433.4100.7671854.8837.2533.4400336.7934666.1448544405829.21193.154336.8514.28304806.102437.29715.3682836.71.66747121.6911704312.456.415107.861113.8873543.597477.09714.92073408.1619380.3463.5431852.2643866500333333539.238908.311245.9201852.132.14490676296673414.52172.09156.9099112.9801760.700.38614815.972171.9542210915679355519412405015.715.818.93834.635.534.846.544.139.326.247.90.121.07378.437.9419.110.3715.233693.908164.4048.72122.641156.9327.6128466667235015918.1469531.84.88233.3140.3929.6707.0913.93198291.8112.22.6553445.233.41700.24601.32.333172.24175.2720.851816.94172.2586.3225.25417.3463963.916.534.12374.51195.5487.728491.744164.538.1184.256221.8136.223.160820.024.951.829101300466671292.46.6465987.1112.55862.3896.734.558.8650142.05313.005229.42105.9023.3905.775729.0520618.33983.128.2436.124913.0253.574553.10649.6713.1234.439722734.234.634.725157666735.35.77136.43011.437.633.374731.764484213.998842.94.3845207333345.694.5694437.7100.9781851.2237.3043.4335936.7270664.9447744413529.16413.151576.8414.30314811.181437.28815.3757837.71.66579121.8260734308.956.468107.764113.7734543.385476.80114.90593405.3219382.7163.5601850.8243854500263333539.017907.952245.8851851.142.14385676323333413.02172.07656.8932112.9431761.070.38614814.828171.9512210915679355519412405015.715.818.93834.635.534.846.544.139.326.247.90.121.07387.237.1219.210.3815.251894.966163.7648.78121.961177.1327.5128500000238322088.1678831.54.94333.3139.8229.7957.0893.93216295.0113.42.6833415.353.39698.14574.82.346172.10175.0620.847917.05471.7996.3025.41417.3673939.416.434.22388.41188.7489.528338.594180.337.9184.166217.6136.053.170818.325.071.823401306633331291.86.6155787.5112.48462.3966.754.548.8761842.08412.964229.36105.7823.3675.773729.1531620.23981.928.3434.656914.4543.565773.09647.6313.1434.384600334.134.634.825147333335.45.78736.53011.237.633.387732.834472914.010642.84.3745308666745.594.7684442.7100.8131854.4037.3263.4391436.7633665.0448054408529.17663.149436.8414.29364812.421436.74415.3573837.31.6655121.7424724307.756.475107.878113.8933543.951477.28414.91033405.9919367.3063.5931851.3043838500016667539.337908.474245.7831852.052.14452675996673413.88172.15056.9175112.9851760.610.3864815.617171.9582210915679355519412405015.715.818.93834.635.534.846.544.139.326.247.90.121.07OpenBenchmarking.org

Zstd Compression

This test measures the time needed to compress/decompress a sample file (a FreeBSD disk image - FreeBSD-12.2-RELEASE-amd64-memstick.img) using Zstd compression with options for different compression levels / settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.4.9Compression Level: 8, Long Mode - Compression Speed32190180270360450SE +/- 2.94, N = 3SE +/- 2.00, N = 3SE +/- 4.78, N = 3387.2378.4395.11. (CC) gcc options: -O3 -pthread -lz -llzma

AOM AV1

This is a test of the AOMedia AV1 encoder (libaom) developed by AOMedia and Google. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterAOM AV1 3.0Encoder Mode: Speed 8 Realtime - Input: Bosphorus 4K321918273645SE +/- 0.51, N = 4SE +/- 0.08, N = 3SE +/- 0.05, N = 337.1237.9438.681. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread

OpenBenchmarking.orgFrames Per Second, More Is BetterAOM AV1 3.0Encoder Mode: Speed 6 Two-Pass - Input: Bosphorus 1080p321510152025SE +/- 0.01, N = 3SE +/- 0.08, N = 3SE +/- 0.17, N = 319.2119.1119.681. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread

OpenBenchmarking.orgFrames Per Second, More Is BetterAOM AV1 3.0Encoder Mode: Speed 0 Two-Pass - Input: Bosphorus 1080p3210.08550.1710.25650.3420.4275SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 30.380.370.381. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: Triad32148121620SE +/- 0.03, N = 3SE +/- 0.01, N = 3SE +/- 0.03, N = 315.2515.2315.611. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -pthread -lmpi_cxx -lmpi

OpenSCAD

OpenSCAD is a programmer-focused solid 3D CAD modeller. OpenSCAD is free software and allows creating 3D CAD objects in a script-based modelling environment. This test profile will use the system-provided OpenSCAD program otherwise and time how long it takes tn render different SCAD assets to PNG output. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterOpenSCADRender: Projector Mount Swivel32120406080100SE +/- 0.08, N = 3SE +/- 0.42, N = 3SE +/- 0.21, N = 394.9793.9192.931. OpenSCAD version 2019.05

dav1d

Dav1d is an open-source, speedy AV1 video decoder. This test profile times how long it takes to decode sample AV1 video content. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFPS, More Is Betterdav1d 0.8.2Video Input: Chimera 1080p 10-bit3214080120160200SE +/- 1.42, N = 3SE +/- 1.75, N = 3SE +/- 0.13, N = 3163.76164.40160.91MIN: 104.03 / MAX: 384.96MIN: 103.73 / MAX: 392.07MIN: 103.41 / MAX: 410.161. (CC) gcc options: -pthread -lm

AOM AV1

This is a test of the AOMedia AV1 encoder (libaom) developed by AOMedia and Google. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterAOM AV1 3.0Encoder Mode: Speed 9 Realtime - Input: Bosphorus 4K3211122334455SE +/- 0.07, N = 3SE +/- 0.10, N = 3SE +/- 0.53, N = 1548.7848.7247.781. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread

OpenBenchmarking.orgFrames Per Second, More Is BetterAOM AV1 3.0Encoder Mode: Speed 8 Realtime - Input: Bosphorus 1080p321306090120150SE +/- 1.32, N = 12SE +/- 1.89, N = 3SE +/- 0.42, N = 3121.96122.64124.391. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread

Zstd Compression

This test measures the time needed to compress/decompress a sample file (a FreeBSD disk image - FreeBSD-12.2-RELEASE-amd64-memstick.img) using Zstd compression with options for different compression levels / settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.4.9Compression Level: 3, Long Mode - Compression Speed32130060090012001500SE +/- 6.67, N = 3SE +/- 10.72, N = 3SE +/- 5.14, N = 31177.11156.91164.01. (CC) gcc options: -O3 -pthread -lz -llzma

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.4.9Compression Level: 8 - Compression Speed32170140210280350SE +/- 0.88, N = 3SE +/- 2.23, N = 3SE +/- 1.66, N = 3327.5327.6332.61. (CC) gcc options: -O3 -pthread -lz -llzma

srsLTE

srsLTE is an open-source LTE software radio suite created by Software Radio Systems (SRS). srsLTE can be used for building your own software defined (SDR) LTE mobile network. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSamples / Second, More Is BettersrsLTE 20.10.1Test: OFDM_Test32130M60M90M120M150MSE +/- 665832.81, N = 3SE +/- 821245.67, N = 3SE +/- 416333.20, N = 31285000001284666671267000001. (CXX) g++ options: -std=c++11 -fno-strict-aliasing -march=native -mfpmath=sse -mavx2 -fvisibility=hidden -O3 -fno-trapping-math -fno-math-errno -rdynamic -lpthread -lbladeRF -lm -lfftw3f -lmbedcrypto

Stockfish

This is a test of Stockfish, an advanced open-source C++11 chess benchmark that can scale up to 512 CPU threads. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgNodes Per Second, More Is BetterStockfish 13Total Time3215M10M15M20M25MSE +/- 377829.22, N = 3SE +/- 299604.32, N = 3SE +/- 294389.71, N = 42383220823501591237778781. (CXX) g++ options: -lgcov -m64 -lpthread -fno-exceptions -std=c++17 -fprofile-use -fno-peel-loops -fno-tracer -pedantic -O3 -msse -msse3 -mpopcnt -mavx2 -msse4.1 -mssse3 -msse2 -mbmi2 -flto -flto=jobserver

oneDNN

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

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: IP Shapes 3D - Data Type: f32 - Engine: CPU321246810SE +/- 0.01559, N = 3SE +/- 0.00679, N = 3SE +/- 0.00549, N = 38.167888.146958.26103MIN: 8.03MIN: 8MIN: 8.141. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

Zstd Compression

This test measures the time needed to compress/decompress a sample file (a FreeBSD disk image - FreeBSD-12.2-RELEASE-amd64-memstick.img) using Zstd compression with options for different compression levels / settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.4.9Compression Level: 19, Long Mode - Compression Speed321714212835SE +/- 0.07, N = 3SE +/- 0.06, N = 3SE +/- 0.12, N = 331.531.831.91. (CC) gcc options: -O3 -pthread -lz -llzma

Mobile Neural Network

MNN is the Mobile Neural Network as a highly efficient, lightweight deep learning framework developed by Alibaba. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterMobile Neural Network 1.1.3Model: SqueezeNetV1.03211.11222.22443.33664.44885.561SE +/- 0.027, N = 3SE +/- 0.030, N = 3SE +/- 0.029, N = 34.9434.8824.899MIN: 4.65 / MAX: 16.5MIN: 4.68 / MAX: 5.58MIN: 4.71 / MAX: 7.41. (CXX) g++ options: -std=c++11 -O3 -fvisibility=hidden -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -ffast-math -fno-rtti -fno-exceptions -rdynamic -pthread -ldl

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - sCOPY321816243240SE +/- 0.06, N = 3SE +/- 0.10, N = 3SE +/- 0.06, N = 333.333.332.91. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

AOM AV1

This is a test of the AOMedia AV1 encoder (libaom) developed by AOMedia and Google. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterAOM AV1 3.0Encoder Mode: Speed 9 Realtime - Input: Bosphorus 1080p321306090120150SE +/- 2.02, N = 4SE +/- 0.46, N = 3SE +/- 1.63, N = 6139.82140.39141.481. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread

Mobile Neural Network

MNN is the Mobile Neural Network as a highly efficient, lightweight deep learning framework developed by Alibaba. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterMobile Neural Network 1.1.3Model: inception-v3321714212835SE +/- 0.16, N = 3SE +/- 0.04, N = 3SE +/- 0.08, N = 329.8029.6729.45MIN: 29.45 / MAX: 42.01MIN: 29.48 / MAX: 42.09MIN: 29.17 / MAX: 40.181. (CXX) g++ options: -std=c++11 -O3 -fvisibility=hidden -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -ffast-math -fno-rtti -fno-exceptions -rdynamic -pthread -ldl

Basis Universal

Basis Universal is a GPU texture codec. This test times how long it takes to convert sRGB PNGs into Basis Univeral assets with various settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBasis Universal 1.13Settings: UASTC Level 0321246810SE +/- 0.005, N = 3SE +/- 0.003, N = 3SE +/- 0.000, N = 37.0897.0917.1721. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread

oneDNN

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

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPU3210.89481.78962.68443.57924.474SE +/- 0.01771, N = 3SE +/- 0.01212, N = 3SE +/- 0.02149, N = 33.932163.931983.97706MIN: 3.62MIN: 3.64MIN: 3.651. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

srsLTE

srsLTE is an open-source LTE software radio suite created by Software Radio Systems (SRS). srsLTE can be used for building your own software defined (SDR) LTE mobile network. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgeNb Mb/s, More Is BettersrsLTE 20.10.1Test: PHY_DL_Test32160120180240300SE +/- 0.84, N = 3SE +/- 0.73, N = 3SE +/- 0.19, N = 3295.0291.8293.21. (CXX) g++ options: -std=c++11 -fno-strict-aliasing -march=native -mfpmath=sse -mavx2 -fvisibility=hidden -O3 -fno-trapping-math -fno-math-errno -rdynamic -lpthread -lbladeRF -lm -lfftw3f -lmbedcrypto

OpenBenchmarking.orgUE Mb/s, More Is BettersrsLTE 20.10.1Test: PHY_DL_Test321306090120150SE +/- 0.12, N = 3SE +/- 0.13, N = 3SE +/- 0.09, N = 3113.4112.2112.71. (CXX) g++ options: -std=c++11 -fno-strict-aliasing -march=native -mfpmath=sse -mavx2 -fvisibility=hidden -O3 -fno-trapping-math -fno-math-errno -rdynamic -lpthread -lbladeRF -lm -lfftw3f -lmbedcrypto

Mobile Neural Network

MNN is the Mobile Neural Network as a highly efficient, lightweight deep learning framework developed by Alibaba. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterMobile Neural Network 1.1.3Model: MobileNetV2_2243210.60371.20741.81112.41483.0185SE +/- 0.063, N = 3SE +/- 0.051, N = 3SE +/- 0.063, N = 32.6832.6552.672MIN: 2.37 / MAX: 4.91MIN: 2.35 / MAX: 3.81MIN: 2.34 / MAX: 4.311. (CXX) g++ options: -std=c++11 -O3 -fvisibility=hidden -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -ffast-math -fno-rtti -fno-exceptions -rdynamic -pthread -ldl

oneDNN

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

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPU3217001400210028003500SE +/- 0.66, N = 3SE +/- 26.13, N = 3SE +/- 2.52, N = 33415.353445.233413.97MIN: 3409.68MIN: 3408.23MIN: 3402.031. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

AOM AV1

This is a test of the AOMedia AV1 encoder (libaom) developed by AOMedia and Google. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterAOM AV1 3.0Encoder Mode: Speed 4 Two-Pass - Input: Bosphorus 4K3210.76951.5392.30853.0783.8475SE +/- 0.01, N = 3SE +/- 0.02, N = 3SE +/- 0.02, N = 33.393.413.421. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread

LuaRadio

LuaRadio is a lightweight software-defined radio (SDR) framework built atop LuaJIT. LuaRadio provides a suite of source, sink, and processing blocks, with a simple API for defining flow graphs, running flow graphs, creating blocks, and creating data types. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterLuaRadio 0.9.1Test: Complex Phase321150300450600750SE +/- 11.13, N = 3SE +/- 6.74, N = 3SE +/- 5.92, N = 3698.1700.2704.2

Zstd Compression

This test measures the time needed to compress/decompress a sample file (a FreeBSD disk image - FreeBSD-12.2-RELEASE-amd64-memstick.img) using Zstd compression with options for different compression levels / settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.4.9Compression Level: 8, Long Mode - Decompression Speed32110002000300040005000SE +/- 22.35, N = 3SE +/- 1.73, N = 3SE +/- 2.70, N = 34574.84601.34609.81. (CC) gcc options: -O3 -pthread -lz -llzma

Mobile Neural Network

MNN is the Mobile Neural Network as a highly efficient, lightweight deep learning framework developed by Alibaba. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterMobile Neural Network 1.1.3Model: mobilenet-v1-1.03210.52791.05581.58372.11162.6395SE +/- 0.010, N = 3SE +/- 0.006, N = 3SE +/- 0.007, N = 32.3462.3332.329MIN: 2.28 / MAX: 3.63MIN: 2.27 / MAX: 3.55MIN: 2.26 / MAX: 4.961. (CXX) g++ options: -std=c++11 -O3 -fvisibility=hidden -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -ffast-math -fno-rtti -fno-exceptions -rdynamic -pthread -ldl

SVT-VP9

This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-VP9 CPU-based multi-threaded video encoder for the VP9 video format with a sample YUV input video file. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-VP9 0.3Tuning: VMAF Optimized - Input: Bosphorus 1080p3214080120160200SE +/- 0.26, N = 3SE +/- 0.43, N = 3SE +/- 0.19, N = 3172.10172.24173.311. (CC) gcc options: -O3 -fcommon -fPIE -fPIC -fvisibility=hidden -pie -rdynamic -lpthread -lrt -lm

OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-VP9 0.3Tuning: PSNR/SSIM Optimized - Input: Bosphorus 1080p3214080120160200SE +/- 0.09, N = 3SE +/- 0.47, N = 3SE +/- 0.31, N = 3175.06175.27176.291. (CC) gcc options: -O3 -fcommon -fPIE -fPIC -fvisibility=hidden -pie -rdynamic -lpthread -lrt -lm

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: S3D321510152025SE +/- 0.02, N = 3SE +/- 0.04, N = 3SE +/- 0.04, N = 320.8520.8520.991. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -pthread -lmpi_cxx -lmpi

OpenSCAD

OpenSCAD is a programmer-focused solid 3D CAD modeller. OpenSCAD is free software and allows creating 3D CAD objects in a script-based modelling environment. This test profile will use the system-provided OpenSCAD program otherwise and time how long it takes tn render different SCAD assets to PNG output. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterOpenSCADRender: Leonardo Phone Case Slim32148121620SE +/- 0.01, N = 3SE +/- 0.02, N = 3SE +/- 0.01, N = 317.0516.9416.981. OpenSCAD version 2019.05

libavif avifenc

This is a test of the AOMedia libavif library testing the encoding of a JPEG image to AV1 Image Format (AVIF). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 0.9.0Encoder Speed: 03211632486480SE +/- 0.26, N = 3SE +/- 0.25, N = 3SE +/- 0.16, N = 371.8072.2672.071. (CXX) g++ options: -O3 -fPIC -lm

AOM AV1

This is a test of the AOMedia AV1 encoder (libaom) developed by AOMedia and Google. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterAOM AV1 3.0Encoder Mode: Speed 6 Two-Pass - Input: Bosphorus 4K321246810SE +/- 0.02, N = 3SE +/- 0.02, N = 3SE +/- 0.01, N = 36.306.326.341. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread

Mobile Neural Network

MNN is the Mobile Neural Network as a highly efficient, lightweight deep learning framework developed by Alibaba. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterMobile Neural Network 1.1.3Model: resnet-v2-50321612182430SE +/- 0.05, N = 3SE +/- 0.02, N = 3SE +/- 0.04, N = 325.4125.2525.31MIN: 25.17 / MAX: 37.2MIN: 25.09 / MAX: 35.04MIN: 25.08 / MAX: 37.211. (CXX) g++ options: -std=c++11 -O3 -fvisibility=hidden -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -ffast-math -fno-rtti -fno-exceptions -rdynamic -pthread -ldl

OpenSCAD

OpenSCAD is a programmer-focused solid 3D CAD modeller. OpenSCAD is free software and allows creating 3D CAD objects in a script-based modelling environment. This test profile will use the system-provided OpenSCAD program otherwise and time how long it takes tn render different SCAD assets to PNG output. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterOpenSCADRender: Retro Car32148121620SE +/- 0.02, N = 3SE +/- 0.10, N = 3SE +/- 0.02, N = 317.3717.3517.261. OpenSCAD version 2019.05

Zstd Compression

This test measures the time needed to compress/decompress a sample file (a FreeBSD disk image - FreeBSD-12.2-RELEASE-amd64-memstick.img) using Zstd compression with options for different compression levels / settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.4.9Compression Level: 19, Long Mode - Decompression Speed3219001800270036004500SE +/- 5.83, N = 3SE +/- 0.96, N = 3SE +/- 9.00, N = 33939.43963.93957.51. (CC) gcc options: -O3 -pthread -lz -llzma

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPs/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dGEMM-NT32148121620SE +/- 0.03, N = 3SE +/- 0.03, N = 3SE +/- 0.00, N = 316.416.516.51. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dCOPY321816243240SE +/- 0.03, N = 3SE +/- 0.03, N = 3SE +/- 0.03, N = 334.234.134.01. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

Zstd Compression

This test measures the time needed to compress/decompress a sample file (a FreeBSD disk image - FreeBSD-12.2-RELEASE-amd64-memstick.img) using Zstd compression with options for different compression levels / settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.4.9Compression Level: 3 - Compression Speed3215001000150020002500SE +/- 12.46, N = 3SE +/- 16.90, N = 3SE +/- 12.12, N = 32388.42374.52379.41. (CC) gcc options: -O3 -pthread -lz -llzma

GNU Radio

GNU Radio is a free software development toolkit providing signal processing blocks to implement software-defined radios (SDR) and signal processing systems. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterGNU RadioTest: Five Back to Back FIR Filters32130060090012001500SE +/- 15.25, N = 3SE +/- 9.72, N = 3SE +/- 12.45, N = 31188.71195.51190.21. 3.8.1.0

LuaRadio

LuaRadio is a lightweight software-defined radio (SDR) framework built atop LuaJIT. LuaRadio provides a suite of source, sink, and processing blocks, with a simple API for defining flow graphs, running flow graphs, creating blocks, and creating data types. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterLuaRadio 0.9.1Test: FM Deemphasis Filter321110220330440550SE +/- 0.27, N = 3SE +/- 1.98, N = 3SE +/- 1.86, N = 3489.5487.7486.8

Sysbench

This is a benchmark of Sysbench with the built-in CPU and memory sub-tests. Sysbench is a scriptable multi-threaded benchmark tool based on LuaJIT. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/sec, More Is BetterSysbench 1.0.20Test: RAM / Memory3216K12K18K24K30KSE +/- 41.95, N = 3SE +/- 138.44, N = 3SE +/- 98.12, N = 328338.5928491.7428391.471. (CC) gcc options: -pthread -O2 -funroll-loops -rdynamic -ldl -laio -lm

Zstd Compression

This test measures the time needed to compress/decompress a sample file (a FreeBSD disk image - FreeBSD-12.2-RELEASE-amd64-memstick.img) using Zstd compression with options for different compression levels / settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.4.9Compression Level: 3 - Decompression Speed3219001800270036004500SE +/- 1.88, N = 3SE +/- 19.69, N = 3SE +/- 4.56, N = 34180.34164.54186.51. (CC) gcc options: -O3 -pthread -lz -llzma

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dGEMV-N321918273645SE +/- 0.00, N = 337.938.137.91. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

dav1d

Dav1d is an open-source, speedy AV1 video decoder. This test profile times how long it takes to decode sample AV1 video content. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFPS, More Is Betterdav1d 0.8.2Video Input: Summer Nature 4K3214080120160200SE +/- 0.19, N = 3SE +/- 0.13, N = 3SE +/- 0.14, N = 3184.16184.25185.13MIN: 173.76 / MAX: 209.16MIN: 173.91 / MAX: 209.75MIN: 174.56 / MAX: 210.521. (CC) gcc options: -pthread -lm

GNU GMP GMPbench

GMPbench is a test of the GNU Multiple Precision Arithmetic (GMP) Library. GMPbench is a single-threaded integer benchmark that leverages the GMP library to stress the CPU with widening integer multiplication. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGMPbench Score, More Is BetterGNU GMP GMPbench 6.2.1Total Time321130026003900520065006217.66221.86250.31. (CC) gcc options: -O3 -fomit-frame-pointer -lm

SVT-VP9

This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-VP9 CPU-based multi-threaded video encoder for the VP9 video format with a sample YUV input video file. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-VP9 0.3Tuning: Visual Quality Optimized - Input: Bosphorus 1080p321306090120150SE +/- 0.35, N = 3SE +/- 0.13, N = 3SE +/- 0.26, N = 3136.05136.22136.761. (CC) gcc options: -O3 -fcommon -fPIE -fPIC -fvisibility=hidden -pie -rdynamic -lpthread -lrt -lm

libavif avifenc

This is a test of the AOMedia libavif library testing the encoding of a JPEG image to AV1 Image Format (AVIF). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 0.9.0Encoder Speed: 103210.71331.42662.13992.85323.5665SE +/- 0.012, N = 3SE +/- 0.010, N = 3SE +/- 0.004, N = 33.1703.1603.1541. (CXX) g++ options: -O3 -fPIC -lm

GNU Radio

GNU Radio is a free software development toolkit providing signal processing blocks to implement software-defined radios (SDR) and signal processing systems. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterGNU RadioTest: FIR Filter3212004006008001000SE +/- 2.94, N = 3SE +/- 3.57, N = 3SE +/- 1.65, N = 3818.3820.0822.31. 3.8.1.0

AOM AV1

This is a test of the AOMedia AV1 encoder (libaom) developed by AOMedia and Google. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterAOM AV1 3.0Encoder Mode: Speed 6 Realtime - Input: Bosphorus 1080p321612182430SE +/- 0.22, N = 3SE +/- 0.29, N = 5SE +/- 0.24, N = 325.0724.9525.061. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread

oneDNN

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

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPU3210.41220.82441.23661.64882.061SE +/- 0.00185, N = 3SE +/- 0.00543, N = 3SE +/- 0.00346, N = 31.823401.829101.83215MIN: 1.78MIN: 1.78MIN: 1.791. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

Liquid-DSP

LiquidSDR's Liquid-DSP is a software-defined radio (SDR) digital signal processing library. This test profile runs a multi-threaded benchmark of this SDR/DSP library focused on embedded platform usage. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgsamples/s, More Is BetterLiquid-DSP 2021.01.31Threads: 2 - Buffer Length: 256 - Filter Length: 5732130M60M90M120M150MSE +/- 35276.68, N = 3SE +/- 367891.89, N = 3SE +/- 40551.75, N = 31306633331300466671301833331. (CC) gcc options: -O3 -pthread -lm -lc -lliquid

LuaRadio

LuaRadio is a lightweight software-defined radio (SDR) framework built atop LuaJIT. LuaRadio provides a suite of source, sink, and processing blocks, with a simple API for defining flow graphs, running flow graphs, creating blocks, and creating data types. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterLuaRadio 0.9.1Test: Five Back to Back FIR Filters32130060090012001500SE +/- 4.98, N = 3SE +/- 3.21, N = 3SE +/- 2.64, N = 31291.81292.41297.9

oneDNN

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

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPU321246810SE +/- 0.00264, N = 3SE +/- 0.01698, N = 3SE +/- 0.00984, N = 36.615576.646596.63993MIN: 5.99MIN: 6MIN: 61. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

LuaRadio

LuaRadio is a lightweight software-defined radio (SDR) framework built atop LuaJIT. LuaRadio provides a suite of source, sink, and processing blocks, with a simple API for defining flow graphs, running flow graphs, creating blocks, and creating data types. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterLuaRadio 0.9.1Test: Hilbert Transform32120406080100SE +/- 0.32, N = 3SE +/- 0.06, N = 3SE +/- 0.03, N = 387.587.187.2

Timed Erlang/OTP Compilation

This test times how long it takes to compile Erlang/OTP. Erlang is a programming language and run-time for massively scalable soft real-time systems with high availability requirements. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Erlang/OTP Compilation 23.2Time To Compile321306090120150SE +/- 0.31, N = 3SE +/- 0.40, N = 3SE +/- 0.14, N = 3112.48112.56113.00

libavif avifenc

This is a test of the AOMedia libavif library testing the encoding of a JPEG image to AV1 Image Format (AVIF). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 0.9.0Encoder Speed: 6, Lossless3211428425670SE +/- 0.14, N = 3SE +/- 0.11, N = 3SE +/- 0.15, N = 362.4062.3962.671. (CXX) g++ options: -O3 -fPIC -lm

AOM AV1

This is a test of the AOMedia AV1 encoder (libaom) developed by AOMedia and Google. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterAOM AV1 3.0Encoder Mode: Speed 4 Two-Pass - Input: Bosphorus 1080p321246810SE +/- 0.01, N = 3SE +/- 0.02, N = 3SE +/- 0.01, N = 36.756.736.761. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread

simdjson

This is a benchmark of SIMDJSON, a high performance JSON parser. SIMDJSON aims to be the fastest JSON parser and is used by projects like Microsoft FishStore, Yandex ClickHouse, Shopify, and others. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is Bettersimdjson 0.8.2Throughput Test: DistinctUserID3211.02382.04763.07144.09525.119SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.00, N = 34.544.554.531. (CXX) g++ options: -O3 -pthread

oneDNN

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

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPU321246810SE +/- 0.00592, N = 3SE +/- 0.01330, N = 3SE +/- 0.00652, N = 38.876188.865018.83898MIN: 4.77MIN: 4.67MIN: 4.681. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

OpenSCAD

OpenSCAD is a programmer-focused solid 3D CAD modeller. OpenSCAD is free software and allows creating 3D CAD objects in a script-based modelling environment. This test profile will use the system-provided OpenSCAD program otherwise and time how long it takes tn render different SCAD assets to PNG output. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterOpenSCADRender: Mini-ITX Case3211020304050SE +/- 0.02, N = 3SE +/- 0.12, N = 3SE +/- 0.07, N = 342.0842.0541.911. OpenSCAD version 2019.05

libavif avifenc

This is a test of the AOMedia libavif library testing the encoding of a JPEG image to AV1 Image Format (AVIF). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 0.9.0Encoder Speed: 63213691215SE +/- 0.01, N = 3SE +/- 0.04, N = 3SE +/- 0.03, N = 312.9613.0112.951. (CXX) g++ options: -O3 -fPIC -lm

SVT-HEVC

This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-HEVC CPU-based multi-threaded video encoder for the HEVC / H.265 video format with a sample 1080p YUV video file. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-HEVC 1.5.0Tuning: 10 - Input: Bosphorus 1080p32150100150200250SE +/- 0.43, N = 3SE +/- 0.19, N = 3SE +/- 0.11, N = 3229.36229.42230.271. (CC) gcc options: -fPIE -fPIC -O3 -O2 -pie -rdynamic -lpthread -lrt

OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-HEVC 1.5.0Tuning: 7 - Input: Bosphorus 1080p32120406080100SE +/- 0.10, N = 3SE +/- 0.06, N = 3SE +/- 0.13, N = 3105.78105.90106.191. (CC) gcc options: -fPIE -fPIC -O3 -O2 -pie -rdynamic -lpthread -lrt

Basis Universal

Basis Universal is a GPU texture codec. This test times how long it takes to convert sRGB PNGs into Basis Univeral assets with various settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBasis Universal 1.13Settings: ETC1S321612182430SE +/- 0.05, N = 3SE +/- 0.08, N = 3SE +/- 0.03, N = 323.3723.3923.461. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread

ASTC Encoder

ASTC Encoder (astcenc) is for the Adaptive Scalable Texture Compression (ASTC) format commonly used with OpenGL, OpenGL ES, and Vulkan graphics APIs. This test profile does a coding test of both compression/decompression. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterASTC Encoder 2.4Preset: Medium3211.30392.60783.91175.21566.5195SE +/- 0.0062, N = 3SE +/- 0.0051, N = 3SE +/- 0.0231, N = 35.77375.77575.79491. (CXX) g++ options: -O3 -flto -pthread

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: Bus Speed Readback321714212835SE +/- 0.20, N = 3SE +/- 0.10, N = 3SE +/- 0.22, N = 329.1529.0529.051. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -pthread -lmpi_cxx -lmpi

GNU Radio

GNU Radio is a free software development toolkit providing signal processing blocks to implement software-defined radios (SDR) and signal processing systems. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterGNU RadioTest: Hilbert Transform321130260390520650SE +/- 1.23, N = 3SE +/- 2.62, N = 3SE +/- 1.07, N = 3620.2618.3618.01. 3.8.1.0

Zstd Compression

This test measures the time needed to compress/decompress a sample file (a FreeBSD disk image - FreeBSD-12.2-RELEASE-amd64-memstick.img) using Zstd compression with options for different compression levels / settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.4.9Compression Level: 19 - Decompression Speed3219001800270036004500SE +/- 4.27, N = 3SE +/- 1.35, N = 3SE +/- 10.42, N = 33981.93983.13969.01. (CC) gcc options: -O3 -pthread -lz -llzma

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: CPU BLAS - sCOPY321714212835SE +/- 0.03, N = 3SE +/- 0.03, N = 3SE +/- 0.03, N = 328.328.228.21. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

Botan

Botan is a BSD-licensed cross-platform open-source C++ crypto library "cryptography toolkit" that supports most publicly known cryptographic algorithms. The project's stated goal is to be "the best option for cryptography in C++ by offering the tools necessary to implement a range of practical systems, such as TLS protocol, X.509 certificates, modern AEAD ciphers, PKCS#11 and TPM hardware support, password hashing, and post quantum crypto schemes." Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterBotan 2.17.3Test: Twofish32190180270360450SE +/- 1.83, N = 3SE +/- 0.12, N = 3SE +/- 0.16, N = 3434.66436.12436.081. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt

OpenBenchmarking.orgMiB/s, More Is BetterBotan 2.17.3Test: ChaCha20Poly13053212004006008001000SE +/- 1.80, N = 3SE +/- 3.02, N = 3SE +/- 0.12, N = 3914.45913.03916.111. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt

oneDNN

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

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: IP Shapes 1D - Data Type: f32 - Engine: CPU3210.80491.60982.41473.21964.0245SE +/- 0.00251, N = 3SE +/- 0.00753, N = 3SE +/- 0.01205, N = 33.565773.574553.57740MIN: 3.26MIN: 3.24MIN: 3.231. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

simdjson

This is a benchmark of SIMDJSON, a high performance JSON parser. SIMDJSON aims to be the fastest JSON parser and is used by projects like Microsoft FishStore, Yandex ClickHouse, Shopify, and others. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is Bettersimdjson 0.8.2Throughput Test: Kostya3210.69751.3952.09252.793.4875SE +/- 0.01, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 33.093.103.101. (CXX) g++ options: -O3 -pthread

dav1d

Dav1d is an open-source, speedy AV1 video decoder. This test profile times how long it takes to decode sample AV1 video content. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFPS, More Is Betterdav1d 0.8.2Video Input: Summer Nature 1080p321140280420560700SE +/- 0.54, N = 3SE +/- 0.67, N = 3SE +/- 1.42, N = 3647.63649.67648.26MIN: 582.84 / MAX: 717.06MIN: 588.3 / MAX: 717.46MIN: 589.55 / MAX: 721.081. (CC) gcc options: -pthread -lm

AOM AV1

This is a test of the AOMedia AV1 encoder (libaom) developed by AOMedia and Google. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterAOM AV1 3.0Encoder Mode: Speed 6 Realtime - Input: Bosphorus 4K3213691215SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.02, N = 313.1413.1213.161. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread

Xcompact3d Incompact3d

Xcompact3d Incompact3d is a Fortran-MPI based, finite difference high-performance code for solving the incompressible Navier-Stokes equation and as many as you need scalar transport equations. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterXcompact3d Incompact3d 2021-03-11Input: input.i3d 129 Cells Per Direction321816243240SE +/- 0.31, N = 3SE +/- 0.34, N = 3SE +/- 0.35, N = 334.3834.4434.341. (F9X) gfortran options: -cpp -O2 -funroll-loops -floop-optimize -fcray-pointer -fbacktrace -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPs/s, More Is BetterViennaCL 1.7.1Test: CPU BLAS - dGEMM-NT321816243240SE +/- 0.03, N = 3SE +/- 0.07, N = 3SE +/- 0.03, N = 334.134.234.11. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

Zstd Compression

This test measures the time needed to compress/decompress a sample file (a FreeBSD disk image - FreeBSD-12.2-RELEASE-amd64-memstick.img) using Zstd compression with options for different compression levels / settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.4.9Compression Level: 19 - Compression Speed321816243240SE +/- 0.06, N = 3SE +/- 0.07, N = 3SE +/- 0.12, N = 334.634.634.51. (CC) gcc options: -O3 -pthread -lz -llzma

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dGEMV-T321816243240SE +/- 0.03, N = 3SE +/- 0.03, N = 3SE +/- 0.00, N = 334.834.734.81. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

Liquid-DSP

LiquidSDR's Liquid-DSP is a software-defined radio (SDR) digital signal processing library. This test profile runs a multi-threaded benchmark of this SDR/DSP library focused on embedded platform usage. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgsamples/s, More Is BetterLiquid-DSP 2021.01.31Threads: 4 - Buffer Length: 256 - Filter Length: 5732150M100M150M200M250MSE +/- 321679.62, N = 3SE +/- 375470.08, N = 3SE +/- 265476.51, N = 32514733332515766672508566671. (CC) gcc options: -O3 -pthread -lm -lc -lliquid

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dAXPY321816243240SE +/- 0.03, N = 3SE +/- 0.00, N = 3SE +/- 0.03, N = 335.435.335.31. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

libavif avifenc

This is a test of the AOMedia libavif library testing the encoding of a JPEG image to AV1 Image Format (AVIF). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 0.9.0Encoder Speed: 10, Lossless3211.30212.60423.90635.20846.5105SE +/- 0.012, N = 3SE +/- 0.013, N = 3SE +/- 0.013, N = 35.7875.7715.7781. (CXX) g++ options: -O3 -fPIC -lm

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPs/s, More Is BetterViennaCL 1.7.1Test: CPU BLAS - dGEMM-TN321816243240SE +/- 0.00, N = 3SE +/- 0.07, N = 3SE +/- 0.03, N = 336.536.436.51. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

GNU Radio

GNU Radio is a free software development toolkit providing signal processing blocks to implement software-defined radios (SDR) and signal processing systems. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterGNU RadioTest: Signal Source (Cosine)3216001200180024003000SE +/- 2.72, N = 3SE +/- 2.10, N = 3SE +/- 3.36, N = 33011.23011.43003.21. 3.8.1.0

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - sDOT321918273645SE +/- 0.03, N = 3SE +/- 0.03, N = 3SE +/- 0.00, N = 337.637.637.71. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

Basis Universal

Basis Universal is a GPU texture codec. This test times how long it takes to convert sRGB PNGs into Basis Univeral assets with various settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBasis Universal 1.13Settings: UASTC Level 2321816243240SE +/- 0.12, N = 3SE +/- 0.10, N = 3SE +/- 0.02, N = 333.3933.3733.461. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread

dav1d

Dav1d is an open-source, speedy AV1 video decoder. This test profile times how long it takes to decode sample AV1 video content. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFPS, More Is Betterdav1d 0.8.2Video Input: Chimera 1080p321160320480640800SE +/- 0.90, N = 3SE +/- 2.01, N = 3SE +/- 1.00, N = 3732.83731.76733.64MIN: 537.85 / MAX: 1141.98MIN: 538.01 / MAX: 1138.02MIN: 538.11 / MAX: 1144.181. (CC) gcc options: -pthread -lm

toyBrot Fractal Generator

ToyBrot is a Mandelbrot fractal generator supporting C++ threads/tasks, OpenMP, Intel Threaded Building Blocks (TBB), and other targets. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BettertoyBrot Fractal Generator 2020-11-18Implementation: C++ Tasks32110K20K30K40K50KSE +/- 38.73, N = 3SE +/- 118.82, N = 3SE +/- 16.42, N = 34472944842447561. (CXX) g++ options: -O3 -lpthread

oneDNN

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

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPU32148121620SE +/- 0.02, N = 3SE +/- 0.01, N = 3SE +/- 0.03, N = 314.0114.0014.03MIN: 13.88MIN: 13.9MIN: 13.921. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: CPU BLAS - sAXPY3211020304050SE +/- 0.09, N = 3SE +/- 0.07, N = 3SE +/- 0.03, N = 342.842.942.81. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

simdjson

This is a benchmark of SIMDJSON, a high performance JSON parser. SIMDJSON aims to be the fastest JSON parser and is used by projects like Microsoft FishStore, Yandex ClickHouse, Shopify, and others. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is Bettersimdjson 0.8.2Throughput Test: PartialTweets3210.98551.9712.95653.9424.9275SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 34.374.384.381. (CXX) g++ options: -O3 -pthread

Liquid-DSP

LiquidSDR's Liquid-DSP is a software-defined radio (SDR) digital signal processing library. This test profile runs a multi-threaded benchmark of this SDR/DSP library focused on embedded platform usage. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgsamples/s, More Is BetterLiquid-DSP 2021.01.31Threads: 8 - Buffer Length: 256 - Filter Length: 57321100M200M300M400M500MSE +/- 2209451.92, N = 3SE +/- 514274.03, N = 3SE +/- 1523508.38, N = 34530866674520733334527133331. (CC) gcc options: -O3 -pthread -lm -lc -lliquid

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: CPU BLAS - dGEMV-N3211020304050SE +/- 0.03, N = 3SE +/- 0.03, N = 3SE +/- 0.00, N = 345.545.645.51. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

Timed Linux Kernel Compilation

This test times how long it takes to build the Linux kernel in a default configuration (defconfig) for the architecture being tested. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Linux Kernel Compilation 5.10.20Time To Compile32120406080100SE +/- 0.47, N = 3SE +/- 0.60, N = 3SE +/- 0.78, N = 394.7794.5794.71

Zstd Compression

This test measures the time needed to compress/decompress a sample file (a FreeBSD disk image - FreeBSD-12.2-RELEASE-amd64-memstick.img) using Zstd compression with options for different compression levels / settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.4.9Compression Level: 3, Long Mode - Decompression Speed32110002000300040005000SE +/- 1.29, N = 3SE +/- 5.89, N = 3SE +/- 7.93, N = 34442.74437.74433.41. (CC) gcc options: -O3 -pthread -lz -llzma

OpenSCAD

OpenSCAD is a programmer-focused solid 3D CAD modeller. OpenSCAD is free software and allows creating 3D CAD objects in a script-based modelling environment. This test profile will use the system-provided OpenSCAD program otherwise and time how long it takes tn render different SCAD assets to PNG output. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterOpenSCADRender: Pistol32120406080100SE +/- 0.09, N = 3SE +/- 0.21, N = 3SE +/- 0.07, N = 3100.81100.98100.771. OpenSCAD version 2019.05

oneDNN

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

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPU321400800120016002000SE +/- 1.53, N = 3SE +/- 1.24, N = 3SE +/- 3.26, N = 31854.401851.221854.88MIN: 1846.77MIN: 1841.63MIN: 1847.251. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

libavif avifenc

This is a test of the AOMedia libavif library testing the encoding of a JPEG image to AV1 Image Format (AVIF). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 0.9.0Encoder Speed: 2321918273645SE +/- 0.09, N = 3SE +/- 0.11, N = 3SE +/- 0.06, N = 337.3337.3037.251. (CXX) g++ options: -O3 -fPIC -lm

oneDNN

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

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPU3210.7741.5482.3223.0963.87SE +/- 0.00188, N = 3SE +/- 0.00274, N = 3SE +/- 0.00410, N = 33.439143.433593.44003MIN: 3.18MIN: 3.17MIN: 3.151. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: Reduction321816243240SE +/- 0.03, N = 3SE +/- 0.07, N = 3SE +/- 0.05, N = 336.7636.7336.791. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -pthread -lmpi_cxx -lmpi

GNU Radio

GNU Radio is a free software development toolkit providing signal processing blocks to implement software-defined radios (SDR) and signal processing systems. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterGNU RadioTest: IIR Filter321140280420560700SE +/- 0.81, N = 3SE +/- 1.34, N = 3SE +/- 1.14, N = 3665.0664.9666.11. 3.8.1.0

toyBrot Fractal Generator

ToyBrot is a Mandelbrot fractal generator supporting C++ threads/tasks, OpenMP, Intel Threaded Building Blocks (TBB), and other targets. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BettertoyBrot Fractal Generator 2020-11-18Implementation: C++ Threads32110K20K30K40K50KSE +/- 36.98, N = 3SE +/- 25.27, N = 3SE +/- 27.00, N = 34480544774448541. (CXX) g++ options: -O3 -lpthread

OpenBenchmarking.orgms, Fewer Is BettertoyBrot Fractal Generator 2020-11-18Implementation: TBB3219K18K27K36K45KSE +/- 82.93, N = 3SE +/- 72.84, N = 3SE +/- 59.38, N = 34408544135440581. (CXX) g++ options: -O3 -lpthread

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: Bus Speed Download321714212835SE +/- 0.16, N = 3SE +/- 0.05, N = 3SE +/- 0.10, N = 329.1829.1629.211. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -pthread -lmpi_cxx -lmpi

oneDNN

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

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU3210.70971.41942.12912.83883.5485SE +/- 0.00951, N = 3SE +/- 0.00442, N = 3SE +/- 0.00701, N = 33.149433.151573.15433MIN: 3.09MIN: 3.1MIN: 3.11. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

SVT-HEVC

This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-HEVC CPU-based multi-threaded video encoder for the HEVC / H.265 video format with a sample 1080p YUV video file. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-HEVC 1.5.0Tuning: 1 - Input: Bosphorus 1080p321246810SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 36.846.846.851. (CC) gcc options: -fPIE -fPIC -O3 -O2 -pie -rdynamic -lpthread -lrt

oneDNN

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

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU32148121620SE +/- 0.02, N = 3SE +/- 0.03, N = 3SE +/- 0.02, N = 314.2914.3014.28MIN: 14.2MIN: 14.2MIN: 14.21. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

Botan

Botan is a BSD-licensed cross-platform open-source C++ crypto library "cryptography toolkit" that supports most publicly known cryptographic algorithms. The project's stated goal is to be "the best option for cryptography in C++ by offering the tools necessary to implement a range of practical systems, such as TLS protocol, X.509 certificates, modern AEAD ciphers, PKCS#11 and TPM hardware support, password hashing, and post quantum crypto schemes." Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterBotan 2.17.3Test: AES-256 - Decrypt32110002000300040005000SE +/- 0.04, N = 3SE +/- 0.85, N = 3SE +/- 4.51, N = 34812.424811.184806.101. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt

OpenBenchmarking.orgMiB/s, More Is BetterBotan 2.17.3Test: Twofish - Decrypt32190180270360450SE +/- 0.95, N = 3SE +/- 0.15, N = 3SE +/- 0.18, N = 3436.74437.29437.301. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: FFT SP32148121620SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 315.3615.3815.371. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -pthread -lmpi_cxx -lmpi

GNU Radio

GNU Radio is a free software development toolkit providing signal processing blocks to implement software-defined radios (SDR) and signal processing systems. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterGNU RadioTest: FM Deemphasis Filter3212004006008001000SE +/- 1.55, N = 3SE +/- 1.84, N = 3SE +/- 2.38, N = 3837.3837.7836.71. 3.8.1.0

oneDNN

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

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: IP Shapes 1D - Data Type: u8s8f32 - Engine: CPU3210.37520.75041.12561.50081.876SE +/- 0.00123, N = 3SE +/- 0.00242, N = 3SE +/- 0.00165, N = 31.665501.665791.66747MIN: 1.51MIN: 1.51MIN: 1.511. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

Xcompact3d Incompact3d

Xcompact3d Incompact3d is a Fortran-MPI based, finite difference high-performance code for solving the incompressible Navier-Stokes equation and as many as you need scalar transport equations. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterXcompact3d Incompact3d 2021-03-11Input: input.i3d 193 Cells Per Direction321306090120150SE +/- 0.05, N = 3SE +/- 0.03, N = 3SE +/- 0.01, N = 3121.74121.83121.691. (F9X) gfortran options: -cpp -O2 -funroll-loops -floop-optimize -fcray-pointer -fbacktrace -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi

Zstd Compression

This test measures the time needed to compress/decompress a sample file (a FreeBSD disk image - FreeBSD-12.2-RELEASE-amd64-memstick.img) using Zstd compression with options for different compression levels / settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.4.9Compression Level: 8 - Decompression Speed3219001800270036004500SE +/- 5.58, N = 3SE +/- 5.34, N = 3SE +/- 11.68, N = 34307.74308.94312.41. (CC) gcc options: -O3 -pthread -lz -llzma

Timed Mesa Compilation

This test profile times how long it takes to compile Mesa with Meson/Ninja. For minimizing build dependencies and avoid versioning conflicts, test this is just the core Mesa build without LLVM or the extra Gallium3D/Mesa drivers enabled. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Mesa Compilation 21.0Time To Compile3211326395265SE +/- 0.04, N = 3SE +/- 0.04, N = 3SE +/- 0.02, N = 356.4856.4756.42

Botan

Botan is a BSD-licensed cross-platform open-source C++ crypto library "cryptography toolkit" that supports most publicly known cryptographic algorithms. The project's stated goal is to be "the best option for cryptography in C++ by offering the tools necessary to implement a range of practical systems, such as TLS protocol, X.509 certificates, modern AEAD ciphers, PKCS#11 and TPM hardware support, password hashing, and post quantum crypto schemes." Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterBotan 2.17.3Test: KASUMI - Decrypt32120406080100SE +/- 0.04, N = 3SE +/- 0.07, N = 3SE +/- 0.03, N = 3107.88107.76107.861. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt

ASTC Encoder

ASTC Encoder (astcenc) is for the Adaptive Scalable Texture Compression (ASTC) format commonly used with OpenGL, OpenGL ES, and Vulkan graphics APIs. This test profile does a coding test of both compression/decompression. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterASTC Encoder 2.4Preset: Exhaustive321306090120150SE +/- 0.16, N = 3SE +/- 0.14, N = 3SE +/- 0.12, N = 3113.89113.77113.891. (CXX) g++ options: -O3 -flto -pthread

Botan

Botan is a BSD-licensed cross-platform open-source C++ crypto library "cryptography toolkit" that supports most publicly known cryptographic algorithms. The project's stated goal is to be "the best option for cryptography in C++ by offering the tools necessary to implement a range of practical systems, such as TLS protocol, X.509 certificates, modern AEAD ciphers, PKCS#11 and TPM hardware support, password hashing, and post quantum crypto schemes." Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterBotan 2.17.3Test: Blowfish321120240360480600SE +/- 0.10, N = 3SE +/- 0.40, N = 3SE +/- 0.05, N = 3543.95543.39543.601. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt

Timed Node.js Compilation

This test profile times how long it takes to build/compile Node.js itself from source. Node.js is a JavaScript run-time built from the Chrome V8 JavaScript engine while itself is written in C/C++. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Node.js Compilation 15.11Time To Compile321100200300400500SE +/- 0.19, N = 3SE +/- 0.34, N = 3SE +/- 0.32, N = 3477.28476.80477.10

ASTC Encoder

ASTC Encoder (astcenc) is for the Adaptive Scalable Texture Compression (ASTC) format commonly used with OpenGL, OpenGL ES, and Vulkan graphics APIs. This test profile does a coding test of both compression/decompression. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterASTC Encoder 2.4Preset: Thorough32148121620SE +/- 0.02, N = 3SE +/- 0.02, N = 3SE +/- 0.01, N = 314.9114.9114.921. (CXX) g++ options: -O3 -flto -pthread

oneDNN

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

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU3217001400210028003500SE +/- 4.35, N = 3SE +/- 3.80, N = 3SE +/- 5.97, N = 33405.993405.323408.16MIN: 3393.27MIN: 3391.39MIN: 3391.461. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

Sysbench

This is a benchmark of Sysbench with the built-in CPU and memory sub-tests. Sysbench is a scriptable multi-threaded benchmark tool based on LuaJIT. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgEvents Per Second, More Is BetterSysbench 1.0.20Test: CPU3214K8K12K16K20KSE +/- 16.54, N = 3SE +/- 20.02, N = 3SE +/- 14.24, N = 319367.3019382.7119380.341. (CC) gcc options: -pthread -O2 -funroll-loops -rdynamic -ldl -laio -lm

Basis Universal

Basis Universal is a GPU texture codec. This test times how long it takes to convert sRGB PNGs into Basis Univeral assets with various settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBasis Universal 1.13Settings: UASTC Level 33211428425670SE +/- 0.02, N = 3SE +/- 0.04, N = 3SE +/- 0.03, N = 363.5963.5663.541. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread

oneDNN

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

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPU321400800120016002000SE +/- 1.14, N = 3SE +/- 0.16, N = 3SE +/- 0.58, N = 31851.301850.821852.26MIN: 1843.57MIN: 1844.82MIN: 1846.851. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

toyBrot Fractal Generator

ToyBrot is a Mandelbrot fractal generator supporting C++ threads/tasks, OpenMP, Intel Threaded Building Blocks (TBB), and other targets. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BettertoyBrot Fractal Generator 2020-11-18Implementation: OpenMP3219K18K27K36K45KSE +/- 19.34, N = 3SE +/- 16.82, N = 3SE +/- 13.86, N = 34383843854438661. (CXX) g++ options: -O3 -lpthread

Liquid-DSP

LiquidSDR's Liquid-DSP is a software-defined radio (SDR) digital signal processing library. This test profile runs a multi-threaded benchmark of this SDR/DSP library focused on embedded platform usage. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgsamples/s, More Is BetterLiquid-DSP 2021.01.31Threads: 16 - Buffer Length: 256 - Filter Length: 57321110M220M330M440M550MSE +/- 199360.09, N = 3SE +/- 188355.46, N = 3SE +/- 324054.18, N = 35000166675002633335003333331. (CC) gcc options: -O3 -pthread -lm -lc -lliquid

Botan

Botan is a BSD-licensed cross-platform open-source C++ crypto library "cryptography toolkit" that supports most publicly known cryptographic algorithms. The project's stated goal is to be "the best option for cryptography in C++ by offering the tools necessary to implement a range of practical systems, such as TLS protocol, X.509 certificates, modern AEAD ciphers, PKCS#11 and TPM hardware support, password hashing, and post quantum crypto schemes." Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterBotan 2.17.3Test: Blowfish - Decrypt321120240360480600SE +/- 0.03, N = 3SE +/- 0.37, N = 3SE +/- 0.04, N = 3539.34539.02539.241. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt

OpenBenchmarking.orgMiB/s, More Is BetterBotan 2.17.3Test: ChaCha20Poly1305 - Decrypt3212004006008001000SE +/- 0.54, N = 3SE +/- 0.82, N = 3SE +/- 0.56, N = 3908.47907.95908.311. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: GEMM SGEMM_N32150100150200250SE +/- 0.06, N = 3SE +/- 0.06, N = 3SE +/- 0.06, N = 3245.78245.89245.921. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -pthread -lmpi_cxx -lmpi

oneDNN

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

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU321400800120016002000SE +/- 0.20, N = 3SE +/- 1.72, N = 3SE +/- 0.88, N = 31852.051851.141852.13MIN: 1846.88MIN: 1843.95MIN: 1846.261. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPU3210.48260.96521.44781.93042.413SE +/- 0.00152, N = 3SE +/- 0.00224, N = 3SE +/- 0.00353, N = 32.144522.143852.14490MIN: 1.97MIN: 1.97MIN: 1.941. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

Liquid-DSP

LiquidSDR's Liquid-DSP is a software-defined radio (SDR) digital signal processing library. This test profile runs a multi-threaded benchmark of this SDR/DSP library focused on embedded platform usage. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgsamples/s, More Is BetterLiquid-DSP 2021.01.31Threads: 1 - Buffer Length: 256 - Filter Length: 5732114M28M42M56M70MSE +/- 27834.83, N = 3SE +/- 1763.83, N = 3SE +/- 881.92, N = 36759966767632333676296671. (CC) gcc options: -O3 -pthread -lm -lc -lliquid

oneDNN

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

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.1.2Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPU3217001400210028003500SE +/- 0.89, N = 3SE +/- 0.58, N = 3SE +/- 3.64, N = 33413.883413.023414.52MIN: 3407.32MIN: 3406.86MIN: 3404.151. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

Botan

Botan is a BSD-licensed cross-platform open-source C++ crypto library "cryptography toolkit" that supports most publicly known cryptographic algorithms. The project's stated goal is to be "the best option for cryptography in C++ by offering the tools necessary to implement a range of practical systems, such as TLS protocol, X.509 certificates, modern AEAD ciphers, PKCS#11 and TPM hardware support, password hashing, and post quantum crypto schemes." Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterBotan 2.17.3Test: CAST-256 - Decrypt3214080120160200SE +/- 0.03, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 3172.15172.08172.091. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: Texture Read Bandwidth3211326395265SE +/- 0.00, N = 3SE +/- 0.01, N = 3SE +/- 0.02, N = 356.9256.8956.911. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -pthread -lmpi_cxx -lmpi

Botan

Botan is a BSD-licensed cross-platform open-source C++ crypto library "cryptography toolkit" that supports most publicly known cryptographic algorithms. The project's stated goal is to be "the best option for cryptography in C++ by offering the tools necessary to implement a range of practical systems, such as TLS protocol, X.509 certificates, modern AEAD ciphers, PKCS#11 and TPM hardware support, password hashing, and post quantum crypto schemes." Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterBotan 2.17.3Test: KASUMI321306090120150SE +/- 0.05, N = 3SE +/- 0.05, N = 3SE +/- 0.05, N = 3112.99112.94112.981. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: Max SP Flops321400800120016002000SE +/- 0.83, N = 3SE +/- 0.55, N = 3SE +/- 1.38, N = 31760.611761.071760.701. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -pthread -lmpi_cxx -lmpi

OpenBenchmarking.orgGHash/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: MD5 Hash3210.08690.17380.26070.34760.4345SE +/- 0.0000, N = 3SE +/- 0.0000, N = 3SE +/- 0.0000, N = 30.38600.38610.38611. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -pthread -lmpi_cxx -lmpi

Botan

Botan is a BSD-licensed cross-platform open-source C++ crypto library "cryptography toolkit" that supports most publicly known cryptographic algorithms. The project's stated goal is to be "the best option for cryptography in C++ by offering the tools necessary to implement a range of practical systems, such as TLS protocol, X.509 certificates, modern AEAD ciphers, PKCS#11 and TPM hardware support, password hashing, and post quantum crypto schemes." Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMiB/s, More Is BetterBotan 2.17.3Test: AES-25632110002000300040005000SE +/- 0.22, N = 3SE +/- 0.54, N = 3SE +/- 0.17, N = 34815.624814.834815.971. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt

OpenBenchmarking.orgMiB/s, More Is BetterBotan 2.17.3Test: CAST-2563214080120160200SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 3171.96171.95171.951. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt

Systemd Total Boot Time

This test uses systemd-analyze to report the entire boot time. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterSystemd Total Boot TimeTest: Userspace3215K10K15K20K25K221092210922109

OpenBenchmarking.orgms, Fewer Is BetterSystemd Total Boot TimeTest: Firmware3213K6K9K12K15K156791567915679

OpenBenchmarking.orgms, Fewer Is BetterSystemd Total Boot TimeTest: Loader3218001600240032004000355535553555

OpenBenchmarking.orgms, Fewer Is BetterSystemd Total Boot TimeTest: Kernel321400800120016002000194119411941

OpenBenchmarking.orgms, Fewer Is BetterSystemd Total Boot TimeTest: Total3215K10K15K20K25K240502405024050

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPs/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dGEMM-TT32148121620SE +/- 0.00, N = 2SE +/- 0.00, N = 2SE +/- 0.00, N = 215.715.715.71. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

OpenBenchmarking.orgGFLOPs/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dGEMM-TN32148121620SE +/- 0.03, N = 3SE +/- 0.03, N = 3SE +/- 0.03, N = 315.815.815.81. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

OpenBenchmarking.orgGFLOPs/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dGEMM-NN321510152025SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 318.918.918.91. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dDOT3219182736453838381. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - sAXPY321816243240SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 334.634.634.61. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

OpenBenchmarking.orgGFLOPs/s, More Is BetterViennaCL 1.7.1Test: CPU BLAS - dGEMM-TT321816243240SE +/- 0.03, N = 3SE +/- 0.03, N = 3SE +/- 0.03, N = 335.535.535.51. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

OpenBenchmarking.orgGFLOPs/s, More Is BetterViennaCL 1.7.1Test: CPU BLAS - dGEMM-NN321816243240SE +/- 0.03, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 334.834.834.81. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: CPU BLAS - dGEMV-T3211122334455SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.03, N = 346.546.546.51. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: CPU BLAS - dDOT3211020304050SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.03, N = 344.144.144.11. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: CPU BLAS - dAXPY321918273645SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 339.339.339.31. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: CPU BLAS - dCOPY321612182430SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.03, N = 326.226.226.21. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: CPU BLAS - sDOT3211122334455SE +/- 0.03, N = 3SE +/- 0.03, N = 3SE +/- 0.06, N = 347.947.947.91. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

AOM AV1

This is a test of the AOMedia AV1 encoder (libaom) developed by AOMedia and Google. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterAOM AV1 3.0Encoder Mode: Speed 0 Two-Pass - Input: Bosphorus 4K3210.0270.0540.0810.1080.135SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 30.120.120.121. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread

simdjson

This is a benchmark of SIMDJSON, a high performance JSON parser. SIMDJSON aims to be the fastest JSON parser and is used by projects like Microsoft FishStore, Yandex ClickHouse, Shopify, and others. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is Bettersimdjson 0.8.2Throughput Test: LargeRandom3210.24080.48160.72240.96321.204SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 31.071.071.071. (CXX) g++ options: -O3 -pthread

162 Results Shown

Zstd Compression
AOM AV1:
  Speed 8 Realtime - Bosphorus 4K
  Speed 6 Two-Pass - Bosphorus 1080p
  Speed 0 Two-Pass - Bosphorus 1080p
SHOC Scalable HeterOgeneous Computing
OpenSCAD
dav1d
AOM AV1:
  Speed 9 Realtime - Bosphorus 4K
  Speed 8 Realtime - Bosphorus 1080p
Zstd Compression:
  3, Long Mode - Compression Speed
  8 - Compression Speed
srsLTE
Stockfish
oneDNN
Zstd Compression
Mobile Neural Network
ViennaCL
AOM AV1
Mobile Neural Network
Basis Universal
oneDNN
srsLTE:
  PHY_DL_Test:
    eNb Mb/s
    UE Mb/s
Mobile Neural Network
oneDNN
AOM AV1
LuaRadio
Zstd Compression
Mobile Neural Network
SVT-VP9:
  VMAF Optimized - Bosphorus 1080p
  PSNR/SSIM Optimized - Bosphorus 1080p
SHOC Scalable HeterOgeneous Computing
OpenSCAD
libavif avifenc
AOM AV1
Mobile Neural Network
OpenSCAD
Zstd Compression
ViennaCL:
  OpenCL BLAS - dGEMM-NT
  OpenCL BLAS - dCOPY
Zstd Compression
GNU Radio
LuaRadio
Sysbench
Zstd Compression
ViennaCL
dav1d
GNU GMP GMPbench
SVT-VP9
libavif avifenc
GNU Radio
AOM AV1
oneDNN
Liquid-DSP
LuaRadio
oneDNN
LuaRadio
Timed Erlang/OTP Compilation
libavif avifenc
AOM AV1
simdjson
oneDNN
OpenSCAD
libavif avifenc
SVT-HEVC:
  10 - Bosphorus 1080p
  7 - Bosphorus 1080p
Basis Universal
ASTC Encoder
SHOC Scalable HeterOgeneous Computing
GNU Radio
Zstd Compression
ViennaCL
Botan:
  Twofish
  ChaCha20Poly1305
oneDNN
simdjson
dav1d
AOM AV1
Xcompact3d Incompact3d
ViennaCL
Zstd Compression
ViennaCL
Liquid-DSP
ViennaCL
libavif avifenc
ViennaCL
GNU Radio
ViennaCL
Basis Universal
dav1d
toyBrot Fractal Generator
oneDNN
ViennaCL
simdjson
Liquid-DSP
ViennaCL
Timed Linux Kernel Compilation
Zstd Compression
OpenSCAD
oneDNN
libavif avifenc
oneDNN
SHOC Scalable HeterOgeneous Computing
GNU Radio
toyBrot Fractal Generator:
  C++ Threads
  TBB
SHOC Scalable HeterOgeneous Computing
oneDNN
SVT-HEVC
oneDNN
Botan:
  AES-256 - Decrypt
  Twofish - Decrypt
SHOC Scalable HeterOgeneous Computing
GNU Radio
oneDNN
Xcompact3d Incompact3d
Zstd Compression
Timed Mesa Compilation
Botan
ASTC Encoder
Botan
Timed Node.js Compilation
ASTC Encoder
oneDNN
Sysbench
Basis Universal
oneDNN
toyBrot Fractal Generator
Liquid-DSP
Botan:
  Blowfish - Decrypt
  ChaCha20Poly1305 - Decrypt
SHOC Scalable HeterOgeneous Computing
oneDNN:
  Recurrent Neural Network Inference - f32 - CPU
  Deconvolution Batch shapes_1d - u8s8f32 - CPU
Liquid-DSP
oneDNN
Botan
SHOC Scalable HeterOgeneous Computing
Botan
SHOC Scalable HeterOgeneous Computing:
  OpenCL - Max SP Flops
  OpenCL - MD5 Hash
Botan:
  AES-256
  CAST-256
Systemd Total Boot Time:
  Userspace
  Firmware
  Loader
  Kernel
  Total
ViennaCL:
  OpenCL BLAS - dGEMM-TT
  OpenCL BLAS - dGEMM-TN
  OpenCL BLAS - dGEMM-NN
  OpenCL BLAS - dDOT
  OpenCL BLAS - sAXPY
  CPU BLAS - dGEMM-TT
  CPU BLAS - dGEMM-NN
  CPU BLAS - dGEMV-T
  CPU BLAS - dDOT
  CPU BLAS - dAXPY
  CPU BLAS - dCOPY
  CPU BLAS - sDOT
AOM AV1
simdjson