Core i3 8100 December

Intel Core i3-8100 testing with a ASRock Z370M-ITX/ac (P4.10 BIOS) and Intel 8th Gen Core Gaussian Mixture Model 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 2012270-HA-COREI381005
Jump To Table - Results

View

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

Limit displaying results to tests within:

Audio Encoding 3 Tests
Bioinformatics 2 Tests
Chess Test Suite 2 Tests
Timed Code Compilation 3 Tests
C/C++ Compiler Tests 6 Tests
CPU Massive 7 Tests
Creator Workloads 9 Tests
Encoding 4 Tests
Game Development 2 Tests
HPC - High Performance Computing 5 Tests
Machine Learning 3 Tests
Multi-Core 9 Tests
NVIDIA GPU Compute 4 Tests
Intel oneAPI 2 Tests
Programmer / Developer System Benchmarks 6 Tests
Scientific Computing 2 Tests
Server 3 Tests
Server CPU Tests 3 Tests
Texture Compression 2 Tests
Vulkan Compute 4 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
December 26 2020
  7 Hours, 53 Minutes
2
December 26 2020
  8 Hours, 11 Minutes
3
December 27 2020
  7 Hours, 49 Minutes
Invert Hiding All Results Option
  7 Hours, 58 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):


Core i3 8100 DecemberProcessorMotherboardChipsetMemoryDiskGraphicsAudioMonitorNetworkOSKernelDesktopDisplay ServerDisplay DriverOpenGLVulkanCompilerFile-SystemScreen Resolution123Intel Core i3-8100 @ 3.60GHz (4 Cores)ASRock Z370M-ITX/ac (P4.10 BIOS)Intel 8th Gen Core 4-core Desktop8GB60GB DREVO X1 SSDIntel 8th Gen Core Gaussian Mixture Model 3GB (1100MHz)Realtek ALC892VA2431Intel I219-V + Intel I211 + Intel Dual Band-AC 3168NGWUbuntu 20.045.9.0-050900rc1daily20200819-generic (x86_64) 20200818GNOME Shell 3.36.4X Server 1.20.8modesetting 1.20.84.6 Mesa 20.0.81.2.131GCC 9.3.0ext41920x1080OpenBenchmarking.orgCompiler 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=/build/gcc-9-HskZEa/gcc-9-9.3.0/debian/tmp-nvptx/usr,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: 0xde - Thermald 1.9.1 Python Details- Python 3.8.5Security Details- itlb_multihit: KVM: Mitigation of VMX disabled + l1tf: Mitigation of PTE Inversion; VMX: conditional cache flushes SMT disabled + mds: Mitigation of Clear buffers; SMT disabled + meltdown: Mitigation of PTI + 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 Full generic retpoline IBPB: conditional IBRS_FW STIBP: disabled RSB filling + srbds: Mitigation of Microcode + tsx_async_abort: Not affected

123Result OverviewPhoronix Test Suite100%102%105%107%CoremarkBetsy GPU CompressorVkResampleNode.js V8 Web Tooling BenchmarkUnpacking FirefoxStockfishasmFishSQLite SpeedtestBRL-CADTimed MAFFT AlignmentVKMarkMonkey Audio EncodingTimed Eigen CompilationoneDNNsimdjsonBuild2OpenVINOrav1eTimed FFmpeg CompilationVkFFTWavPack Audio EncodingTimed HMMer SearchLZ4 CompressionASTC EncoderNCNNOpus Codec EncodingCLOMP

Core i3 8100 Decembervkresample: 2x - Singleonednn: IP Shapes 3D - f32 - CPUnode-web-tooling: openvino: Age Gender Recognition Retail 0013 FP16 - CPUunpack-firefox: firefox-84.0.source.tar.xzsimdjson: Kostyastockfish: Total Timeopenvino: Person Detection 0106 FP32 - CPUvkmark: 1280 x 1024onednn: IP Shapes 3D - u8s8f32 - CPUasmfish: 1024 Hash Memory, 26 Depthonednn: Convolution Batch Shapes Auto - f32 - CPUsqlite-speedtest: Timed Time - Size 1,000brl-cad: VGR Performance Metricopenvino: Face Detection 0106 FP16 - CPUcompress-lz4: 9 - Compression Speedonednn: Deconvolution Batch shapes_3d - f32 - CPUncnn: CPU-v2-v2 - mobilenet-v2openvino: Age Gender Recognition Retail 0013 FP32 - CPUopenvino: Age Gender Recognition Retail 0013 FP16 - CPUonednn: Recurrent Neural Network Training - bf16bf16bf16 - CPUmafft: Multiple Sequence Alignment - LSU RNAncnn: CPU - mnasnetencode-ape: WAV To APEonednn: Convolution Batch Shapes Auto - u8s8f32 - CPUbuild-eigen: Time To Compileonednn: Deconvolution Batch shapes_1d - f32 - CPUbuild2: Time To Compileonednn: Recurrent Neural Network Inference - f32 - CPUncnn: CPU - regnety_400mopenvino: Age Gender Recognition Retail 0013 FP32 - CPUvkresample: 2x - Doublencnn: Vulkan GPU - shufflenet-v2ncnn: Vulkan GPU - resnet50compress-lz4: 1 - Compression Speedrav1e: 1ncnn: CPU - efficientnet-b0ncnn: Vulkan GPU-v2-v2 - mobilenet-v2ncnn: CPU - googlenetcompress-lz4: 9 - Decompression Speedncnn: CPU - shufflenet-v2openvino: Person Detection 0106 FP32 - CPUncnn: Vulkan GPU - regnety_400mncnn: Vulkan GPU - squeezenet_ssdncnn: Vulkan GPU - efficientnet-b0ncnn: Vulkan GPU - googlenetastcenc: Fastcompress-lz4: 1 - Decompression Speedonednn: IP Shapes 1D - u8s8f32 - CPUonednn: IP Shapes 1D - f32 - CPUonednn: Recurrent Neural Network Inference - u8s8f32 - CPUonednn: Matrix Multiply Batch Shapes Transformer - f32 - CPUncnn: CPU-v3-v3 - mobilenet-v3ncnn: Vulkan GPU-v3-v3 - mobilenet-v3ncnn: Vulkan GPU - alexnetncnn: Vulkan GPU - mnasnetvkmark: 1920 x 1080ncnn: CPU - squeezenet_ssdncnn: CPU - resnet50vkfft: onednn: Recurrent Neural Network Training - u8s8f32 - CPUencode-wavpack: WAV To WavPackcompress-lz4: 3 - Compression Speedonednn: Recurrent Neural Network Inference - bf16bf16bf16 - CPUonednn: Deconvolution Batch shapes_1d - u8s8f32 - CPUncnn: Vulkan GPU - mobilenetrav1e: 10ncnn: CPU - alexnetrav1e: 5build-ffmpeg: Time To Compileonednn: Deconvolution Batch shapes_3d - u8s8f32 - CPUopenvino: Face Detection 0106 FP16 - CPUastcenc: Exhaustivencnn: Vulkan GPU - yolov4-tinyastcenc: Thoroughcompress-lz4: 3 - Decompression Speedncnn: CPU - mobilenetopenvino: Person Detection 0106 FP16 - CPUastcenc: Mediumonednn: Matrix Multiply Batch Shapes Transformer - u8s8f32 - CPUncnn: Vulkan GPU - vgg16ncnn: CPU - yolov4-tinyncnn: Vulkan GPU - resnet18ncnn: CPU - resnet18encode-opus: WAV To Opus Encodehmmer: Pfam Database Searchncnn: CPU - vgg16onednn: Recurrent Neural Network Training - f32 - CPUopenvino: Face Detection 0106 FP32 - CPUopenvino: Person Detection 0106 FP16 - CPUopenvino: Face Detection 0106 FP32 - CPUncnn: Vulkan GPU - blazefacencnn: CPU - blazefacerav1e: 6simdjson: DistinctUserIDsimdjson: PartialTweetssimdjson: LargeRandclomp: Static OMP Speedupbetsy: ETC2 RGB - Highestcoremark: CoreMark Size 666 - Iterations Per Secondbetsy: ETC1 - Highest123478.8339.0614210.172549.4923.2560.5764565850.709773.59647898802118.534679.432365801.1640.0814.46107.181.461.456758.8411.9396.5513.62417.708693.50811.5173338.8643873.7514.552543.511013.8279.3647.847608.280.34410.397.1621.699077.69.345648.2014.5540.5210.3721.685.769201.55.588768.470603855.794.711866.246.2619.286.5465740.4047.8313676758.2517.96540.973856.2014.554828.162.89019.320.984151.63011.24543426.15738.3738.9792.929058.028.155641.2814.287.5011487.2838.9922.5222.5510.648126.15787.356759.553436.650.711.152.362.361.3080.540.520.351.77.54198117.50907510.906422.2279.5288410.392516.5123.3930.5865519820.719693.63768900838018.712579.660362471.1739.7614.35837.141.451.456804.3411.9226.5113.70417.808593.15011.4995337.9883860.5214.502551.411010.4969.3747.797630.560.34410.367.1521.759053.29.345638.5914.5340.4610.3721.725.759212.45.586718.479883858.994.719316.256.2619.316.5465840.4647.9013696759.2517.94241.023860.1514.539028.142.88719.340.984151.67811.23403422.79737.7938.9692.889051.328.155638.3214.297.5063987.2438.9722.5222.5610.648126.19887.326761.683436.560.711.152.362.361.3080.540.520.351.7100468.34611911.082421.6359.1888910.292562.1723.6690.5864517270.709643.59274891245818.543078.924365691.1639.9814.41457.131.451.466764.3311.9976.5313.64317.760193.09011.4730337.5763874.2314.552552.271011.0059.3447.937612.810.34310.377.1421.749066.99.365650.5614.5640.4410.3921.695.759216.95.579588.466033852.804.719436.246.2519.306.5365840.4547.8813686767.3917.94540.993855.4914.540228.132.88919.330.983151.78411.23433426.05738.4438.9492.959057.328.175642.3214.297.5048287.2338.9722.5322.5610.652126.15487.346761.793437.370.711.152.362.361.3080.540.520.351.792000.14054310.370OpenBenchmarking.org

VkResample

VkResample is a Vulkan-based image upscaling library based on VkFFT. The sample input file is upscaling a 4K image to 8K using Vulkan-based GPU acceleration. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterVkResample 1.0Upscale: 2x - Precision: Single123100200300400500SE +/- 0.46, N = 3SE +/- 0.41, N = 3SE +/- 0.49, N = 3478.83422.23421.641. (CXX) g++ options: -O3 -pthread
OpenBenchmarking.orgms, Fewer Is BetterVkResample 1.0Upscale: 2x - Precision: Single12380160240320400Min: 478.16 / Avg: 478.83 / Max: 479.72Min: 421.71 / Avg: 422.23 / Max: 423.03Min: 421.01 / Avg: 421.63 / Max: 422.591. (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 2.0Harness: IP Shapes 3D - Data Type: f32 - Engine: CPU1233691215SE +/- 0.01406, N = 3SE +/- 0.02685, N = 3SE +/- 0.02174, N = 39.061429.528849.18889MIN: 8.95MIN: 9.43MIN: 9.11. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: IP Shapes 3D - Data Type: f32 - Engine: CPU1233691215Min: 9.04 / Avg: 9.06 / Max: 9.09Min: 9.48 / Avg: 9.53 / Max: 9.56Min: 9.15 / Avg: 9.19 / Max: 9.221. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

Node.js V8 Web Tooling Benchmark

Running the V8 project's Web-Tooling-Benchmark under Node.js. The Web-Tooling-Benchmark stresses JavaScript-related workloads common to web developers like Babel and TypeScript and Babylon. This test profile can test the system's JavaScript performance with Node.js. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgruns/s, More Is BetterNode.js V8 Web Tooling Benchmark1233691215SE +/- 0.06, N = 3SE +/- 0.00, N = 3SE +/- 0.06, N = 310.1710.3910.291. Nodejs v10.19.0
OpenBenchmarking.orgruns/s, More Is BetterNode.js V8 Web Tooling Benchmark1233691215Min: 10.05 / Avg: 10.17 / Max: 10.23Min: 10.38 / Avg: 10.39 / Max: 10.39Min: 10.18 / Avg: 10.29 / Max: 10.381. Nodejs v10.19.0

OpenVINO

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

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2021.1Model: Age Gender Recognition Retail 0013 FP16 - Device: CPU1235001000150020002500SE +/- 10.55, N = 3SE +/- 32.43, N = 12SE +/- 18.26, N = 32549.492516.512562.17
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2021.1Model: Age Gender Recognition Retail 0013 FP16 - Device: CPU123400800120016002000Min: 2533.93 / Avg: 2549.49 / Max: 2569.6Min: 2161.99 / Avg: 2516.51 / Max: 2575.89Min: 2534.89 / Avg: 2562.17 / Max: 2596.83

Unpacking Firefox

This simple test profile measures how long it takes to extract the .tar.xz source package of the Mozilla Firefox Web Browser. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterUnpacking Firefox 84.0Extracting: firefox-84.0.source.tar.xz123612182430SE +/- 0.18, N = 4SE +/- 0.24, N = 4SE +/- 0.30, N = 423.2623.3923.67
OpenBenchmarking.orgSeconds, Fewer Is BetterUnpacking Firefox 84.0Extracting: firefox-84.0.source.tar.xz123612182430Min: 23.02 / Avg: 23.26 / Max: 23.78Min: 23.02 / Avg: 23.39 / Max: 24.06Min: 23.01 / Avg: 23.67 / Max: 24.43

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 0.7.1Throughput Test: Kostya1230.13050.2610.39150.5220.6525SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 30.570.580.581. (CXX) g++ options: -O3 -pthread
OpenBenchmarking.orgGB/s, More Is Bettersimdjson 0.7.1Throughput Test: Kostya123246810Min: 0.57 / Avg: 0.57 / Max: 0.58Min: 0.57 / Avg: 0.58 / Max: 0.58Min: 0.57 / Avg: 0.58 / Max: 0.581. (CXX) g++ options: -O3 -pthread

Stockfish

This is a test of Stockfish, an advanced C++11 chess benchmark that can scale up to 128 CPU cores. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgNodes Per Second, More Is BetterStockfish 12Total Time1231.4M2.8M4.2M5.6M7MSE +/- 59897.76, N = 3SE +/- 39431.93, N = 3SE +/- 73617.81, N = 36456585655198264517271. (CXX) g++ options: -m64 -lpthread -fno-exceptions -std=c++17 -pedantic -O3 -msse -msse3 -mpopcnt -msse4.1 -mssse3 -msse2 -flto -flto=jobserver
OpenBenchmarking.orgNodes Per Second, More Is BetterStockfish 12Total Time1231.1M2.2M3.3M4.4M5.5MMin: 6350666 / Avg: 6456585 / Max: 6558012Min: 6500126 / Avg: 6551981.67 / Max: 6629367Min: 6312827 / Avg: 6451726.67 / Max: 65634721. (CXX) g++ options: -m64 -lpthread -fno-exceptions -std=c++17 -pedantic -O3 -msse -msse3 -mpopcnt -msse4.1 -mssse3 -msse2 -flto -flto=jobserver

OpenVINO

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

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2021.1Model: Person Detection 0106 FP32 - Device: CPU1230.15980.31960.47940.63920.799SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 30.700.710.70
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2021.1Model: Person Detection 0106 FP32 - Device: CPU123246810Min: 0.7 / Avg: 0.7 / Max: 0.71Min: 0.7 / Avg: 0.71 / Max: 0.71Min: 0.7 / Avg: 0.7 / Max: 0.71

VKMark

VKMark is a collection of Vulkan tests/benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgVKMark Score, More Is BetterVKMark 2020-05-21Resolution: 1280 x 10241232004006008001000SE +/- 1.00, N = 3SE +/- 3.93, N = 39779699641. (CXX) g++ options: -pthread -ldl -pipe -std=c++14 -MD -MQ -MF
OpenBenchmarking.orgVKMark Score, More Is BetterVKMark 2020-05-21Resolution: 1280 x 10241232004006008001000Min: 968 / Avg: 969 / Max: 971Min: 959 / Avg: 964.33 / Max: 9721. (CXX) g++ options: -pthread -ldl -pipe -std=c++14 -MD -MQ -MF

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 2.0Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPU1230.81851.6372.45553.2744.0925SE +/- 0.01043, N = 3SE +/- 0.03180, N = 3SE +/- 0.01609, N = 33.596473.637683.59274MIN: 3.39MIN: 3.18MIN: 3.341. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPU123246810Min: 3.58 / Avg: 3.6 / Max: 3.61Min: 3.59 / Avg: 3.64 / Max: 3.7Min: 3.57 / Avg: 3.59 / Max: 3.621. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

asmFish

This is a test of asmFish, an advanced chess benchmark written in Assembly. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgNodes/second, More Is BetterasmFish 2018-07-231024 Hash Memory, 26 Depth1232M4M6M8M10MSE +/- 98059.95, N = 3SE +/- 71523.89, N = 3SE +/- 46897.59, N = 3898802190083808912458
OpenBenchmarking.orgNodes/second, More Is BetterasmFish 2018-07-231024 Hash Memory, 26 Depth1231.6M3.2M4.8M6.4M8MMin: 8802963 / Avg: 8988021.33 / Max: 9136785Min: 8910895 / Avg: 9008380 / Max: 9147784Min: 8841446 / Avg: 8912457.67 / Max: 9001031

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 2.0Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU123510152025SE +/- 0.02, N = 3SE +/- 0.03, N = 3SE +/- 0.02, N = 318.5318.7118.54MIN: 18.42MIN: 18.57MIN: 18.431. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU123510152025Min: 18.51 / Avg: 18.53 / Max: 18.58Min: 18.67 / Avg: 18.71 / Max: 18.77Min: 18.5 / Avg: 18.54 / Max: 18.581. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

SQLite Speedtest

This is a benchmark of SQLite's speedtest1 benchmark program with an increased problem size of 1,000. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterSQLite Speedtest 3.30Timed Time - Size 1,00012320406080100SE +/- 0.03, N = 3SE +/- 0.34, N = 3SE +/- 0.10, N = 379.4379.6678.921. (CC) gcc options: -O2 -ldl -lz -lpthread
OpenBenchmarking.orgSeconds, Fewer Is BetterSQLite Speedtest 3.30Timed Time - Size 1,0001231530456075Min: 79.39 / Avg: 79.43 / Max: 79.5Min: 79.07 / Avg: 79.66 / Max: 80.26Min: 78.77 / Avg: 78.92 / Max: 79.11. (CC) gcc options: -O2 -ldl -lz -lpthread

BRL-CAD

BRL-CAD 7.28.0 is a cross-platform, open-source solid modeling system with built-in benchmark mode. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgVGR Performance Metric, More Is BetterBRL-CAD 7.30.8VGR Performance Metric1238K16K24K32K40K3658036247365691. (CXX) g++ options: -std=c++11 -pipe -fno-strict-aliasing -fno-common -fexceptions -ftemplate-depth-128 -m64 -ggdb3 -O3 -fipa-pta -fstrength-reduce -finline-functions -flto -pedantic -rdynamic -lSM -lICE -lXi -lGLU -lGL -lGLdispatch -lX11 -lXext -lXrender -lpthread -ldl -luuid -lm

OpenVINO

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

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2021.1Model: Face Detection 0106 FP16 - Device: CPU1230.26330.52660.78991.05321.3165SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 31.161.171.16
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2021.1Model: Face Detection 0106 FP16 - Device: CPU123246810Min: 1.16 / Avg: 1.16 / Max: 1.17Min: 1.16 / Avg: 1.17 / Max: 1.17Min: 1.16 / Avg: 1.16 / Max: 1.17

LZ4 Compression

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

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 9 - Compression Speed123918273645SE +/- 0.06, N = 3SE +/- 0.21, N = 3SE +/- 0.01, N = 340.0839.7639.981. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 9 - Compression Speed123816243240Min: 39.96 / Avg: 40.08 / Max: 40.14Min: 39.35 / Avg: 39.76 / Max: 39.97Min: 39.97 / Avg: 39.98 / Max: 39.991. (CC) gcc options: -O3

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 2.0Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPU12348121620SE +/- 0.08, N = 3SE +/- 0.01, N = 3SE +/- 0.04, N = 314.4614.3614.41MIN: 14.22MIN: 14.23MIN: 14.251. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPU12348121620Min: 14.32 / Avg: 14.46 / Max: 14.59Min: 14.35 / Avg: 14.36 / Max: 14.37Min: 14.35 / Avg: 14.41 / Max: 14.481. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU-v2-v2 - Model: mobilenet-v2123246810SE +/- 0.05, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 37.187.147.13MIN: 7.07 / MAX: 29.75MIN: 7.07 / MAX: 8.36MIN: 7.07 / MAX: 8.561. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU-v2-v2 - Model: mobilenet-v21233691215Min: 7.12 / Avg: 7.18 / Max: 7.28Min: 7.13 / Avg: 7.14 / Max: 7.16Min: 7.12 / Avg: 7.13 / Max: 7.141. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenVINO

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

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2021.1Model: Age Gender Recognition Retail 0013 FP32 - Device: CPU1230.32850.6570.98551.3141.6425SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 31.461.451.45
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2021.1Model: Age Gender Recognition Retail 0013 FP32 - Device: CPU123246810Min: 1.46 / Avg: 1.46 / Max: 1.46Min: 1.45 / Avg: 1.45 / Max: 1.46Min: 1.45 / Avg: 1.45 / Max: 1.45

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2021.1Model: Age Gender Recognition Retail 0013 FP16 - Device: CPU1230.32850.6570.98551.3141.6425SE +/- 0.00, N = 3SE +/- 0.00, N = 12SE +/- 0.00, N = 31.451.451.46
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2021.1Model: Age Gender Recognition Retail 0013 FP16 - Device: CPU123246810Min: 1.45 / Avg: 1.45 / Max: 1.45Min: 1.45 / Avg: 1.45 / Max: 1.46Min: 1.46 / Avg: 1.46 / Max: 1.46

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 2.0Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPU12315003000450060007500SE +/- 7.71, N = 3SE +/- 37.00, N = 3SE +/- 4.92, N = 36758.846804.346764.33MIN: 6740.34MIN: 6761.06MIN: 6750.921. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPU12312002400360048006000Min: 6748.84 / Avg: 6758.84 / Max: 6774Min: 6763.88 / Avg: 6804.34 / Max: 6878.22Min: 6754.88 / Avg: 6764.33 / Max: 6771.451. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

Timed MAFFT Alignment

This test performs an alignment of 100 pyruvate decarboxylase sequences. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed MAFFT Alignment 7.471Multiple Sequence Alignment - LSU RNA1233691215SE +/- 0.04, N = 3SE +/- 0.04, N = 3SE +/- 0.07, N = 311.9411.9212.001. (CC) gcc options: -std=c99 -O3 -lm -lpthread
OpenBenchmarking.orgSeconds, Fewer Is BetterTimed MAFFT Alignment 7.471Multiple Sequence Alignment - LSU RNA1233691215Min: 11.86 / Avg: 11.94 / Max: 11.99Min: 11.88 / Avg: 11.92 / Max: 12Min: 11.86 / Avg: 12 / Max: 12.071. (CC) gcc options: -std=c99 -O3 -lm -lpthread

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: mnasnet123246810SE +/- 0.02, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 36.556.516.53MIN: 6.5 / MAX: 16.64MIN: 6.47 / MAX: 6.73MIN: 6.49 / MAX: 6.661. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: mnasnet1233691215Min: 6.53 / Avg: 6.55 / Max: 6.58Min: 6.5 / Avg: 6.51 / Max: 6.52Min: 6.52 / Avg: 6.53 / Max: 6.541. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

Monkey Audio Encoding

This test times how long it takes to encode a sample WAV file to Monkey's Audio APE format. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterMonkey Audio Encoding 3.99.6WAV To APE12348121620SE +/- 0.03, N = 5SE +/- 0.06, N = 5SE +/- 0.03, N = 513.6213.7013.641. (CXX) g++ options: -O3 -pedantic -rdynamic -lrt
OpenBenchmarking.orgSeconds, Fewer Is BetterMonkey Audio Encoding 3.99.6WAV To APE12348121620Min: 13.55 / Avg: 13.62 / Max: 13.74Min: 13.58 / Avg: 13.7 / Max: 13.9Min: 13.58 / Avg: 13.64 / Max: 13.731. (CXX) g++ options: -O3 -pedantic -rdynamic -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 2.0Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPU12348121620SE +/- 0.03, N = 3SE +/- 0.02, N = 3SE +/- 0.01, N = 317.7117.8117.76MIN: 17.57MIN: 17.67MIN: 17.611. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPU123510152025Min: 17.67 / Avg: 17.71 / Max: 17.77Min: 17.78 / Avg: 17.81 / Max: 17.84Min: 17.74 / Avg: 17.76 / Max: 17.771. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

Timed Eigen Compilation

This test times how long it takes to build all Eigen examples. The Eigen examples are compiled serially. Eigen is a C++ template library for linear algebra. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Eigen Compilation 3.3.9Time To Compile12320406080100SE +/- 0.06, N = 3SE +/- 0.07, N = 3SE +/- 0.06, N = 393.5193.1593.09
OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Eigen Compilation 3.3.9Time To Compile12320406080100Min: 93.39 / Avg: 93.51 / Max: 93.6Min: 93.05 / Avg: 93.15 / Max: 93.27Min: 93 / Avg: 93.09 / Max: 93.21

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 2.0Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPU1233691215SE +/- 0.03, N = 3SE +/- 0.03, N = 3SE +/- 0.01, N = 311.5211.5011.47MIN: 11.42MIN: 11.4MIN: 11.381. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPU1233691215Min: 11.48 / Avg: 11.52 / Max: 11.59Min: 11.46 / Avg: 11.5 / Max: 11.56Min: 11.45 / Avg: 11.47 / Max: 11.491. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

Build2

This test profile measures the time to bootstrap/install the build2 C++ build toolchain from source. Build2 is a cross-platform build toolchain for C/C++ code and features Cargo-like features. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBuild2 0.13Time To Compile12370140210280350SE +/- 0.66, N = 3SE +/- 0.28, N = 3SE +/- 0.64, N = 3338.86337.99337.58
OpenBenchmarking.orgSeconds, Fewer Is BetterBuild2 0.13Time To Compile12360120180240300Min: 338.17 / Avg: 338.86 / Max: 340.18Min: 337.43 / Avg: 337.99 / Max: 338.35Min: 336.38 / Avg: 337.58 / Max: 338.54

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 2.0Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU1238001600240032004000SE +/- 20.54, N = 3SE +/- 0.72, N = 3SE +/- 18.83, N = 33873.753860.523874.23MIN: 3848.44MIN: 3856.22MIN: 3849.241. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU1237001400210028003500Min: 3852.94 / Avg: 3873.75 / Max: 3914.83Min: 3859.09 / Avg: 3860.52 / Max: 3861.37Min: 3853.73 / Avg: 3874.23 / Max: 3911.851. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: regnety_400m12348121620SE +/- 0.03, N = 3SE +/- 0.03, N = 3SE +/- 0.02, N = 314.5514.5014.55MIN: 14.43 / MAX: 15.76MIN: 14.4 / MAX: 15.22MIN: 14.46 / MAX: 15.191. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: regnety_400m12348121620Min: 14.49 / Avg: 14.55 / Max: 14.61Min: 14.44 / Avg: 14.5 / Max: 14.55Min: 14.52 / Avg: 14.55 / Max: 14.581. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenVINO

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

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2021.1Model: Age Gender Recognition Retail 0013 FP32 - Device: CPU1235001000150020002500SE +/- 4.46, N = 3SE +/- 2.44, N = 3SE +/- 6.24, N = 32543.512551.412552.27
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2021.1Model: Age Gender Recognition Retail 0013 FP32 - Device: CPU123400800120016002000Min: 2536.6 / Avg: 2543.51 / Max: 2551.85Min: 2546.68 / Avg: 2551.41 / Max: 2554.79Min: 2544.16 / Avg: 2552.27 / Max: 2564.53

VkResample

VkResample is a Vulkan-based image upscaling library based on VkFFT. The sample input file is upscaling a 4K image to 8K using Vulkan-based GPU acceleration. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterVkResample 1.0Upscale: 2x - Precision: Double1232004006008001000SE +/- 1.41, N = 3SE +/- 2.85, N = 3SE +/- 4.21, N = 31013.831010.501011.011. (CXX) g++ options: -O3 -pthread
OpenBenchmarking.orgms, Fewer Is BetterVkResample 1.0Upscale: 2x - Precision: Double1232004006008001000Min: 1011.45 / Avg: 1013.83 / Max: 1016.32Min: 1005.21 / Avg: 1010.5 / Max: 1015.01Min: 1003.22 / Avg: 1011.01 / Max: 1017.691. (CXX) g++ options: -O3 -pthread

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: shufflenet-v21233691215SE +/- 0.01, N = 3SE +/- 0.02, N = 3SE +/- 0.01, N = 39.369.379.34MIN: 9.29 / MAX: 10.85MIN: 9.31 / MAX: 10.64MIN: 9.29 / MAX: 10.751. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: shufflenet-v21233691215Min: 9.35 / Avg: 9.36 / Max: 9.37Min: 9.33 / Avg: 9.37 / Max: 9.39Min: 9.31 / Avg: 9.34 / Max: 9.351. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: resnet501231122334455SE +/- 0.02, N = 3SE +/- 0.03, N = 3SE +/- 0.11, N = 347.8447.7947.93MIN: 47.59 / MAX: 59.98MIN: 47.59 / MAX: 57.75MIN: 47.56 / MAX: 120.581. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: resnet501231020304050Min: 47.8 / Avg: 47.84 / Max: 47.86Min: 47.74 / Avg: 47.79 / Max: 47.84Min: 47.77 / Avg: 47.93 / Max: 48.131. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

LZ4 Compression

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

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 1 - Compression Speed12316003200480064008000SE +/- 4.40, N = 3SE +/- 19.00, N = 3SE +/- 7.93, N = 37608.287630.567612.811. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 1 - Compression Speed12313002600390052006500Min: 7599.53 / Avg: 7608.28 / Max: 7613.44Min: 7595.51 / Avg: 7630.56 / Max: 7660.81Min: 7599.05 / Avg: 7612.81 / Max: 7626.511. (CC) gcc options: -O3

rav1e

Xiph rav1e is a Rust-written AV1 video encoder. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is Betterrav1e 0.4 AlphaSpeed: 11230.07740.15480.23220.30960.387SE +/- 0.000, N = 3SE +/- 0.001, N = 3SE +/- 0.001, N = 30.3440.3440.343
OpenBenchmarking.orgFrames Per Second, More Is Betterrav1e 0.4 AlphaSpeed: 112312345Min: 0.34 / Avg: 0.34 / Max: 0.34Min: 0.34 / Avg: 0.34 / Max: 0.35Min: 0.34 / Avg: 0.34 / Max: 0.34

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: efficientnet-b01233691215SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 310.3910.3610.37MIN: 10.35 / MAX: 11.79MIN: 10.33 / MAX: 10.6MIN: 10.32 / MAX: 10.491. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: efficientnet-b01233691215Min: 10.38 / Avg: 10.39 / Max: 10.39Min: 10.36 / Avg: 10.36 / Max: 10.36Min: 10.36 / Avg: 10.37 / Max: 10.381. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU-v2-v2 - Model: mobilenet-v2123246810SE +/- 0.01, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 37.167.157.14MIN: 7.08 / MAX: 8.7MIN: 7.07 / MAX: 9.43MIN: 7.08 / MAX: 8.351. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU-v2-v2 - Model: mobilenet-v21233691215Min: 7.15 / Avg: 7.16 / Max: 7.17Min: 7.15 / Avg: 7.15 / Max: 7.15Min: 7.14 / Avg: 7.14 / Max: 7.151. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: googlenet123510152025SE +/- 0.02, N = 3SE +/- 0.05, N = 3SE +/- 0.06, N = 321.6921.7521.74MIN: 21.59 / MAX: 29.99MIN: 21.6 / MAX: 56.79MIN: 21.62 / MAX: 52.651. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: googlenet123510152025Min: 21.66 / Avg: 21.69 / Max: 21.73Min: 21.7 / Avg: 21.75 / Max: 21.85Min: 21.68 / Avg: 21.74 / Max: 21.861. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

LZ4 Compression

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

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 9 - Decompression Speed1232K4K6K8K10KSE +/- 15.30, N = 3SE +/- 13.38, N = 3SE +/- 14.57, N = 39077.69053.29066.91. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 9 - Decompression Speed12316003200480064008000Min: 9058.7 / Avg: 9077.6 / Max: 9107.9Min: 9030.6 / Avg: 9053.17 / Max: 9076.9Min: 9043.9 / Avg: 9066.93 / Max: 9093.91. (CC) gcc options: -O3

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: shufflenet-v21233691215SE +/- 0.02, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 39.349.349.36MIN: 9.28 / MAX: 10.73MIN: 9.28 / MAX: 10.73MIN: 9.29 / MAX: 11.041. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: shufflenet-v21233691215Min: 9.31 / Avg: 9.34 / Max: 9.36Min: 9.32 / Avg: 9.34 / Max: 9.37Min: 9.33 / Avg: 9.36 / Max: 9.371. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenVINO

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

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2021.1Model: Person Detection 0106 FP32 - Device: CPU12312002400360048006000SE +/- 7.38, N = 3SE +/- 5.55, N = 3SE +/- 8.84, N = 35648.205638.595650.56
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2021.1Model: Person Detection 0106 FP32 - Device: CPU12310002000300040005000Min: 5635.96 / Avg: 5648.2 / Max: 5661.45Min: 5630.51 / Avg: 5638.59 / Max: 5649.23Min: 5638.9 / Avg: 5650.56 / Max: 5667.91

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: regnety_400m12348121620SE +/- 0.02, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 314.5514.5314.56MIN: 14.45 / MAX: 28.03MIN: 14.46 / MAX: 15.24MIN: 14.49 / MAX: 15.221. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: regnety_400m12348121620Min: 14.52 / Avg: 14.55 / Max: 14.57Min: 14.51 / Avg: 14.53 / Max: 14.55Min: 14.55 / Avg: 14.56 / Max: 14.581. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: squeezenet_ssd123918273645SE +/- 0.08, N = 3SE +/- 0.05, N = 3SE +/- 0.05, N = 340.5240.4640.44MIN: 40.36 / MAX: 90.92MIN: 40.33 / MAX: 67.8MIN: 40.32 / MAX: 50.61. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: squeezenet_ssd123816243240Min: 40.43 / Avg: 40.52 / Max: 40.68Min: 40.4 / Avg: 40.46 / Max: 40.55Min: 40.37 / Avg: 40.44 / Max: 40.541. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: efficientnet-b01233691215SE +/- 0.01, N = 3SE +/- 0.00, N = 3SE +/- 0.03, N = 310.3710.3710.39MIN: 10.32 / MAX: 11.49MIN: 10.34 / MAX: 11.79MIN: 10.32 / MAX: 20.271. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: efficientnet-b01233691215Min: 10.35 / Avg: 10.37 / Max: 10.39Min: 10.37 / Avg: 10.37 / Max: 10.37Min: 10.35 / Avg: 10.39 / Max: 10.441. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: googlenet123510152025SE +/- 0.02, N = 3SE +/- 0.03, N = 3SE +/- 0.02, N = 321.6821.7221.69MIN: 21.59 / MAX: 31.15MIN: 21.61 / MAX: 31.67MIN: 21.59 / MAX: 31.581. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: googlenet123510152025Min: 21.65 / Avg: 21.68 / Max: 21.73Min: 21.67 / Avg: 21.72 / Max: 21.77Min: 21.66 / Avg: 21.69 / Max: 21.721. (CXX) g++ options: -O3 -rdynamic -lgomp -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 2.0Preset: Fast1231.2962.5923.8885.1846.48SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 35.765.755.751. (CXX) g++ options: -std=c++14 -fvisibility=hidden -O3 -flto -mfpmath=sse -mavx2 -mpopcnt -lpthread
OpenBenchmarking.orgSeconds, Fewer Is BetterASTC Encoder 2.0Preset: Fast123246810Min: 5.74 / Avg: 5.76 / Max: 5.78Min: 5.74 / Avg: 5.75 / Max: 5.76Min: 5.74 / Avg: 5.75 / Max: 5.771. (CXX) g++ options: -std=c++14 -fvisibility=hidden -O3 -flto -mfpmath=sse -mavx2 -mpopcnt -lpthread

LZ4 Compression

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

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 1 - Decompression Speed1232K4K6K8K10KSE +/- 38.89, N = 3SE +/- 38.65, N = 3SE +/- 21.43, N = 39201.59212.49216.91. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 1 - Decompression Speed12316003200480064008000Min: 9128.6 / Avg: 9201.53 / Max: 9261.4Min: 9154.9 / Avg: 9212.4 / Max: 9285.9Min: 9178.6 / Avg: 9216.87 / Max: 9252.71. (CC) gcc options: -O3

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 2.0Harness: IP Shapes 1D - Data Type: u8s8f32 - Engine: CPU1231.25752.5153.77255.036.2875SE +/- 0.00294, N = 3SE +/- 0.00388, N = 3SE +/- 0.00020, N = 35.588765.586715.57958MIN: 5.55MIN: 5.55MIN: 5.561. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: IP Shapes 1D - Data Type: u8s8f32 - Engine: CPU123246810Min: 5.58 / Avg: 5.59 / Max: 5.59Min: 5.58 / Avg: 5.59 / Max: 5.59Min: 5.58 / Avg: 5.58 / Max: 5.581. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: IP Shapes 1D - Data Type: f32 - Engine: CPU123246810SE +/- 0.01592, N = 3SE +/- 0.00926, N = 3SE +/- 0.01958, N = 38.470608.479888.46603MIN: 8.4MIN: 8.39MIN: 8.371. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: IP Shapes 1D - Data Type: f32 - Engine: CPU1233691215Min: 8.44 / Avg: 8.47 / Max: 8.49Min: 8.46 / Avg: 8.48 / Max: 8.5Min: 8.44 / Avg: 8.47 / Max: 8.51. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPU1238001600240032004000SE +/- 5.10, N = 3SE +/- 1.86, N = 3SE +/- 1.54, N = 33855.793858.993852.80MIN: 3845.17MIN: 3851.44MIN: 3847.921. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPU1237001400210028003500Min: 3847.12 / Avg: 3855.79 / Max: 3864.78Min: 3856.53 / Avg: 3858.99 / Max: 3862.63Min: 3849.79 / Avg: 3852.8 / Max: 3854.851. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU1231.06192.12383.18574.24765.3095SE +/- 0.00413, N = 3SE +/- 0.00367, N = 3SE +/- 0.00063, N = 34.711864.719314.71943MIN: 4.66MIN: 4.66MIN: 4.661. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU123246810Min: 4.7 / Avg: 4.71 / Max: 4.72Min: 4.71 / Avg: 4.72 / Max: 4.72Min: 4.72 / Avg: 4.72 / Max: 4.721. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU-v3-v3 - Model: mobilenet-v3123246810SE +/- 0.01, N = 3SE +/- 0.02, N = 3SE +/- 0.02, N = 36.246.256.24MIN: 6.17 / MAX: 7.41MIN: 6.17 / MAX: 15.83MIN: 6.17 / MAX: 7.81. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU-v3-v3 - Model: mobilenet-v31233691215Min: 6.23 / Avg: 6.24 / Max: 6.25Min: 6.23 / Avg: 6.25 / Max: 6.29Min: 6.21 / Avg: 6.24 / Max: 6.261. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU-v3-v3 - Model: mobilenet-v3123246810SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 36.266.266.25MIN: 6.18 / MAX: 7.85MIN: 6.19 / MAX: 7.44MIN: 6.18 / MAX: 7.71. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU-v3-v3 - Model: mobilenet-v31233691215Min: 6.25 / Avg: 6.26 / Max: 6.26Min: 6.25 / Avg: 6.26 / Max: 6.26Min: 6.24 / Avg: 6.25 / Max: 6.251. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: alexnet123510152025SE +/- 0.01, N = 3SE +/- 0.02, N = 3SE +/- 0.01, N = 319.2819.3119.30MIN: 19.2 / MAX: 19.63MIN: 19.23 / MAX: 21.66MIN: 19.2 / MAX: 21.551. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: alexnet123510152025Min: 19.26 / Avg: 19.28 / Max: 19.3Min: 19.27 / Avg: 19.31 / Max: 19.34Min: 19.28 / Avg: 19.3 / Max: 19.321. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: mnasnet123246810SE +/- 0.01, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 36.546.546.53MIN: 6.5 / MAX: 8.09MIN: 6.51 / MAX: 7.17MIN: 6.49 / MAX: 7.161. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: mnasnet1233691215Min: 6.53 / Avg: 6.54 / Max: 6.55Min: 6.53 / Avg: 6.54 / Max: 6.54Min: 6.52 / Avg: 6.53 / Max: 6.541. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

VKMark

VKMark is a collection of Vulkan tests/benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgVKMark Score, More Is BetterVKMark 2020-05-21Resolution: 1920 x 1080123140280420560700SE +/- 1.76, N = 3SE +/- 0.58, N = 36576586581. (CXX) g++ options: -pthread -ldl -pipe -std=c++14 -MD -MQ -MF
OpenBenchmarking.orgVKMark Score, More Is BetterVKMark 2020-05-21Resolution: 1920 x 1080123120240360480600Min: 654 / Avg: 657.33 / Max: 660Min: 657 / Avg: 658 / Max: 6591. (CXX) g++ options: -pthread -ldl -pipe -std=c++14 -MD -MQ -MF

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: squeezenet_ssd123918273645SE +/- 0.03, N = 3SE +/- 0.04, N = 3SE +/- 0.02, N = 340.4040.4640.45MIN: 40.28 / MAX: 42.66MIN: 40.32 / MAX: 57.2MIN: 40.35 / MAX: 52.521. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: squeezenet_ssd123816243240Min: 40.35 / Avg: 40.4 / Max: 40.43Min: 40.42 / Avg: 40.46 / Max: 40.54Min: 40.41 / Avg: 40.45 / Max: 40.471. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: resnet501231122334455SE +/- 0.01, N = 3SE +/- 0.06, N = 3SE +/- 0.07, N = 347.8347.9047.88MIN: 47.62 / MAX: 57.56MIN: 47.62 / MAX: 88.54MIN: 47.6 / MAX: 95.31. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: resnet501231020304050Min: 47.82 / Avg: 47.83 / Max: 47.84Min: 47.83 / Avg: 47.9 / Max: 48.01Min: 47.81 / Avg: 47.88 / Max: 48.011. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

VkFFT

VkFFT is a Fast Fourier Transform (FFT) Library that is GPU accelerated by means of the Vulkan API. The VkFFT benchmark runs FFT performance differences of many different sizes before returning an overall benchmark score. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBenchmark Score, More Is BetterVkFFT 1.1.112330060090012001500SE +/- 2.33, N = 3SE +/- 1.20, N = 3SE +/- 1.00, N = 31367136913681. (CXX) g++ options: -O3 -pthread
OpenBenchmarking.orgBenchmark Score, More Is BetterVkFFT 1.1.11232004006008001000Min: 1362 / Avg: 1366.67 / Max: 1369Min: 1367 / Avg: 1369.33 / Max: 1371Min: 1366 / Avg: 1368 / Max: 13691. (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 2.0Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPU12315003000450060007500SE +/- 7.88, N = 3SE +/- 2.23, N = 3SE +/- 3.49, N = 36758.256759.256767.39MIN: 6740.42MIN: 6750.8MIN: 6755.11. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPU12312002400360048006000Min: 6743.04 / Avg: 6758.25 / Max: 6769.45Min: 6756.53 / Avg: 6759.25 / Max: 6763.68Min: 6760.41 / Avg: 6767.39 / Max: 6771.151. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

WavPack Audio Encoding

This test times how long it takes to encode a sample WAV file to WavPack format with very high quality settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterWavPack Audio Encoding 5.3WAV To WavPack12348121620SE +/- 0.02, N = 5SE +/- 0.00, N = 5SE +/- 0.00, N = 517.9717.9417.951. (CXX) g++ options: -rdynamic
OpenBenchmarking.orgSeconds, Fewer Is BetterWavPack Audio Encoding 5.3WAV To WavPack123510152025Min: 17.94 / Avg: 17.97 / Max: 18.05Min: 17.94 / Avg: 17.94 / Max: 17.95Min: 17.94 / Avg: 17.94 / Max: 17.951. (CXX) g++ options: -rdynamic

LZ4 Compression

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

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 3 - Compression Speed123918273645SE +/- 0.08, N = 3SE +/- 0.05, N = 3SE +/- 0.08, N = 340.9741.0240.991. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 3 - Compression Speed123918273645Min: 40.84 / Avg: 40.97 / Max: 41.13Min: 40.96 / Avg: 41.02 / Max: 41.12Min: 40.86 / Avg: 40.99 / Max: 41.141. (CC) gcc options: -O3

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 2.0Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPU1238001600240032004000SE +/- 1.72, N = 3SE +/- 1.45, N = 3SE +/- 2.63, N = 33856.203860.153855.49MIN: 3850.81MIN: 3854.62MIN: 3847.911. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPU1237001400210028003500Min: 3853.62 / Avg: 3856.2 / Max: 3859.47Min: 3858.54 / Avg: 3860.15 / Max: 3863.05Min: 3850.23 / Avg: 3855.49 / Max: 3858.311. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPU12348121620SE +/- 0.03, N = 3SE +/- 0.01, N = 3SE +/- 0.03, N = 314.5514.5414.54MIN: 14.46MIN: 14.47MIN: 14.441. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPU12348121620Min: 14.5 / Avg: 14.55 / Max: 14.6Min: 14.52 / Avg: 14.54 / Max: 14.56Min: 14.49 / Avg: 14.54 / Max: 14.571. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: mobilenet123714212835SE +/- 0.03, N = 3SE +/- 0.04, N = 3SE +/- 0.02, N = 328.1628.1428.13MIN: 28 / MAX: 30.49MIN: 27.98 / MAX: 33.11MIN: 28 / MAX: 30.421. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: mobilenet123612182430Min: 28.13 / Avg: 28.16 / Max: 28.23Min: 28.07 / Avg: 28.14 / Max: 28.18Min: 28.1 / Avg: 28.13 / Max: 28.171. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

rav1e

Xiph rav1e is a Rust-written AV1 video encoder. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is Betterrav1e 0.4 AlphaSpeed: 101230.65031.30061.95092.60123.2515SE +/- 0.004, N = 3SE +/- 0.006, N = 3SE +/- 0.006, N = 32.8902.8872.889
OpenBenchmarking.orgFrames Per Second, More Is Betterrav1e 0.4 AlphaSpeed: 10123246810Min: 2.88 / Avg: 2.89 / Max: 2.89Min: 2.88 / Avg: 2.89 / Max: 2.89Min: 2.88 / Avg: 2.89 / Max: 2.9

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: alexnet123510152025SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.03, N = 319.3219.3419.33MIN: 19.23 / MAX: 21.26MIN: 19.23 / MAX: 21.66MIN: 19.23 / MAX: 28.731. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: alexnet123510152025Min: 19.3 / Avg: 19.32 / Max: 19.33Min: 19.32 / Avg: 19.34 / Max: 19.37Min: 19.29 / Avg: 19.33 / Max: 19.41. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

rav1e

Xiph rav1e is a Rust-written AV1 video encoder. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is Betterrav1e 0.4 AlphaSpeed: 51230.22140.44280.66420.88561.107SE +/- 0.001, N = 3SE +/- 0.002, N = 3SE +/- 0.002, N = 30.9840.9840.983
OpenBenchmarking.orgFrames Per Second, More Is Betterrav1e 0.4 AlphaSpeed: 5123246810Min: 0.98 / Avg: 0.98 / Max: 0.99Min: 0.98 / Avg: 0.98 / Max: 0.99Min: 0.98 / Avg: 0.98 / Max: 0.99

Timed FFmpeg Compilation

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

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed FFmpeg Compilation 4.2.2Time To Compile123306090120150SE +/- 0.07, N = 3SE +/- 0.05, N = 3SE +/- 0.12, N = 3151.63151.68151.78
OpenBenchmarking.orgSeconds, Fewer Is BetterTimed FFmpeg Compilation 4.2.2Time To Compile123306090120150Min: 151.52 / Avg: 151.63 / Max: 151.75Min: 151.6 / Avg: 151.68 / Max: 151.77Min: 151.61 / Avg: 151.78 / Max: 152.02

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 2.0Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPU1233691215SE +/- 0.01, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 311.2511.2311.23MIN: 11.19MIN: 11.2MIN: 11.191. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPU1233691215Min: 11.23 / Avg: 11.25 / Max: 11.27Min: 11.23 / Avg: 11.23 / Max: 11.24Min: 11.22 / Avg: 11.23 / Max: 11.251. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

OpenVINO

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

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2021.1Model: Face Detection 0106 FP16 - Device: CPU1237001400210028003500SE +/- 2.94, N = 3SE +/- 2.63, N = 3SE +/- 5.13, N = 33426.153422.793426.05
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2021.1Model: Face Detection 0106 FP16 - Device: CPU1236001200180024003000Min: 3421.63 / Avg: 3426.15 / Max: 3431.68Min: 3417.94 / Avg: 3422.79 / Max: 3426.99Min: 3416.4 / Avg: 3426.05 / Max: 3433.89

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 2.0Preset: Exhaustive123160320480640800SE +/- 0.13, N = 3SE +/- 0.07, N = 3SE +/- 0.14, N = 3738.37737.79738.441. (CXX) g++ options: -std=c++14 -fvisibility=hidden -O3 -flto -mfpmath=sse -mavx2 -mpopcnt -lpthread
OpenBenchmarking.orgSeconds, Fewer Is BetterASTC Encoder 2.0Preset: Exhaustive123130260390520650Min: 738.12 / Avg: 738.37 / Max: 738.56Min: 737.72 / Avg: 737.79 / Max: 737.92Min: 738.21 / Avg: 738.44 / Max: 738.691. (CXX) g++ options: -std=c++14 -fvisibility=hidden -O3 -flto -mfpmath=sse -mavx2 -mpopcnt -lpthread

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: yolov4-tiny123918273645SE +/- 0.03, N = 3SE +/- 0.04, N = 3SE +/- 0.03, N = 338.9738.9638.94MIN: 38.82 / MAX: 49.15MIN: 38.78 / MAX: 48.94MIN: 38.78 / MAX: 41.061. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: yolov4-tiny123816243240Min: 38.93 / Avg: 38.97 / Max: 39.03Min: 38.89 / Avg: 38.96 / Max: 39.01Min: 38.88 / Avg: 38.94 / Max: 38.981. (CXX) g++ options: -O3 -rdynamic -lgomp -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 2.0Preset: Thorough12320406080100SE +/- 0.02, N = 3SE +/- 0.02, N = 3SE +/- 0.03, N = 392.9292.8892.951. (CXX) g++ options: -std=c++14 -fvisibility=hidden -O3 -flto -mfpmath=sse -mavx2 -mpopcnt -lpthread
OpenBenchmarking.orgSeconds, Fewer Is BetterASTC Encoder 2.0Preset: Thorough12320406080100Min: 92.89 / Avg: 92.92 / Max: 92.95Min: 92.85 / Avg: 92.88 / Max: 92.9Min: 92.89 / Avg: 92.95 / Max: 92.991. (CXX) g++ options: -std=c++14 -fvisibility=hidden -O3 -flto -mfpmath=sse -mavx2 -mpopcnt -lpthread

LZ4 Compression

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

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 3 - Decompression Speed1232K4K6K8K10KSE +/- 9.10, N = 3SE +/- 18.24, N = 3SE +/- 2.77, N = 39058.09051.39057.31. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 3 - Decompression Speed12316003200480064008000Min: 9045 / Avg: 9057.97 / Max: 9075.5Min: 9018.4 / Avg: 9051.3 / Max: 9081.4Min: 9052.1 / Avg: 9057.33 / Max: 9061.51. (CC) gcc options: -O3

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: mobilenet123714212835SE +/- 0.02, N = 3SE +/- 0.01, N = 3SE +/- 0.04, N = 328.1528.1528.17MIN: 28.03 / MAX: 30.22MIN: 28.04 / MAX: 30.44MIN: 28.04 / MAX: 38.211. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: mobilenet123612182430Min: 28.13 / Avg: 28.15 / Max: 28.18Min: 28.14 / Avg: 28.15 / Max: 28.17Min: 28.13 / Avg: 28.17 / Max: 28.241. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenVINO

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

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2021.1Model: Person Detection 0106 FP16 - Device: CPU12312002400360048006000SE +/- 3.62, N = 3SE +/- 4.42, N = 3SE +/- 3.88, N = 35641.285638.325642.32
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2021.1Model: Person Detection 0106 FP16 - Device: CPU12310002000300040005000Min: 5634.26 / Avg: 5641.28 / Max: 5646.33Min: 5632.23 / Avg: 5638.32 / Max: 5646.9Min: 5636.64 / Avg: 5642.32 / Max: 5649.74

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 2.0Preset: Medium12348121620SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 314.2814.2914.291. (CXX) g++ options: -std=c++14 -fvisibility=hidden -O3 -flto -mfpmath=sse -mavx2 -mpopcnt -lpthread
OpenBenchmarking.orgSeconds, Fewer Is BetterASTC Encoder 2.0Preset: Medium12348121620Min: 14.27 / Avg: 14.28 / Max: 14.29Min: 14.28 / Avg: 14.29 / Max: 14.31Min: 14.27 / Avg: 14.29 / Max: 14.311. (CXX) g++ options: -std=c++14 -fvisibility=hidden -O3 -flto -mfpmath=sse -mavx2 -mpopcnt -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 2.0Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPU123246810SE +/- 0.00777, N = 3SE +/- 0.00589, N = 3SE +/- 0.00679, N = 37.501147.506397.50482MIN: 7.46MIN: 7.46MIN: 7.461. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPU1233691215Min: 7.49 / Avg: 7.5 / Max: 7.51Min: 7.5 / Avg: 7.51 / Max: 7.52Min: 7.49 / Avg: 7.5 / Max: 7.521. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: vgg1612320406080100SE +/- 0.01, N = 3SE +/- 0.09, N = 3SE +/- 0.01, N = 387.2887.2487.23MIN: 87.01 / MAX: 97.33MIN: 86.94 / MAX: 96.82MIN: 87.02 / MAX: 97.071. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: vgg1612320406080100Min: 87.26 / Avg: 87.28 / Max: 87.3Min: 87.12 / Avg: 87.24 / Max: 87.41Min: 87.21 / Avg: 87.23 / Max: 87.251. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: yolov4-tiny123918273645SE +/- 0.06, N = 3SE +/- 0.01, N = 3SE +/- 0.02, N = 338.9938.9738.97MIN: 38.79 / MAX: 49.09MIN: 38.81 / MAX: 49.04MIN: 38.81 / MAX: 41.211. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: yolov4-tiny123816243240Min: 38.91 / Avg: 38.99 / Max: 39.11Min: 38.95 / Avg: 38.97 / Max: 38.99Min: 38.93 / Avg: 38.97 / Max: 39.011. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: resnet18123510152025SE +/- 0.03, N = 3SE +/- 0.02, N = 3SE +/- 0.01, N = 322.5222.5222.53MIN: 22.4 / MAX: 23.98MIN: 22.41 / MAX: 23.18MIN: 22.46 / MAX: 23.241. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: resnet18123510152025Min: 22.46 / Avg: 22.52 / Max: 22.56Min: 22.48 / Avg: 22.52 / Max: 22.55Min: 22.52 / Avg: 22.53 / Max: 22.541. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: resnet18123510152025SE +/- 0.01, N = 3SE +/- 0.02, N = 3SE +/- 0.02, N = 322.5522.5622.56MIN: 22.49 / MAX: 23.15MIN: 22.46 / MAX: 23.21MIN: 22.46 / MAX: 23.291. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: resnet18123510152025Min: 22.54 / Avg: 22.55 / Max: 22.56Min: 22.53 / Avg: 22.56 / Max: 22.6Min: 22.52 / Avg: 22.56 / Max: 22.591. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

Opus Codec Encoding

Opus is an open audio codec. Opus is a lossy audio compression format designed primarily for interactive real-time applications over the Internet. This test uses Opus-Tools and measures the time required to encode a WAV file to Opus. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterOpus Codec Encoding 1.3.1WAV To Opus Encode1233691215SE +/- 0.02, N = 5SE +/- 0.02, N = 5SE +/- 0.02, N = 510.6510.6510.651. (CXX) g++ options: -fvisibility=hidden -logg -lm
OpenBenchmarking.orgSeconds, Fewer Is BetterOpus Codec Encoding 1.3.1WAV To Opus Encode1233691215Min: 10.62 / Avg: 10.65 / Max: 10.72Min: 10.62 / Avg: 10.65 / Max: 10.73Min: 10.62 / Avg: 10.65 / Max: 10.721. (CXX) g++ options: -fvisibility=hidden -logg -lm

Timed HMMer Search

This test searches through the Pfam database of profile hidden markov models. The search finds the domain structure of Drosophila Sevenless protein. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed HMMer Search 3.3.1Pfam Database Search123306090120150SE +/- 0.02, N = 3SE +/- 0.04, N = 3SE +/- 0.06, N = 3126.16126.20126.151. (CC) gcc options: -O3 -pthread -lhmmer -leasel -lm
OpenBenchmarking.orgSeconds, Fewer Is BetterTimed HMMer Search 3.3.1Pfam Database Search12320406080100Min: 126.13 / Avg: 126.16 / Max: 126.19Min: 126.15 / Avg: 126.2 / Max: 126.28Min: 126.06 / Avg: 126.15 / Max: 126.261. (CC) gcc options: -O3 -pthread -lhmmer -leasel -lm

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: vgg1612320406080100SE +/- 0.03, N = 3SE +/- 0.09, N = 3SE +/- 0.11, N = 387.3587.3287.34MIN: 87.09 / MAX: 96.91MIN: 87 / MAX: 97.85MIN: 87.03 / MAX: 96.311. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: vgg1612320406080100Min: 87.29 / Avg: 87.35 / Max: 87.4Min: 87.16 / Avg: 87.32 / Max: 87.45Min: 87.15 / Avg: 87.34 / Max: 87.521. (CXX) g++ options: -O3 -rdynamic -lgomp -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 2.0Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU12314002800420056007000SE +/- 1.61, N = 3SE +/- 7.18, N = 3SE +/- 0.71, N = 36759.556761.686761.79MIN: 6754.19MIN: 6742.39MIN: 6755.521. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
OpenBenchmarking.orgms, Fewer Is BetteroneDNN 2.0Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU12312002400360048006000Min: 6757.26 / Avg: 6759.55 / Max: 6762.66Min: 6748.09 / Avg: 6761.68 / Max: 6772.5Min: 6760.6 / Avg: 6761.79 / Max: 6763.041. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

OpenVINO

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

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2021.1Model: Face Detection 0106 FP32 - Device: CPU1237001400210028003500SE +/- 2.49, N = 3SE +/- 6.17, N = 3SE +/- 9.57, N = 33436.653436.563437.37
OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2021.1Model: Face Detection 0106 FP32 - Device: CPU1236001200180024003000Min: 3432.22 / Avg: 3436.65 / Max: 3440.83Min: 3424.34 / Avg: 3436.56 / Max: 3444.1Min: 3419.31 / Avg: 3437.37 / Max: 3451.87

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2021.1Model: Person Detection 0106 FP16 - Device: CPU1230.15980.31960.47940.63920.799SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 30.710.710.71
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2021.1Model: Person Detection 0106 FP16 - Device: CPU123246810Min: 0.7 / Avg: 0.71 / Max: 0.71Min: 0.7 / Avg: 0.71 / Max: 0.71Min: 0.7 / Avg: 0.71 / Max: 0.71

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2021.1Model: Face Detection 0106 FP32 - Device: CPU1230.25880.51760.77641.03521.294SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 31.151.151.15
OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2021.1Model: Face Detection 0106 FP32 - Device: CPU123246810Min: 1.15 / Avg: 1.15 / Max: 1.15Min: 1.15 / Avg: 1.15 / Max: 1.15Min: 1.14 / Avg: 1.15 / Max: 1.15

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: blazeface1230.5311.0621.5932.1242.655SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 32.362.362.36MIN: 2.34 / MAX: 3.09MIN: 2.35 / MAX: 2.57MIN: 2.34 / MAX: 2.491. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: blazeface123246810Min: 2.36 / Avg: 2.36 / Max: 2.37Min: 2.36 / Avg: 2.36 / Max: 2.37Min: 2.36 / Avg: 2.36 / Max: 2.371. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: blazeface1230.5311.0621.5932.1242.655SE +/- 0.01, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 32.362.362.36MIN: 2.34 / MAX: 2.49MIN: 2.34 / MAX: 2.38MIN: 2.34 / MAX: 2.471. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: blazeface123246810Min: 2.35 / Avg: 2.36 / Max: 2.37Min: 2.35 / Avg: 2.36 / Max: 2.36Min: 2.36 / Avg: 2.36 / Max: 2.361. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

rav1e

Xiph rav1e is a Rust-written AV1 video encoder. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is Betterrav1e 0.4 AlphaSpeed: 61230.29430.58860.88291.17721.4715SE +/- 0.001, N = 3SE +/- 0.001, N = 3SE +/- 0.001, N = 31.3081.3081.308
OpenBenchmarking.orgFrames Per Second, More Is Betterrav1e 0.4 AlphaSpeed: 6123246810Min: 1.31 / Avg: 1.31 / Max: 1.31Min: 1.31 / Avg: 1.31 / Max: 1.31Min: 1.31 / Avg: 1.31 / Max: 1.31

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 0.7.1Throughput Test: DistinctUserID1230.12150.2430.36450.4860.6075SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 30.540.540.541. (CXX) g++ options: -O3 -pthread
OpenBenchmarking.orgGB/s, More Is Bettersimdjson 0.7.1Throughput Test: DistinctUserID123246810Min: 0.54 / Avg: 0.54 / Max: 0.54Min: 0.54 / Avg: 0.54 / Max: 0.54Min: 0.54 / Avg: 0.54 / Max: 0.541. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGB/s, More Is Bettersimdjson 0.7.1Throughput Test: PartialTweets1230.1170.2340.3510.4680.585SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 30.520.520.521. (CXX) g++ options: -O3 -pthread
OpenBenchmarking.orgGB/s, More Is Bettersimdjson 0.7.1Throughput Test: PartialTweets123246810Min: 0.52 / Avg: 0.52 / Max: 0.53Min: 0.52 / Avg: 0.52 / Max: 0.53Min: 0.52 / Avg: 0.52 / Max: 0.521. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGB/s, More Is Bettersimdjson 0.7.1Throughput Test: LargeRandom1230.07880.15760.23640.31520.394SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 30.350.350.351. (CXX) g++ options: -O3 -pthread
OpenBenchmarking.orgGB/s, More Is Bettersimdjson 0.7.1Throughput Test: LargeRandom12312345Min: 0.35 / Avg: 0.35 / Max: 0.35Min: 0.35 / Avg: 0.35 / Max: 0.35Min: 0.35 / Avg: 0.35 / Max: 0.351. (CXX) g++ options: -O3 -pthread

CLOMP

CLOMP is the C version of the Livermore OpenMP benchmark developed to measure OpenMP overheads and other performance impacts due to threading in order to influence future system designs. This particular test profile configuration is currently set to look at the OpenMP static schedule speed-up across all available CPU cores using the recommended test configuration. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSpeedup, More Is BetterCLOMP 1.2Static OMP Speedup1230.38250.7651.14751.531.9125SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 31.71.71.71. (CC) gcc options: -fopenmp -O3 -lm
OpenBenchmarking.orgSpeedup, More Is BetterCLOMP 1.2Static OMP Speedup123246810Min: 1.7 / Avg: 1.7 / Max: 1.7Min: 1.7 / Avg: 1.7 / Max: 1.7Min: 1.7 / Avg: 1.7 / Max: 1.71. (CC) gcc options: -fopenmp -O3 -lm

Betsy GPU Compressor

Betsy is an open-source GPU compressor of various GPU compression techniques. Betsy is written in GLSL for Vulkan/OpenGL (compute shader) support for GPU-based texture compression. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBetsy GPU Compressor 1.1 BetaCodec: ETC2 RGB - Quality: Highest12468107.5411. (CXX) g++ options: -O3 -O2 -lpthread -ldl

Coremark

This is a test of EEMBC CoreMark processor benchmark. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgIterations/Sec, More Is BetterCoremark 1.0CoreMark Size 666 - Iterations Per Second12320K40K60K80K100KSE +/- 1397.84, N = 15SE +/- 1569.03, N = 15SE +/- 971.22, N = 898117.51100468.3592000.141. (CC) gcc options: -O2 -lrt" -lrt
OpenBenchmarking.orgIterations/Sec, More Is BetterCoremark 1.0CoreMark Size 666 - Iterations Per Second12320K40K60K80K100KMin: 90059.66 / Avg: 98117.51 / Max: 106560.11Min: 88849.4 / Avg: 100468.35 / Max: 107570.26Min: 87772.23 / Avg: 92000.14 / Max: 95710.951. (CC) gcc options: -O2 -lrt" -lrt

Betsy GPU Compressor

Betsy is an open-source GPU compressor of various GPU compression techniques. Betsy is written in GLSL for Vulkan/OpenGL (compute shader) support for GPU-based texture compression. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBetsy GPU Compressor 1.1 BetaCodec: ETC1 - Quality: Highest1233691215SE +/- 0.78, N = 12SE +/- 0.08, N = 15SE +/- 0.63, N = 1210.9111.0810.371. (CXX) g++ options: -O3 -O2 -lpthread -ldl
OpenBenchmarking.orgSeconds, Fewer Is BetterBetsy GPU Compressor 1.1 BetaCodec: ETC1 - Quality: Highest1233691215Min: 4.31 / Avg: 10.91 / Max: 16.9Min: 10.81 / Avg: 11.08 / Max: 12.16Min: 3.48 / Avg: 10.37 / Max: 11.591. (CXX) g++ options: -O3 -O2 -lpthread -ldl

101 Results Shown

VkResample
oneDNN
Node.js V8 Web Tooling Benchmark
OpenVINO
Unpacking Firefox
simdjson
Stockfish
OpenVINO
VKMark
oneDNN
asmFish
oneDNN
SQLite Speedtest
BRL-CAD
OpenVINO
LZ4 Compression
oneDNN
NCNN
OpenVINO:
  Age Gender Recognition Retail 0013 FP32 - CPU
  Age Gender Recognition Retail 0013 FP16 - CPU
oneDNN
Timed MAFFT Alignment
NCNN
Monkey Audio Encoding
oneDNN
Timed Eigen Compilation
oneDNN
Build2
oneDNN
NCNN
OpenVINO
VkResample
NCNN:
  Vulkan GPU - shufflenet-v2
  Vulkan GPU - resnet50
LZ4 Compression
rav1e
NCNN:
  CPU - efficientnet-b0
  Vulkan GPU-v2-v2 - mobilenet-v2
  CPU - googlenet
LZ4 Compression
NCNN
OpenVINO
NCNN:
  Vulkan GPU - regnety_400m
  Vulkan GPU - squeezenet_ssd
  Vulkan GPU - efficientnet-b0
  Vulkan GPU - googlenet
ASTC Encoder
LZ4 Compression
oneDNN:
  IP Shapes 1D - u8s8f32 - CPU
  IP Shapes 1D - f32 - CPU
  Recurrent Neural Network Inference - u8s8f32 - CPU
  Matrix Multiply Batch Shapes Transformer - f32 - CPU
NCNN:
  CPU-v3-v3 - mobilenet-v3
  Vulkan GPU-v3-v3 - mobilenet-v3
  Vulkan GPU - alexnet
  Vulkan GPU - mnasnet
VKMark
NCNN:
  CPU - squeezenet_ssd
  CPU - resnet50
VkFFT
oneDNN
WavPack Audio Encoding
LZ4 Compression
oneDNN:
  Recurrent Neural Network Inference - bf16bf16bf16 - CPU
  Deconvolution Batch shapes_1d - u8s8f32 - CPU
NCNN
rav1e
NCNN
rav1e
Timed FFmpeg Compilation
oneDNN
OpenVINO
ASTC Encoder
NCNN
ASTC Encoder
LZ4 Compression
NCNN
OpenVINO
ASTC Encoder
oneDNN
NCNN:
  Vulkan GPU - vgg16
  CPU - yolov4-tiny
  Vulkan GPU - resnet18
  CPU - resnet18
Opus Codec Encoding
Timed HMMer Search
NCNN
oneDNN
OpenVINO:
  Face Detection 0106 FP32 - CPU
  Person Detection 0106 FP16 - CPU
  Face Detection 0106 FP32 - CPU
NCNN:
  Vulkan GPU - blazeface
  CPU - blazeface
rav1e
simdjson:
  DistinctUserID
  PartialTweets
  LargeRand
CLOMP
Betsy GPU Compressor
Coremark
Betsy GPU Compressor