Core i9 9900K Xmas

Intel Core i9-9900K testing with a ASRock Z390M Pro4 (P4.20 BIOS) and Intel UHD 630 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 2012229-HA-COREI999049
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 4 Tests
Timed Code Compilation 2 Tests
CPU Massive 2 Tests
Creator Workloads 5 Tests
Encoding 4 Tests
HPC - High Performance Computing 2 Tests
Machine Learning 2 Tests
Multi-Core 3 Tests
NVIDIA GPU Compute 3 Tests
Programmer / Developer System Benchmarks 4 Tests
Server 2 Tests
Vulkan Compute 3 Tests

Statistics

Show Overall Harmonic Mean(s)
Show Overall Geometric Mean
Show Geometric Means Per-Suite/Category
Show Wins / Losses Counts (Pie Chart)
Normalize Results
Remove Outliers Before Calculating Averages

Graph Settings

Force Line Graphs Where Applicable
Convert To Scalar Where Applicable
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 22 2020
  3 Hours, 31 Minutes
2
December 22 2020
  3 Hours, 33 Minutes
3
December 22 2020
  3 Hours, 38 Minutes
Invert Hiding All Results Option
  3 Hours, 34 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):

1

Compiler Notes: --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 Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xd6 - Thermald 1.9.1
Security Notes: itlb_multihit: KVM: Mitigation of VMX disabled + l1tf: Not affected + mds: Vulnerable; SMT vulnerable + meltdown: Not affected + spec_store_bypass: Vulnerable + spectre_v1: Vulnerable: __user pointer sanitization and usercopy barriers only; no swapgs barriers + spectre_v2: Vulnerable IBPB: disabled STIBP: disabled + srbds: Vulnerable + tsx_async_abort: Vulnerable

2

3

Processor: Intel Core i9-9900K @ 5.00GHz (8 Cores / 16 Threads), Motherboard: ASRock Z390M Pro4 (P4.20 BIOS), Chipset: Intel Cannon Lake PCH, Memory: 16GB, Disk: 240GB Corsair Force MP510, Graphics: Intel UHD 630 3GB (1200MHz), Audio: Realtek ALC892, Monitor: G237HL, Network: Intel I219-V

OS: Ubuntu 20.04, Kernel: 5.9.0-050900rc1daily20200819-generic (x86_64) 20200818, Desktop: GNOME Shell 3.36.4, Display Server: X Server 1.20.8, Display Driver: modesetting 1.20.8, OpenGL: 4.6 Mesa 20.0.4, OpenCL: OpenCL 2.1, Compiler: GCC 9.3.0, File-System: ext4, Screen Resolution: 1920x1080

Core i9 9900K XmasProcessorMotherboardChipsetMemoryDiskGraphicsAudioMonitorNetworkOSKernelDesktopDisplay ServerDisplay DriverOpenGLOpenCLCompilerFile-SystemScreen Resolution123Intel Core i9-9900K @ 5.00GHz (8 Cores / 16 Threads)ASRock Z390M Pro4 (P4.20 BIOS)Intel Cannon Lake PCH16GB240GB Corsair Force MP510Intel UHD 630 3GB (1200MHz)Realtek ALC892G237HLIntel I219-VUbuntu 20.045.9.0-050900rc1daily20200819-generic (x86_64) 20200818GNOME Shell 3.36.4X Server 1.20.8modesetting 1.20.84.6 Mesa 20.0.4OpenCL 2.1GCC 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,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: 0xd6 - Thermald 1.9.1 Security Details- itlb_multihit: KVM: Mitigation of VMX disabled + l1tf: Not affected + mds: Vulnerable; SMT vulnerable + meltdown: Not affected + spec_store_bypass: Vulnerable + spectre_v1: Vulnerable: __user pointer sanitization and usercopy barriers only; no swapgs barriers + spectre_v2: Vulnerable IBPB: disabled STIBP: disabled + srbds: Vulnerable + tsx_async_abort: Vulnerable

123Result OverviewPhoronix Test Suite100%102%105%107%CLOMPNode.js V8 Web Tooling BenchmarkTimed Eigen CompilationBuild2VkResampleOpus Codec EncodingNCNNMonkey Audio EncodingOgg Audio EncodingVkFFTWavPack Audio EncodingoneDNNsimdjson

Core i9 9900K Xmasclomp: Static OMP Speedupncnn: Vulkan GPU - yolov4-tinyncnn: CPU - regnety_400mnode-web-tooling: ncnn: Vulkan GPU - mnasnetncnn: Vulkan GPU - blazefaceonednn: IP Shapes 3D - f32 - CPUncnn: Vulkan GPU-v3-v3 - mobilenet-v3ncnn: CPU-v3-v3 - mobilenet-v3ncnn: Vulkan GPU - shufflenet-v2ncnn: Vulkan GPU - resnet50ncnn: Vulkan GPU - squeezenet_ssdvkresample: 2x - Doubleonednn: IP Shapes 1D - u8s8f32 - CPUncnn: Vulkan GPU - resnet18onednn: IP Shapes 3D - u8s8f32 - CPUncnn: CPU - blazefacencnn: Vulkan GPU - efficientnet-b0build-eigen: Time To Compileonednn: Deconvolution Batch shapes_1d - u8s8f32 - CPUonednn: Matrix Multiply Batch Shapes Transformer - f32 - CPUbuild2: Time To Compilencnn: CPU - mnasnetonednn: IP Shapes 1D - f32 - CPUonednn: Recurrent Neural Network Inference - u8s8f32 - CPUonednn: Convolution Batch Shapes Auto - u8s8f32 - CPUonednn: Convolution Batch Shapes Auto - f32 - CPUncnn: Vulkan GPU-v2-v2 - mobilenet-v2ncnn: CPU - googlenetonednn: Deconvolution Batch shapes_3d - u8s8f32 - CPUncnn: CPU - resnet50encode-opus: WAV To Opus Encodencnn: Vulkan GPU - mobilenetncnn: CPU - alexnetncnn: CPU - shufflenet-v2onednn: Deconvolution Batch shapes_3d - f32 - CPUonednn: Deconvolution Batch shapes_1d - f32 - CPUncnn: Vulkan GPU - googlenetencode-ape: WAV To APEncnn: Vulkan GPU - vgg16ncnn: Vulkan GPU - alexnetncnn: CPU - efficientnet-b0vkresample: 2x - Singleonednn: Recurrent Neural Network Inference - f32 - CPUencode-ogg: WAV To Oggvkfft: ncnn: CPU - squeezenet_ssdncnn: CPU - yolov4-tinyonednn: Recurrent Neural Network Training - bf16bf16bf16 - CPUonednn: Recurrent Neural Network Inference - bf16bf16bf16 - CPUncnn: CPU-v2-v2 - mobilenet-v2encode-wavpack: WAV To WavPackonednn: Recurrent Neural Network Training - f32 - CPUncnn: CPU - vgg16ncnn: CPU - resnet18ncnn: CPU - mobilenetonednn: Recurrent Neural Network Training - u8s8f32 - CPUonednn: Matrix Multiply Batch Shapes Transformer - u8s8f32 - CPUsimdjson: DistinctUserIDsimdjson: PartialTweetssimdjson: LargeRandsimdjson: Kostyancnn: Vulkan GPU - regnety_400m1233.625.6312.7814.203.721.7510.90803.743.794.7625.4618.93903.6521.7567713.262.028791.786.2166.6785.767293.87754130.4203.703.842972066.0917.262319.49684.8913.513.1131625.087.70917.9011.274.756.247835.2386513.479.86959.1611.296.15388.5782073.4818.105150118.8825.563763.982072.874.9013.0523758.9259.1013.1717.843765.373.194670.750.730.490.812.693.625.5912.4913.723.841.811.09753.813.844.8425.0418.62918.5331.7432213.162.004191.766.2466.8355.736493.84547130.2143.723.861242074.9717.150419.58204.9213.453.1297325.057.73417.8211.254.776.258475.2602013.529.86758.9511.256.13387.3242070.8118.114150418.9325.573763.032074.814.9113.0263760.8359.0113.1717.823767.163.194030.750.730.490.813.403.326.6912.3314.053.721.7611.20723.793.774.8025.1918.72903.6891.7288513.332.029401.766.1767.2755.719313.85413131.2983.693.831812081.0217.263819.62154.9213.533.1183125.177.69917.8711.304.766.232655.2390213.489.90358.9511.276.14388.4332077.3218.156150018.8825.623771.302077.344.9013.0463765.1059.0113.1917.823767.183.194530.750.730.490.812.41OpenBenchmarking.org

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 Speedup1230.811.622.433.244.05SE +/- 0.00, N = 3SE +/- 0.03, N = 3SE +/- 0.04, N = 153.63.63.31. (CC) gcc options: -fopenmp -O3 -lm
OpenBenchmarking.orgSpeedup, More Is BetterCLOMP 1.2Static OMP Speedup123246810Min: 3.6 / Avg: 3.6 / Max: 3.6Min: 3.6 / Avg: 3.63 / Max: 3.7Min: 3.1 / Avg: 3.29 / Max: 3.71. (CC) gcc options: -fopenmp -O3 -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: Vulkan GPU - Model: yolov4-tiny123612182430SE +/- 0.03, N = 3SE +/- 0.02, N = 3SE +/- 1.12, N = 225.6325.5926.69MIN: 25.45 / MAX: 26.86MIN: 25.42 / MAX: 35.3MIN: 25.44 / MAX: 170.171. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: yolov4-tiny123612182430Min: 25.59 / Avg: 25.63 / Max: 25.69Min: 25.56 / Avg: 25.59 / Max: 25.63Min: 25.57 / Avg: 26.69 / Max: 27.811. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: regnety_400m1233691215SE +/- 0.04, N = 3SE +/- 0.08, N = 3SE +/- 0.19, N = 312.7812.4912.33MIN: 12.1 / MAX: 21.92MIN: 12.09 / MAX: 13.69MIN: 11.9 / MAX: 13.181. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: regnety_400m12348121620Min: 12.73 / Avg: 12.78 / Max: 12.85Min: 12.4 / Avg: 12.49 / Max: 12.65Min: 11.98 / Avg: 12.33 / Max: 12.631. (CXX) g++ options: -O3 -rdynamic -lgomp -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 Benchmark12348121620SE +/- 0.07, N = 3SE +/- 0.05, N = 3SE +/- 0.11, N = 314.2013.7214.051. Nodejs v10.19.0
OpenBenchmarking.orgruns/s, More Is BetterNode.js V8 Web Tooling Benchmark12348121620Min: 14.06 / Avg: 14.2 / Max: 14.29Min: 13.64 / Avg: 13.72 / Max: 13.8Min: 13.85 / Avg: 14.05 / Max: 14.221. Nodejs v10.19.0

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: mnasnet1230.8641.7282.5923.4564.32SE +/- 0.00, N = 3SE +/- 0.13, N = 2SE +/- 0.03, N = 33.723.843.72MIN: 3.69 / MAX: 4.57MIN: 3.68 / MAX: 12.58MIN: 3.66 / MAX: 5.031. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: mnasnet123246810Min: 3.71 / Avg: 3.72 / Max: 3.72Min: 3.71 / Avg: 3.84 / Max: 3.97Min: 3.68 / Avg: 3.72 / Max: 3.781. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: blazeface1230.4050.811.2151.622.025SE +/- 0.05, N = 3SE +/- 0.03, N = 3SE +/- 0.04, N = 31.751.801.76MIN: 1.63 / MAX: 2.06MIN: 1.63 / MAX: 2.02MIN: 1.67 / MAX: 2.011. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: blazeface123246810Min: 1.66 / Avg: 1.75 / Max: 1.8Min: 1.7 / Avg: 1.77 / Max: 1.8Min: 1.68 / Avg: 1.76 / Max: 1.811. (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: IP Shapes 3D - Data Type: f32 - Engine: CPU1233691215SE +/- 0.02, N = 3SE +/- 0.03, N = 3SE +/- 0.04, N = 310.9111.1011.21MIN: 10.73MIN: 10.87MIN: 10.991. (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: 10.87 / Avg: 10.91 / Max: 10.94Min: 11.04 / Avg: 11.1 / Max: 11.14Min: 11.13 / Avg: 11.21 / Max: 11.281. (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-v3-v3 - Model: mobilenet-v31230.85731.71462.57193.42924.2865SE +/- 0.01, N = 3SE +/- 0.08, N = 3SE +/- 0.05, N = 33.743.813.79MIN: 3.7 / MAX: 4.76MIN: 3.7 / MAX: 4.85MIN: 3.7 / MAX: 4.791. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU-v3-v3 - Model: mobilenet-v3123246810Min: 3.73 / Avg: 3.74 / Max: 3.76Min: 3.73 / Avg: 3.81 / Max: 3.96Min: 3.74 / Avg: 3.79 / Max: 3.891. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU-v3-v3 - Model: mobilenet-v31230.8641.7282.5923.4564.32SE +/- 0.03, N = 3SE +/- 0.11, N = 3SE +/- 0.04, N = 33.793.843.77MIN: 3.72 / MAX: 4.73MIN: 3.71 / MAX: 4.74MIN: 3.69 / MAX: 4.61. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU-v3-v3 - Model: mobilenet-v3123246810Min: 3.76 / Avg: 3.79 / Max: 3.85Min: 3.73 / Avg: 3.84 / Max: 4.05Min: 3.72 / Avg: 3.77 / Max: 3.841. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: shufflenet-v21231.0892.1783.2674.3565.445SE +/- 0.01, N = 3SE +/- 0.09, N = 3SE +/- 0.04, N = 34.764.844.80MIN: 4.73 / MAX: 6.69MIN: 4.71 / MAX: 6.78MIN: 4.67 / MAX: 5.751. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: shufflenet-v2123246810Min: 4.75 / Avg: 4.76 / Max: 4.77Min: 4.74 / Avg: 4.84 / Max: 5.02Min: 4.76 / Avg: 4.8 / Max: 4.891. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: resnet50123612182430SE +/- 0.02, N = 3SE +/- 0.37, N = 3SE +/- 0.37, N = 325.4625.0425.19MIN: 24.03 / MAX: 26.02MIN: 24.07 / MAX: 26.3MIN: 24.27 / MAX: 35.771. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: resnet50123612182430Min: 25.43 / Avg: 25.46 / Max: 25.5Min: 24.31 / Avg: 25.04 / Max: 25.45Min: 24.46 / Avg: 25.19 / Max: 25.641. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: squeezenet_ssd123510152025SE +/- 0.03, N = 3SE +/- 0.23, N = 3SE +/- 0.22, N = 318.9318.6218.72MIN: 18.69 / MAX: 21.76MIN: 18.08 / MAX: 19.05MIN: 18.2 / MAX: 30.11. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: squeezenet_ssd123510152025Min: 18.88 / Avg: 18.93 / Max: 18.97Min: 18.17 / Avg: 18.62 / Max: 18.86Min: 18.29 / Avg: 18.72 / Max: 18.961. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

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 +/- 1.21, N = 3SE +/- 0.93, N = 3SE +/- 0.41, N = 3903.65918.53903.691. (CXX) g++ options: -O3 -pthread
OpenBenchmarking.orgms, Fewer Is BetterVkResample 1.0Upscale: 2x - Precision: Double123160320480640800Min: 902.33 / Avg: 903.65 / Max: 906.08Min: 916.93 / Avg: 918.53 / Max: 920.16Min: 903.17 / Avg: 903.69 / Max: 904.51. (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 1D - Data Type: u8s8f32 - Engine: CPU1230.39530.79061.18591.58121.9765SE +/- 0.01447, N = 13SE +/- 0.01201, N = 15SE +/- 0.01630, N = 101.756771.743221.72885MIN: 1.53MIN: 1.53MIN: 1.551. (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: 1.6 / Avg: 1.76 / Max: 1.82Min: 1.58 / Avg: 1.74 / Max: 1.78Min: 1.58 / Avg: 1.73 / Max: 1.751. (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: resnet181233691215SE +/- 0.12, N = 3SE +/- 0.16, N = 3SE +/- 0.02, N = 213.2613.1613.33MIN: 12.86 / MAX: 14.3MIN: 12.64 / MAX: 14.54MIN: 12.89 / MAX: 14.111. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: resnet1812348121620Min: 13.02 / Avg: 13.26 / Max: 13.38Min: 12.84 / Avg: 13.16 / Max: 13.35Min: 13.31 / Avg: 13.33 / Max: 13.351. (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: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPU1230.45660.91321.36981.82642.283SE +/- 0.00340, N = 3SE +/- 0.00370, N = 3SE +/- 0.00254, N = 32.028792.004192.02940MIN: 1.97MIN: 1.94MIN: 1.971. (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: 2.02 / Avg: 2.03 / Max: 2.03Min: 2 / Avg: 2 / Max: 2.01Min: 2.02 / Avg: 2.03 / Max: 2.031. (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: blazeface1230.40050.8011.20151.6022.0025SE +/- 0.02, N = 3SE +/- 0.04, N = 3SE +/- 0.05, N = 31.781.761.76MIN: 1.63 / MAX: 15.8MIN: 1.67 / MAX: 2.05MIN: 1.63 / MAX: 2.051. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: blazeface123246810Min: 1.74 / Avg: 1.78 / Max: 1.81Min: 1.69 / Avg: 1.76 / Max: 1.81Min: 1.67 / Avg: 1.76 / Max: 1.821. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: efficientnet-b0123246810SE +/- 0.04, N = 3SE +/- 0.07, N = 3SE +/- 0.01, N = 36.216.246.17MIN: 6.11 / MAX: 7.67MIN: 6.1 / MAX: 8.11MIN: 6.03 / MAX: 7.421. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: efficientnet-b0123246810Min: 6.14 / Avg: 6.21 / Max: 6.28Min: 6.17 / Avg: 6.24 / Max: 6.38Min: 6.15 / Avg: 6.17 / Max: 6.191. (CXX) g++ options: -O3 -rdynamic -lgomp -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 Compile1231530456075SE +/- 0.12, N = 3SE +/- 0.20, N = 3SE +/- 0.28, N = 366.6866.8467.28
OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Eigen Compilation 3.3.9Time To Compile1231326395265Min: 66.51 / Avg: 66.68 / Max: 66.91Min: 66.58 / Avg: 66.83 / Max: 67.22Min: 66.71 / Avg: 67.27 / Max: 67.62

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: u8s8f32 - Engine: CPU1231.29762.59523.89285.19046.488SE +/- 0.04711, N = 15SE +/- 0.05212, N = 13SE +/- 0.04354, N = 155.767295.736495.71931MIN: 5.03MIN: 5.03MIN: 5.031. (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: CPU123246810Min: 5.15 / Avg: 5.77 / Max: 5.96Min: 5.15 / Avg: 5.74 / Max: 5.89Min: 5.15 / Avg: 5.72 / Max: 5.871. (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: CPU1230.87241.74482.61723.48964.362SE +/- 0.00398, N = 3SE +/- 0.01019, N = 3SE +/- 0.00211, N = 33.877543.845473.85413MIN: 3.81MIN: 3.77MIN: 3.781. (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: 3.87 / Avg: 3.88 / Max: 3.88Min: 3.83 / Avg: 3.85 / Max: 3.86Min: 3.85 / Avg: 3.85 / Max: 3.861. (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 Compile123306090120150SE +/- 0.20, N = 3SE +/- 0.30, N = 3SE +/- 0.13, N = 3130.42130.21131.30
OpenBenchmarking.orgSeconds, Fewer Is BetterBuild2 0.13Time To Compile12320406080100Min: 130.02 / Avg: 130.42 / Max: 130.68Min: 129.91 / Avg: 130.21 / Max: 130.81Min: 131.05 / Avg: 131.3 / Max: 131.46

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: mnasnet1230.8371.6742.5113.3484.185SE +/- 0.02, N = 3SE +/- 0.03, N = 3SE +/- 0.02, N = 33.703.723.69MIN: 3.62 / MAX: 5.08MIN: 3.6 / MAX: 4.76MIN: 3.6 / MAX: 4.751. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: mnasnet123246810Min: 3.67 / Avg: 3.7 / Max: 3.73Min: 3.68 / Avg: 3.72 / Max: 3.77Min: 3.65 / Avg: 3.69 / Max: 3.711. (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: IP Shapes 1D - Data Type: f32 - Engine: CPU1230.86881.73762.60643.47524.344SE +/- 0.04278, N = 7SE +/- 0.03519, N = 10SE +/- 0.04260, N = 73.842973.861243.83181MIN: 3.47MIN: 3.49MIN: 3.421. (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: CPU123246810Min: 3.61 / Avg: 3.84 / Max: 3.93Min: 3.63 / Avg: 3.86 / Max: 3.94Min: 3.62 / Avg: 3.83 / Max: 3.911. (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: u8s8f32 - Engine: CPU123400800120016002000SE +/- 5.12, N = 3SE +/- 1.26, N = 3SE +/- 1.64, N = 32066.092074.972081.02MIN: 2046.99MIN: 2058.74MIN: 2063.521. (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: CPU123400800120016002000Min: 2059.32 / Avg: 2066.09 / Max: 2076.13Min: 2072.5 / Avg: 2074.97 / Max: 2076.62Min: 2077.82 / Avg: 2081.02 / Max: 2083.241. (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: CPU12348121620SE +/- 0.02, N = 3SE +/- 0.03, N = 3SE +/- 0.02, N = 317.2617.1517.26MIN: 16.85MIN: 16.71MIN: 16.81. (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: CPU12348121620Min: 17.24 / Avg: 17.26 / Max: 17.31Min: 17.1 / Avg: 17.15 / Max: 17.21Min: 17.24 / Avg: 17.26 / Max: 17.31. (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: CPU123510152025SE +/- 0.02, N = 3SE +/- 0.02, N = 3SE +/- 0.02, N = 319.5019.5819.62MIN: 19.39MIN: 19.45MIN: 19.481. (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: 19.46 / Avg: 19.5 / Max: 19.52Min: 19.56 / Avg: 19.58 / Max: 19.62Min: 19.59 / Avg: 19.62 / Max: 19.671. (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-v2-v2 - Model: mobilenet-v21231.1072.2143.3214.4285.535SE +/- 0.02, N = 3SE +/- 0.05, N = 3SE +/- 0.05, N = 34.894.924.92MIN: 4.76 / MAX: 14.13MIN: 4.75 / MAX: 6.03MIN: 4.76 / MAX: 6.221. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU-v2-v2 - Model: mobilenet-v2123246810Min: 4.86 / Avg: 4.89 / Max: 4.93Min: 4.86 / Avg: 4.92 / Max: 5.01Min: 4.87 / Avg: 4.92 / Max: 5.011. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: googlenet1233691215SE +/- 0.28, N = 3SE +/- 0.25, N = 3SE +/- 0.28, N = 313.5113.4513.53MIN: 12.77 / MAX: 14.18MIN: 12.78 / MAX: 14.09MIN: 12.81 / MAX: 14.351. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: googlenet12348121620Min: 12.96 / Avg: 13.51 / Max: 13.87Min: 12.96 / Avg: 13.45 / Max: 13.75Min: 12.99 / Avg: 13.53 / Max: 13.931. (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: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPU1230.70421.40842.11262.81683.521SE +/- 0.00605, N = 3SE +/- 0.00115, N = 3SE +/- 0.00835, N = 33.113163.129733.11831MIN: 3.04MIN: 3.05MIN: 3.051. (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: CPU123246810Min: 3.1 / Avg: 3.11 / Max: 3.12Min: 3.13 / Avg: 3.13 / Max: 3.13Min: 3.11 / Avg: 3.12 / Max: 3.131. (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: resnet50123612182430SE +/- 0.37, N = 3SE +/- 0.37, N = 3SE +/- 0.36, N = 325.0825.0525.17MIN: 24.17 / MAX: 26.32MIN: 24.09 / MAX: 37.48MIN: 24.27 / MAX: 35.81. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: resnet50123612182430Min: 24.33 / Avg: 25.08 / Max: 25.47Min: 24.32 / Avg: 25.05 / Max: 25.49Min: 24.45 / Avg: 25.17 / Max: 25.541. (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 Encode123246810SE +/- 0.005, N = 5SE +/- 0.025, N = 5SE +/- 0.009, N = 57.7097.7347.6991. (CXX) g++ options: -fvisibility=hidden -logg -lm
OpenBenchmarking.orgSeconds, Fewer Is BetterOpus Codec Encoding 1.3.1WAV To Opus Encode1233691215Min: 7.7 / Avg: 7.71 / Max: 7.72Min: 7.68 / Avg: 7.73 / Max: 7.83Min: 7.67 / Avg: 7.7 / Max: 7.721. (CXX) g++ options: -fvisibility=hidden -logg -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: Vulkan GPU - Model: mobilenet12348121620SE +/- 0.02, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 317.9017.8217.87MIN: 17.81 / MAX: 19.41MIN: 17.55 / MAX: 18.33MIN: 17.62 / MAX: 19.751. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: mobilenet123510152025Min: 17.88 / Avg: 17.9 / Max: 17.93Min: 17.8 / Avg: 17.82 / Max: 17.83Min: 17.85 / Avg: 17.87 / Max: 17.891. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: alexnet1233691215SE +/- 0.06, N = 3SE +/- 0.07, N = 3SE +/- 0.10, N = 311.2711.2511.30MIN: 11.03 / MAX: 13.48MIN: 11.02 / MAX: 11.94MIN: 11.05 / MAX: 11.641. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: alexnet1233691215Min: 11.14 / Avg: 11.27 / Max: 11.34Min: 11.11 / Avg: 11.25 / Max: 11.34Min: 11.1 / Avg: 11.3 / Max: 11.451. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: shufflenet-v21231.07332.14663.21994.29325.3665SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 34.754.774.76MIN: 4.72 / MAX: 5.82MIN: 4.71 / MAX: 5.82MIN: 4.73 / MAX: 5.741. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: shufflenet-v2123246810Min: 4.74 / Avg: 4.75 / Max: 4.75Min: 4.77 / Avg: 4.77 / Max: 4.77Min: 4.76 / Avg: 4.76 / Max: 4.761. (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: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPU123246810SE +/- 0.00100, N = 3SE +/- 0.01417, N = 3SE +/- 0.01139, N = 36.247836.258476.23265MIN: 5.97MIN: 5.96MIN: 5.941. (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: CPU1233691215Min: 6.25 / Avg: 6.25 / Max: 6.25Min: 6.24 / Avg: 6.26 / Max: 6.28Min: 6.22 / Avg: 6.23 / Max: 6.261. (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: CPU1231.18352.3673.55054.7345.9175SE +/- 0.06499, N = 12SE +/- 0.05219, N = 14SE +/- 0.06401, N = 125.238655.260205.23902MIN: 4.34MIN: 4.42MIN: 4.351. (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: CPU123246810Min: 4.53 / Avg: 5.24 / Max: 5.32Min: 4.59 / Avg: 5.26 / Max: 5.35Min: 4.54 / Avg: 5.24 / Max: 5.381. (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: googlenet1233691215SE +/- 0.23, N = 3SE +/- 0.36, N = 3SE +/- 0.30, N = 313.4713.5213.48MIN: 12.89 / MAX: 16MIN: 12.59 / MAX: 14.8MIN: 12.7 / MAX: 14.541. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: googlenet12348121620Min: 13 / Avg: 13.47 / Max: 13.73Min: 12.83 / Avg: 13.52 / Max: 14.04Min: 12.89 / Avg: 13.48 / Max: 13.811. (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 APE1233691215SE +/- 0.009, N = 5SE +/- 0.013, N = 5SE +/- 0.024, N = 59.8699.8679.9031. (CXX) g++ options: -O3 -pedantic -rdynamic -lrt
OpenBenchmarking.orgSeconds, Fewer Is BetterMonkey Audio Encoding 3.99.6WAV To APE1233691215Min: 9.84 / Avg: 9.87 / Max: 9.89Min: 9.82 / Avg: 9.87 / Max: 9.9Min: 9.86 / Avg: 9.9 / Max: 9.991. (CXX) g++ options: -O3 -pedantic -rdynamic -lrt

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: vgg161231326395265SE +/- 0.01, N = 3SE +/- 0.06, N = 3SE +/- 0.08, N = 359.1658.9558.95MIN: 58.94 / MAX: 60.55MIN: 58.61 / MAX: 59.89MIN: 58.53 / MAX: 67.61. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: vgg161231224364860Min: 59.15 / Avg: 59.16 / Max: 59.17Min: 58.83 / Avg: 58.95 / Max: 59.03Min: 58.78 / Avg: 58.95 / Max: 59.051. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: alexnet1233691215SE +/- 0.10, N = 3SE +/- 0.08, N = 3SE +/- 0.08, N = 311.2911.2511.27MIN: 11.05 / MAX: 12.62MIN: 11.02 / MAX: 12.47MIN: 11.05 / MAX: 11.641. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: alexnet1233691215Min: 11.08 / Avg: 11.29 / Max: 11.41Min: 11.09 / Avg: 11.25 / Max: 11.35Min: 11.1 / Avg: 11.27 / Max: 11.361. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: efficientnet-b0123246810SE +/- 0.03, N = 3SE +/- 0.05, N = 3SE +/- 0.04, N = 36.156.136.14MIN: 6.05 / MAX: 8.81MIN: 5.97 / MAX: 7.36MIN: 5.98 / MAX: 7.251. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: efficientnet-b0123246810Min: 6.1 / Avg: 6.15 / Max: 6.19Min: 6.03 / Avg: 6.13 / Max: 6.2Min: 6.07 / Avg: 6.14 / Max: 6.181. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

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: Single12380160240320400SE +/- 0.48, N = 3SE +/- 0.23, N = 3SE +/- 0.46, N = 3388.58387.32388.431. (CXX) g++ options: -O3 -pthread
OpenBenchmarking.orgms, Fewer Is BetterVkResample 1.0Upscale: 2x - Precision: Single12370140210280350Min: 387.65 / Avg: 388.58 / Max: 389.27Min: 386.9 / Avg: 387.32 / Max: 387.69Min: 387.97 / Avg: 388.43 / Max: 389.351. (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 Inference - Data Type: f32 - Engine: CPU123400800120016002000SE +/- 4.23, N = 3SE +/- 2.61, N = 3SE +/- 2.04, N = 32073.482070.812077.32MIN: 2044.29MIN: 2057.1MIN: 2063.571. (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: CPU123400800120016002000Min: 2065.45 / Avg: 2073.48 / Max: 2079.78Min: 2066.96 / Avg: 2070.81 / Max: 2075.79Min: 2073.34 / Avg: 2077.32 / Max: 2080.091. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread

Ogg Audio Encoding

This test times how long it takes to encode a sample WAV file to Ogg format using the reference Xiph.org tools/libraries. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterOgg Audio Encoding 1.3.4WAV To Ogg12348121620SE +/- 0.01, N = 3SE +/- 0.02, N = 3SE +/- 0.05, N = 318.1118.1118.161. (CC) gcc options: -O2 -ffast-math -fsigned-char
OpenBenchmarking.orgSeconds, Fewer Is BetterOgg Audio Encoding 1.3.4WAV To Ogg123510152025Min: 18.08 / Avg: 18.1 / Max: 18.13Min: 18.09 / Avg: 18.11 / Max: 18.14Min: 18.05 / Avg: 18.16 / Max: 18.231. (CC) gcc options: -O2 -ffast-math -fsigned-char

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 +/- 1.76, N = 3SE +/- 0.67, N = 3SE +/- 1.00, N = 31501150415001. (CXX) g++ options: -O3 -pthread
OpenBenchmarking.orgBenchmark Score, More Is BetterVkFFT 1.1.112330060090012001500Min: 1498 / Avg: 1500.67 / Max: 1504Min: 1503 / Avg: 1503.67 / Max: 1505Min: 1498 / Avg: 1500 / Max: 15011. (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: CPU - Model: squeezenet_ssd123510152025SE +/- 0.04, N = 3SE +/- 0.02, N = 3SE +/- 0.03, N = 318.8818.9318.88MIN: 18.15 / MAX: 19.57MIN: 18.65 / MAX: 29.58MIN: 18.25 / MAX: 19.41. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: squeezenet_ssd123510152025Min: 18.8 / Avg: 18.88 / Max: 18.93Min: 18.9 / Avg: 18.93 / Max: 18.95Min: 18.83 / Avg: 18.88 / Max: 18.921. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: yolov4-tiny123612182430SE +/- 0.02, N = 3SE +/- 0.06, N = 3SE +/- 0.02, N = 325.5625.5725.62MIN: 24.68 / MAX: 26.45MIN: 24.68 / MAX: 26.46MIN: 25.42 / MAX: 26.941. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: yolov4-tiny123612182430Min: 25.53 / Avg: 25.56 / Max: 25.59Min: 25.45 / Avg: 25.57 / Max: 25.63Min: 25.59 / Avg: 25.62 / Max: 25.651. (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: bf16bf16bf16 - Engine: CPU1238001600240032004000SE +/- 1.57, N = 3SE +/- 2.08, N = 3SE +/- 5.09, N = 33763.983763.033771.30MIN: 3751.93MIN: 3749.98MIN: 3754.361. (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: CPU1237001400210028003500Min: 3761.23 / Avg: 3763.98 / Max: 3766.66Min: 3759.43 / Avg: 3763.03 / Max: 3766.64Min: 3763.87 / Avg: 3771.3 / Max: 3781.051. (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: CPU123400800120016002000SE +/- 5.42, N = 3SE +/- 4.58, N = 3SE +/- 1.94, N = 32072.872074.812077.34MIN: 2051.09MIN: 2055.39MIN: 2060.341. (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: CPU123400800120016002000Min: 2062.95 / Avg: 2072.87 / Max: 2081.61Min: 2066.28 / Avg: 2074.81 / Max: 2081.96Min: 2074.94 / Avg: 2077.34 / Max: 2081.181. (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-v21231.10482.20963.31444.41925.524SE +/- 0.03, N = 3SE +/- 0.04, N = 3SE +/- 0.03, N = 34.904.914.90MIN: 4.76 / MAX: 6.22MIN: 4.74 / MAX: 5.86MIN: 4.74 / MAX: 6.251. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU-v2-v2 - Model: mobilenet-v2123246810Min: 4.87 / Avg: 4.9 / Max: 4.95Min: 4.85 / Avg: 4.91 / Max: 4.99Min: 4.87 / Avg: 4.9 / Max: 4.951. (CXX) g++ options: -O3 -rdynamic -lgomp -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 WavPack1233691215SE +/- 0.02, N = 5SE +/- 0.01, N = 5SE +/- 0.01, N = 513.0513.0313.051. (CXX) g++ options: -rdynamic
OpenBenchmarking.orgSeconds, Fewer Is BetterWavPack Audio Encoding 5.3WAV To WavPack12348121620Min: 12.99 / Avg: 13.05 / Max: 13.1Min: 13 / Avg: 13.03 / Max: 13.04Min: 13.04 / Avg: 13.05 / Max: 13.081. (CXX) g++ options: -rdynamic

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: CPU1238001600240032004000SE +/- 1.66, N = 3SE +/- 0.87, N = 3SE +/- 1.70, N = 33758.923760.833765.10MIN: 3744.47MIN: 3750.11MIN: 3745.821. (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: CPU1237001400210028003500Min: 3755.71 / Avg: 3758.92 / Max: 3761.27Min: 3759.38 / Avg: 3760.83 / Max: 3762.4Min: 3761.75 / Avg: 3765.1 / Max: 3767.221. (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: vgg161231326395265SE +/- 0.06, N = 3SE +/- 0.04, N = 3SE +/- 0.05, N = 359.1059.0159.01MIN: 58.72 / MAX: 60.77MIN: 58.7 / MAX: 59.8MIN: 58.68 / MAX: 62.221. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: vgg161231224364860Min: 58.98 / Avg: 59.1 / Max: 59.17Min: 58.94 / Avg: 59.01 / Max: 59.08Min: 58.93 / Avg: 59.01 / Max: 59.11. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: resnet181233691215SE +/- 0.13, N = 3SE +/- 0.11, N = 3SE +/- 0.14, N = 313.1713.1713.19MIN: 12.77 / MAX: 15.6MIN: 12.75 / MAX: 14.11MIN: 12.73 / MAX: 14.21. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: resnet1812348121620Min: 12.92 / Avg: 13.17 / Max: 13.35Min: 12.95 / Avg: 13.17 / Max: 13.3Min: 12.91 / Avg: 13.19 / Max: 13.351. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: mobilenet12348121620SE +/- 0.03, N = 3SE +/- 0.02, N = 3SE +/- 0.03, N = 317.8417.8217.82MIN: 17.55 / MAX: 18.34MIN: 17.61 / MAX: 19.74MIN: 17.52 / MAX: 19.191. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: mobilenet123510152025Min: 17.79 / Avg: 17.84 / Max: 17.87Min: 17.77 / Avg: 17.82 / Max: 17.85Min: 17.76 / Avg: 17.82 / Max: 17.871. (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: u8s8f32 - Engine: CPU1238001600240032004000SE +/- 0.87, N = 3SE +/- 3.77, N = 3SE +/- 1.19, N = 33765.373767.163767.18MIN: 3754.23MIN: 3750.11MIN: 3748.261. (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: CPU1237001400210028003500Min: 3763.78 / Avg: 3765.37 / Max: 3766.76Min: 3760.29 / Avg: 3767.16 / Max: 3773.29Min: 3765.42 / Avg: 3767.18 / Max: 3769.461. (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: CPU1230.71881.43762.15642.87523.594SE +/- 0.04563, N = 3SE +/- 0.03293, N = 8SE +/- 0.03391, N = 73.194673.194033.19453MIN: 2.88MIN: 2.72MIN: 2.71. (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: CPU123246810Min: 3.1 / Avg: 3.19 / Max: 3.24Min: 2.96 / Avg: 3.19 / Max: 3.24Min: 2.99 / Avg: 3.19 / Max: 3.231. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -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 0.7.1Throughput Test: DistinctUserID1230.16880.33760.50640.67520.844SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 30.750.750.751. (CXX) g++ options: -O3 -pthread
OpenBenchmarking.orgGB/s, More Is Bettersimdjson 0.7.1Throughput Test: DistinctUserID123246810Min: 0.75 / Avg: 0.75 / Max: 0.75Min: 0.75 / Avg: 0.75 / Max: 0.75Min: 0.75 / Avg: 0.75 / Max: 0.751. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGB/s, More Is Bettersimdjson 0.7.1Throughput Test: PartialTweets1230.16430.32860.49290.65720.8215SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 30.730.730.731. (CXX) g++ options: -O3 -pthread
OpenBenchmarking.orgGB/s, More Is Bettersimdjson 0.7.1Throughput Test: PartialTweets123246810Min: 0.73 / Avg: 0.73 / Max: 0.73Min: 0.73 / Avg: 0.73 / Max: 0.73Min: 0.73 / Avg: 0.73 / Max: 0.731. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGB/s, More Is Bettersimdjson 0.7.1Throughput Test: LargeRandom1230.11030.22060.33090.44120.5515SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 30.490.490.491. (CXX) g++ options: -O3 -pthread
OpenBenchmarking.orgGB/s, More Is Bettersimdjson 0.7.1Throughput Test: LargeRandom123246810Min: 0.49 / Avg: 0.49 / Max: 0.49Min: 0.49 / Avg: 0.49 / Max: 0.49Min: 0.49 / Avg: 0.49 / Max: 0.491. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGB/s, More Is Bettersimdjson 0.7.1Throughput Test: Kostya1230.180.360.540.720.9SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 30.80.80.81. (CXX) g++ options: -O3 -pthread
OpenBenchmarking.orgGB/s, More Is Bettersimdjson 0.7.1Throughput Test: Kostya123246810Min: 0.8 / Avg: 0.8 / Max: 0.8Min: 0.8 / Avg: 0.8 / Max: 0.8Min: 0.8 / Avg: 0.8 / Max: 0.81. (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: regnety_400m1233691215SE +/- 0.12, N = 3SE +/- 0.63, N = 3SE +/- 0.06, N = 312.6913.4012.41MIN: 12.4 / MAX: 14.98MIN: 12 / MAX: 16.63MIN: 11.89 / MAX: 13.641. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: Vulkan GPU - Model: regnety_400m12348121620Min: 12.45 / Avg: 12.69 / Max: 12.82Min: 12.74 / Avg: 13.4 / Max: 14.65Min: 12.3 / Avg: 12.41 / Max: 12.51. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

63 Results Shown

CLOMP
NCNN:
  Vulkan GPU - yolov4-tiny
  CPU - regnety_400m
Node.js V8 Web Tooling Benchmark
NCNN:
  Vulkan GPU - mnasnet
  Vulkan GPU - blazeface
oneDNN
NCNN:
  Vulkan GPU-v3-v3 - mobilenet-v3
  CPU-v3-v3 - mobilenet-v3
  Vulkan GPU - shufflenet-v2
  Vulkan GPU - resnet50
  Vulkan GPU - squeezenet_ssd
VkResample
oneDNN
NCNN
oneDNN
NCNN:
  CPU - blazeface
  Vulkan GPU - efficientnet-b0
Timed Eigen Compilation
oneDNN:
  Deconvolution Batch shapes_1d - u8s8f32 - CPU
  Matrix Multiply Batch Shapes Transformer - f32 - CPU
Build2
NCNN
oneDNN:
  IP Shapes 1D - f32 - CPU
  Recurrent Neural Network Inference - u8s8f32 - CPU
  Convolution Batch Shapes Auto - u8s8f32 - CPU
  Convolution Batch Shapes Auto - f32 - CPU
NCNN:
  Vulkan GPU-v2-v2 - mobilenet-v2
  CPU - googlenet
oneDNN
NCNN
Opus Codec Encoding
NCNN:
  Vulkan GPU - mobilenet
  CPU - alexnet
  CPU - shufflenet-v2
oneDNN:
  Deconvolution Batch shapes_3d - f32 - CPU
  Deconvolution Batch shapes_1d - f32 - CPU
NCNN
Monkey Audio Encoding
NCNN:
  Vulkan GPU - vgg16
  Vulkan GPU - alexnet
  CPU - efficientnet-b0
VkResample
oneDNN
Ogg Audio Encoding
VkFFT
NCNN:
  CPU - squeezenet_ssd
  CPU - yolov4-tiny
oneDNN:
  Recurrent Neural Network Training - bf16bf16bf16 - CPU
  Recurrent Neural Network Inference - bf16bf16bf16 - CPU
NCNN
WavPack Audio Encoding
oneDNN
NCNN:
  CPU - vgg16
  CPU - resnet18
  CPU - mobilenet
oneDNN:
  Recurrent Neural Network Training - u8s8f32 - CPU
  Matrix Multiply Batch Shapes Transformer - u8s8f32 - CPU
simdjson:
  DistinctUserID
  PartialTweets
  LargeRand
  Kostya
NCNN

1

Compiler Notes: --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 Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xd6 - Thermald 1.9.1
Security Notes: itlb_multihit: KVM: Mitigation of VMX disabled + l1tf: Not affected + mds: Vulnerable; SMT vulnerable + meltdown: Not affected + spec_store_bypass: Vulnerable + spectre_v1: Vulnerable: __user pointer sanitization and usercopy barriers only; no swapgs barriers + spectre_v2: Vulnerable IBPB: disabled STIBP: disabled + srbds: Vulnerable + tsx_async_abort: Vulnerable

Testing initiated at 22 December 2020 05:54 by user phoronix.

2

Compiler Notes: --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 Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xd6 - Thermald 1.9.1
Security Notes: itlb_multihit: KVM: Mitigation of VMX disabled + l1tf: Not affected + mds: Vulnerable; SMT vulnerable + meltdown: Not affected + spec_store_bypass: Vulnerable + spectre_v1: Vulnerable: __user pointer sanitization and usercopy barriers only; no swapgs barriers + spectre_v2: Vulnerable IBPB: disabled STIBP: disabled + srbds: Vulnerable + tsx_async_abort: Vulnerable

Testing initiated at 22 December 2020 09:08 by user phoronix.

3

Processor: Intel Core i9-9900K @ 5.00GHz (8 Cores / 16 Threads), Motherboard: ASRock Z390M Pro4 (P4.20 BIOS), Chipset: Intel Cannon Lake PCH, Memory: 16GB, Disk: 240GB Corsair Force MP510, Graphics: Intel UHD 630 3GB (1200MHz), Audio: Realtek ALC892, Monitor: G237HL, Network: Intel I219-V

OS: Ubuntu 20.04, Kernel: 5.9.0-050900rc1daily20200819-generic (x86_64) 20200818, Desktop: GNOME Shell 3.36.4, Display Server: X Server 1.20.8, Display Driver: modesetting 1.20.8, OpenGL: 4.6 Mesa 20.0.4, OpenCL: OpenCL 2.1, Compiler: GCC 9.3.0, File-System: ext4, Screen Resolution: 1920x1080

Compiler Notes: --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 Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xd6 - Thermald 1.9.1
Security Notes: itlb_multihit: KVM: Mitigation of VMX disabled + l1tf: Not affected + mds: Vulnerable; SMT vulnerable + meltdown: Not affected + spec_store_bypass: Vulnerable + spectre_v1: Vulnerable: __user pointer sanitization and usercopy barriers only; no swapgs barriers + spectre_v2: Vulnerable IBPB: disabled STIBP: disabled + srbds: Vulnerable + tsx_async_abort: Vulnerable

Testing initiated at 22 December 2020 11:53 by user phoronix.