AMD Ryzen 9 3900X 12-Core testing with a ASUS TUF GAMING X570-PLUS (WI-FI) (2203 BIOS) and MSI AMD Radeon RX 470/480/570/570X/580/580X/590 8GB on Ubuntu 20.04 via the Phoronix Test Suite.
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 -vProcessor Notes: Scaling Governor: acpi-cpufreq ondemand - CPU Microcode: 0x8701021Graphics Notes: GLAMORSecurity Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional STIBP: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
2 3 Processor: AMD Ryzen 9 3900X 12-Core @ 3.80GHz (12 Cores / 24 Threads), Motherboard: ASUS TUF GAMING X570-PLUS (WI-FI) (2203 BIOS), Chipset: AMD Starship/Matisse, Memory: 16GB, Disk: Samsung SSD 970 EVO Plus 250GB, Graphics: MSI AMD Radeon RX 470/480/570/570X/580/580X/590 8GB (1366/2000MHz), Audio: AMD Ellesmere HDMI Audio, Monitor: G237HL, Network: Realtek RTL8111/8168/8411 + Intel-AC 9260
OS: Ubuntu 20.04, Kernel: 5.9.0-050900rc6daily20200922-generic (x86_64) 20200921, Desktop: GNOME Shell 3.36.4, Display Server: X Server 1.20.8, Display Driver: modesetting 1.20.8, OpenGL: 4.6 Mesa 20.2.0-devel (git-64cdc13 2020-07-02 focal-oibaf-ppa) (LLVM 10.0.0), Vulkan: 1.2.131, Compiler: GCC 9.3.0, File-System: ext4, Screen Resolution: 1920x1080
3900X New Tests Processor Motherboard Chipset Memory Disk Graphics Audio Monitor Network OS Kernel Desktop Display Server Display Driver OpenGL Vulkan Compiler File-System Screen Resolution 1 2 3 AMD Ryzen 9 3900X 12-Core @ 3.80GHz (12 Cores / 24 Threads) ASUS TUF GAMING X570-PLUS (WI-FI) (2203 BIOS) AMD Starship/Matisse 16GB Samsung SSD 970 EVO Plus 250GB MSI AMD Radeon RX 470/480/570/570X/580/580X/590 8GB (1366/2000MHz) AMD Ellesmere HDMI Audio G237HL Realtek RTL8111/8168/8411 + Intel-AC 9260 Ubuntu 20.04 5.9.0-050900rc6daily20200922-generic (x86_64) 20200921 GNOME Shell 3.36.4 X Server 1.20.8 modesetting 1.20.8 4.6 Mesa 20.2.0-devel (git-64cdc13 2020-07-02 focal-oibaf-ppa) (LLVM 10.0.0) 1.2.131 GCC 9.3.0 ext4 1920x1080 OpenBenchmarking.org Compiler 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: acpi-cpufreq ondemand - CPU Microcode: 0x8701021 Graphics Details - GLAMOR Security Details - itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional STIBP: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
1 2 3 Result Overview Phoronix Test Suite 100% 101% 102% 103% NCNN dcraw OSBench LibRaw TNN eSpeak-NG Speech Engine WebP Image Encode MPV RealSR-NCNN InfluxDB OpenCV
3900X New Tests osbench: Memory Allocations webp: Quality 100, Lossless, Highest Compression osbench: Create Files webp: Quality 100 opencv: Object Detection dcraw: RAW To PPM Image Conversion ncnn: CPU - squeezenet osbench: Create Processes webp: Quality 100, Highest Compression webp: Quality 100, Lossless osbench: Launch Programs ncnn: Vulkan GPU - resnet50 ncnn: Vulkan GPU - squeezenet ncnn: Vulkan GPU-v2-v2 - mobilenet-v2 tnn: CPU - MobileNet v2 webp: Default ncnn: CPU - mobilenet ncnn: Vulkan GPU - efficientnet-b0 ncnn: CPU - yolov4-tiny tnn: CPU - SqueezeNet v1.1 ncnn: Vulkan GPU - resnet18 ncnn: Vulkan GPU - blazeface libraw: Post-Processing Benchmark ncnn: Vulkan GPU - mobilenet ncnn: CPU - resnet18 espeak: Text-To-Speech Synthesis mpv: Big Buck Bunny Sunflower 1080p - Software Only ncnn: CPU-v2-v2 - mobilenet-v2 ncnn: CPU - googlenet ncnn: Vulkan GPU - mnasnet ncnn: CPU - alexnet ncnn: Vulkan GPU - vgg16 ncnn: CPU - vgg16 ncnn: CPU - blazeface ncnn: Vulkan GPU-v3-v3 - mobilenet-v3 ncnn: Vulkan GPU - yolov4-tiny ncnn: CPU - mnasnet ncnn: CPU - shufflenet-v2 ncnn: Vulkan GPU - shufflenet-v2 mpv: Big Buck Bunny Sunflower 4K - Software Only ncnn: CPU - efficientnet-b0 influxdb: 1024 - 10000 - 2,5000,1 - 10000 influxdb: 4 - 10000 - 2,5000,1 - 10000 realsr-ncnn: 4x - Yes realsr-ncnn: 4x - No ncnn: CPU-v3-v3 - mobilenet-v3 influxdb: 64 - 10000 - 2,5000,1 - 10000 ncnn: Vulkan GPU - googlenet ncnn: Vulkan GPU - alexnet opencv: DNN - Deep Neural Network opencv: Features 2D ncnn: CPU - resnet50 osbench: Create Threads 1 2 3 65.431913 32.017 12.166849 2.240 66325 39.505 16.32 33.159256 6.944 15.390 41.739146 7.71 5.71 2.90 253.966 1.455 16.68 11.03 28.70 233.480 2.73 0.93 42.42 6.11 16.40 27.298 2345.26 5.45 17.16 3.06 16.24 15.15 67.17 1.95 4.07 8.61 4.85 4.90 2.49 698.08 6.65 1472018.5 1310447.9 111.642 15.580 4.78 1444861.4 6.32 5.58 4384 143206 81.33 14.062723 68.543672 32.533 12.420693 2.175 65007 39.085 15.93 32.787323 6.938 15.543 40.913423 7.74 5.65 2.93 250.077 1.433 16.43 10.90 28.34 233.491 2.76 0.93 42.32 6.05 16.31 27.316 2334.36 5.41 17.13 3.04 16.31 15.11 66.98 1.96 4.08 8.57 4.83 4.90 2.48 700.67 6.63 1468089.0 1313053.3 111.925 15.605 4.79 1447435.3 6.33 5.58 4354 135808 27.24 12.553533 68.105936 33.093 12.050966 2.223 66757 38.500 16.10 32.373269 6.780 15.204 40.993690 7.85 5.61 2.88 252.154 1.445 16.54 11.06 28.55 230.640 2.73 0.94 42.74 6.06 16.44 27.513 2327.51 5.43 17.25 3.06 16.21 15.06 67.37 1.96 4.09 8.58 4.84 4.88 2.49 697.91 6.63 1471736.4 1313875.7 111.696 15.616 4.78 1445347.6 6.32 5.58 4443 142960 27.69 12.506962 OpenBenchmarking.org
OSBench OSBench is a collection of micro-benchmarks for measuring operating system primitives like time to create threads/processes, launching programs, creating files, and memory allocation. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Ns Per Event, Fewer Is Better OSBench Test: Memory Allocations 1 3 2 15 30 45 60 75 SE +/- 0.18, N = 3 SE +/- 0.22, N = 3 SE +/- 0.10, N = 3 65.43 68.11 68.54 1. (CC) gcc options: -lm
WebP Image Encode This is a test of Google's libwebp with the cwebp image encode utility and using a sample 6000x4000 pixel JPEG image as the input. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Encode Time - Seconds, Fewer Is Better WebP Image Encode 1.1 Encode Settings: Quality 100, Lossless, Highest Compression 1 2 3 8 16 24 32 40 SE +/- 0.16, N = 3 SE +/- 0.42, N = 3 SE +/- 0.29, N = 3 32.02 32.53 33.09 1. (CC) gcc options: -fvisibility=hidden -O2 -pthread -lm -ljpeg -lpng16 -ltiff
OSBench OSBench is a collection of micro-benchmarks for measuring operating system primitives like time to create threads/processes, launching programs, creating files, and memory allocation. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org us Per Event, Fewer Is Better OSBench Test: Create Files 3 1 2 3 6 9 12 15 SE +/- 0.10, N = 3 SE +/- 0.19, N = 3 SE +/- 0.12, N = 3 12.05 12.17 12.42 1. (CC) gcc options: -lm
WebP Image Encode This is a test of Google's libwebp with the cwebp image encode utility and using a sample 6000x4000 pixel JPEG image as the input. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Encode Time - Seconds, Fewer Is Better WebP Image Encode 1.1 Encode Settings: Quality 100 2 3 1 0.504 1.008 1.512 2.016 2.52 SE +/- 0.017, N = 3 SE +/- 0.013, N = 3 SE +/- 0.019, N = 3 2.175 2.223 2.240 1. (CC) gcc options: -fvisibility=hidden -O2 -pthread -lm -ljpeg -lpng16 -ltiff
OpenCV This is a benchmark of the OpenCV (Computer Vision) library's built-in performance tests. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OpenCV 4.4 Test: Object Detection 2 1 3 14K 28K 42K 56K 70K SE +/- 974.24, N = 3 SE +/- 962.41, N = 3 SE +/- 941.51, N = 15 65007 66325 66757 1. (CXX) g++ options: -fPIC -fsigned-char -pthread -fomit-frame-pointer -ffunction-sections -fdata-sections -msse -msse2 -msse3 -fvisibility=hidden -O3 -shared
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.org ms, Fewer Is Better NCNN 20200916 Target: CPU - Model: squeezenet 2 3 1 4 8 12 16 20 SE +/- 0.14, N = 3 SE +/- 0.13, N = 3 SE +/- 0.03, N = 3 15.93 16.10 16.32 MIN: 15.47 / MAX: 19.58 MIN: 15.7 / MAX: 16.75 MIN: 15.64 / MAX: 91.99 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OSBench OSBench is a collection of micro-benchmarks for measuring operating system primitives like time to create threads/processes, launching programs, creating files, and memory allocation. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org us Per Event, Fewer Is Better OSBench Test: Create Processes 3 2 1 8 16 24 32 40 SE +/- 0.27, N = 3 SE +/- 0.20, N = 3 SE +/- 0.49, N = 3 32.37 32.79 33.16 1. (CC) gcc options: -lm
WebP Image Encode This is a test of Google's libwebp with the cwebp image encode utility and using a sample 6000x4000 pixel JPEG image as the input. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Encode Time - Seconds, Fewer Is Better WebP Image Encode 1.1 Encode Settings: Quality 100, Highest Compression 3 2 1 2 4 6 8 10 SE +/- 0.046, N = 3 SE +/- 0.097, N = 4 SE +/- 0.076, N = 3 6.780 6.938 6.944 1. (CC) gcc options: -fvisibility=hidden -O2 -pthread -lm -ljpeg -lpng16 -ltiff
OpenBenchmarking.org Encode Time - Seconds, Fewer Is Better WebP Image Encode 1.1 Encode Settings: Quality 100, Lossless 3 1 2 4 8 12 16 20 SE +/- 0.08, N = 3 SE +/- 0.18, N = 3 SE +/- 0.21, N = 3 15.20 15.39 15.54 1. (CC) gcc options: -fvisibility=hidden -O2 -pthread -lm -ljpeg -lpng16 -ltiff
OSBench OSBench is a collection of micro-benchmarks for measuring operating system primitives like time to create threads/processes, launching programs, creating files, and memory allocation. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org us Per Event, Fewer Is Better OSBench Test: Launch Programs 2 3 1 10 20 30 40 50 SE +/- 0.35, N = 3 SE +/- 0.40, N = 3 SE +/- 0.44, N = 3 40.91 40.99 41.74 1. (CC) gcc options: -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.org ms, Fewer Is Better NCNN 20200916 Target: Vulkan GPU - Model: resnet50 1 2 3 2 4 6 8 10 SE +/- 0.03, N = 7 SE +/- 0.12, N = 3 SE +/- 0.01, N = 3 7.71 7.74 7.85 MIN: 7.36 / MAX: 20.01 MIN: 7.36 / MAX: 20.93 MIN: 7.37 / MAX: 20.26 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: Vulkan GPU - Model: squeezenet 3 2 1 1.2848 2.5696 3.8544 5.1392 6.424 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.06, N = 7 5.61 5.65 5.71 MIN: 5.52 / MAX: 6.15 MIN: 5.52 / MAX: 13.59 MIN: 5.52 / MAX: 49.8 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: Vulkan GPU-v2-v2 - Model: mobilenet-v2 3 1 2 0.6593 1.3186 1.9779 2.6372 3.2965 SE +/- 0.05, N = 3 SE +/- 0.01, N = 7 SE +/- 0.00, N = 3 2.88 2.90 2.93 MIN: 2.68 / MAX: 4.69 MIN: 2.68 / MAX: 3.38 MIN: 2.68 / MAX: 4.09 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
TNN TNN is an open-source deep learning reasoning framework developed by Tencent. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better TNN 0.2.3 Target: CPU - Model: MobileNet v2 2 3 1 60 120 180 240 300 SE +/- 1.56, N = 3 SE +/- 1.60, N = 3 SE +/- 0.28, N = 3 250.08 252.15 253.97 MIN: 242.75 / MAX: 269.72 MIN: 242.11 / MAX: 284.91 MIN: 245.27 / MAX: 273.35 1. (CXX) g++ options: -fopenmp -pthread -fvisibility=hidden -O3 -rdynamic -ldl
WebP Image Encode This is a test of Google's libwebp with the cwebp image encode utility and using a sample 6000x4000 pixel JPEG image as the input. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Encode Time - Seconds, Fewer Is Better WebP Image Encode 1.1 Encode Settings: Default 2 3 1 0.3274 0.6548 0.9822 1.3096 1.637 SE +/- 0.023, N = 3 SE +/- 0.022, N = 3 SE +/- 0.018, N = 4 1.433 1.445 1.455 1. (CC) gcc options: -fvisibility=hidden -O2 -pthread -lm -ljpeg -lpng16 -ltiff
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.org ms, Fewer Is Better NCNN 20200916 Target: CPU - Model: mobilenet 2 3 1 4 8 12 16 20 SE +/- 0.08, N = 3 SE +/- 0.11, N = 3 SE +/- 0.09, N = 3 16.43 16.54 16.68 MIN: 16.12 / MAX: 16.83 MIN: 16.09 / MAX: 41.76 MIN: 16.31 / MAX: 20.34 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: Vulkan GPU - Model: efficientnet-b0 2 1 3 3 6 9 12 15 SE +/- 0.03, N = 3 SE +/- 0.04, N = 7 SE +/- 0.08, N = 3 10.90 11.03 11.06 MIN: 9.76 / MAX: 23.27 MIN: 9.73 / MAX: 24.03 MIN: 9.72 / MAX: 24.39 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: CPU - Model: yolov4-tiny 2 3 1 7 14 21 28 35 SE +/- 0.14, N = 3 SE +/- 0.03, N = 3 SE +/- 0.08, N = 3 28.34 28.55 28.70 MIN: 27.9 / MAX: 29.03 MIN: 28.23 / MAX: 29.25 MIN: 28.44 / MAX: 29.82 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
TNN TNN is an open-source deep learning reasoning framework developed by Tencent. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better TNN 0.2.3 Target: CPU - Model: SqueezeNet v1.1 3 1 2 50 100 150 200 250 SE +/- 0.82, N = 3 SE +/- 0.44, N = 3 SE +/- 0.21, N = 3 230.64 233.48 233.49 MIN: 227.29 / MAX: 233.37 MIN: 231.3 / MAX: 241.85 MIN: 230.13 / MAX: 234.78 1. (CXX) g++ options: -fopenmp -pthread -fvisibility=hidden -O3 -rdynamic -ldl
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.org ms, Fewer Is Better NCNN 20200916 Target: Vulkan GPU - Model: resnet18 1 3 2 0.621 1.242 1.863 2.484 3.105 SE +/- 0.01, N = 7 SE +/- 0.00, N = 3 SE +/- 0.04, N = 3 2.73 2.73 2.76 MIN: 2.62 / MAX: 2.95 MIN: 2.65 / MAX: 3.26 MIN: 2.65 / MAX: 18.85 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: Vulkan GPU - Model: blazeface 1 2 3 0.2115 0.423 0.6345 0.846 1.0575 SE +/- 0.01, N = 7 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 0.93 0.93 0.94 MIN: 0.87 / MAX: 1.52 MIN: 0.87 / MAX: 1.38 MIN: 0.87 / MAX: 1.44 1. (CXX) g++ options: -O3 -rdynamic -lgomp -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.org ms, Fewer Is Better NCNN 20200916 Target: Vulkan GPU - Model: mobilenet 2 3 1 2 4 6 8 10 SE +/- 0.03, N = 3 SE +/- 0.02, N = 3 SE +/- 0.03, N = 7 6.05 6.06 6.11 MIN: 5.98 / MAX: 6.65 MIN: 5.99 / MAX: 6.31 MIN: 6 / MAX: 44.29 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: CPU - Model: resnet18 2 1 3 4 8 12 16 20 SE +/- 0.11, N = 3 SE +/- 0.15, N = 3 SE +/- 0.09, N = 3 16.31 16.40 16.44 MIN: 16.09 / MAX: 19.82 MIN: 16.08 / MAX: 26.36 MIN: 16.16 / MAX: 21.66 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
MPV MPV is an open-source, cross-platform media player. This test profile tests the frame-rate that can be achieved unsynchronized in a desynchronized mode. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org FPS, More Is Better MPV Video Input: Big Buck Bunny Sunflower 1080p - Decode: Software Only 1 2 3 500 1000 1500 2000 2500 SE +/- 6.47, N = 3 SE +/- 11.23, N = 3 SE +/- 10.07, N = 3 2345.26 2334.36 2327.51 MIN: 1333.31 / MAX: 4000 MIN: 1333.35 / MAX: 4000.24 MIN: 1333.32 / MAX: 4000.16 1. mpv 0.32.0
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.org ms, Fewer Is Better NCNN 20200916 Target: CPU-v2-v2 - Model: mobilenet-v2 2 3 1 1.2263 2.4526 3.6789 4.9052 6.1315 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 SE +/- 0.03, N = 3 5.41 5.43 5.45 MIN: 5.3 / MAX: 6.62 MIN: 5.26 / MAX: 7.02 MIN: 5.28 / MAX: 6.75 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: CPU - Model: googlenet 2 1 3 4 8 12 16 20 SE +/- 0.11, N = 3 SE +/- 0.08, N = 3 SE +/- 0.16, N = 3 17.13 17.16 17.25 MIN: 16.63 / MAX: 43.37 MIN: 16.73 / MAX: 20.56 MIN: 16.73 / MAX: 45.28 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: Vulkan GPU - Model: mnasnet 2 1 3 0.6885 1.377 2.0655 2.754 3.4425 SE +/- 0.00, N = 3 SE +/- 0.01, N = 7 SE +/- 0.02, N = 3 3.04 3.06 3.06 MIN: 2.84 / MAX: 4.92 MIN: 2.83 / MAX: 4.89 MIN: 2.83 / MAX: 4.92 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: CPU - Model: alexnet 3 1 2 4 8 12 16 20 SE +/- 0.05, N = 3 SE +/- 0.01, N = 3 SE +/- 0.09, N = 3 16.21 16.24 16.31 MIN: 15.99 / MAX: 17.45 MIN: 16.13 / MAX: 16.37 MIN: 16.09 / MAX: 44.78 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: Vulkan GPU - Model: vgg16 3 2 1 4 8 12 16 20 SE +/- 0.14, N = 3 SE +/- 0.01, N = 3 SE +/- 0.08, N = 7 15.06 15.11 15.15 MIN: 14.67 / MAX: 24 MIN: 14.7 / MAX: 24.03 MIN: 14.65 / MAX: 32 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: CPU - Model: vgg16 2 1 3 15 30 45 60 75 SE +/- 0.36, N = 3 SE +/- 0.13, N = 3 SE +/- 0.28, N = 3 66.98 67.17 67.37 MIN: 65.64 / MAX: 126.21 MIN: 66.36 / MAX: 73.24 MIN: 66.06 / MAX: 73.16 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: CPU - Model: blazeface 1 2 3 0.441 0.882 1.323 1.764 2.205 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 1.95 1.96 1.96 MIN: 1.91 / MAX: 2.38 MIN: 1.93 / MAX: 2.01 MIN: 1.91 / MAX: 2.03 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: Vulkan GPU-v3-v3 - Model: mobilenet-v3 1 2 3 0.9203 1.8406 2.7609 3.6812 4.6015 SE +/- 0.00, N = 7 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 4.07 4.08 4.09 MIN: 3.81 / MAX: 6.15 MIN: 3.81 / MAX: 6.2 MIN: 3.81 / MAX: 6.11 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: Vulkan GPU - Model: yolov4-tiny 2 3 1 2 4 6 8 10 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 7 8.57 8.58 8.61 MIN: 8.51 / MAX: 8.81 MIN: 8.52 / MAX: 18.15 MIN: 8.53 / MAX: 9.25 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: CPU - Model: mnasnet 2 3 1 1.0913 2.1826 3.2739 4.3652 5.4565 SE +/- 0.01, N = 2 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 4.83 4.84 4.85 MIN: 4.77 / MAX: 5.94 MIN: 4.75 / MAX: 6.83 MIN: 4.74 / MAX: 17.84 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: CPU - Model: shufflenet-v2 3 1 2 1.1025 2.205 3.3075 4.41 5.5125 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 4.88 4.90 4.90 MIN: 4.82 / MAX: 6.14 MIN: 4.83 / MAX: 5.5 MIN: 4.82 / MAX: 5.78 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: Vulkan GPU - Model: shufflenet-v2 2 1 3 0.5603 1.1206 1.6809 2.2412 2.8015 SE +/- 0.00, N = 3 SE +/- 0.01, N = 7 SE +/- 0.01, N = 3 2.48 2.49 2.49 MIN: 2.34 / MAX: 3.41 MIN: 2.33 / MAX: 3.42 MIN: 2.33 / MAX: 3.42 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
MPV MPV is an open-source, cross-platform media player. This test profile tests the frame-rate that can be achieved unsynchronized in a desynchronized mode. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org FPS, More Is Better MPV Video Input: Big Buck Bunny Sunflower 4K - Decode: Software Only 2 1 3 150 300 450 600 750 SE +/- 0.52, N = 3 SE +/- 2.78, N = 3 SE +/- 0.55, N = 3 700.67 698.08 697.91 MIN: 444.45 / MAX: 857.17 MIN: 444.45 / MAX: 857.17 MIN: 461.55 / MAX: 857.17 1. mpv 0.32.0
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.org ms, Fewer Is Better NCNN 20200916 Target: CPU - Model: efficientnet-b0 2 3 1 2 4 6 8 10 SE +/- 0.05, N = 3 SE +/- 0.04, N = 3 SE +/- 0.02, N = 3 6.63 6.63 6.65 MIN: 6.51 / MAX: 7.64 MIN: 6.49 / MAX: 7.39 MIN: 6.57 / MAX: 7.57 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
InfluxDB This is a benchmark of the InfluxDB open-source time-series database optimized for fast, high-availability storage for IoT and other use-cases. The InfluxDB test profile makes use of InfluxDB Inch for facilitating the benchmarks. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org val/sec, More Is Better InfluxDB 1.8.2 Concurrent Streams: 1024 - Batch Size: 10000 - Tags: 2,5000,1 - Points Per Series: 10000 1 3 2 300K 600K 900K 1200K 1500K SE +/- 415.31, N = 3 SE +/- 2749.38, N = 3 SE +/- 1895.88, N = 3 1472018.5 1471736.4 1468089.0
OpenBenchmarking.org val/sec, More Is Better InfluxDB 1.8.2 Concurrent Streams: 4 - Batch Size: 10000 - Tags: 2,5000,1 - Points Per Series: 10000 3 2 1 300K 600K 900K 1200K 1500K SE +/- 2904.05, N = 3 SE +/- 1422.20, N = 3 SE +/- 1715.06, N = 3 1313875.7 1313053.3 1310447.9
RealSR-NCNN RealSR-NCNN is an NCNN neural network implementation of the RealSR project and accelerated using the Vulkan API. RealSR is the Real-World Super Resolution via Kernel Estimation and Noise Injection. NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. This test profile times how long it takes to increase the resolution of a sample image by a scale of 4x with Vulkan. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better RealSR-NCNN 20200818 Scale: 4x - TAA: Yes 1 3 2 30 60 90 120 150 SE +/- 0.60, N = 3 SE +/- 0.63, N = 3 SE +/- 0.80, N = 3 111.64 111.70 111.93
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.org ms, Fewer Is Better NCNN 20200916 Target: CPU-v3-v3 - Model: mobilenet-v3 1 3 2 1.0778 2.1556 3.2334 4.3112 5.389 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 4.78 4.78 4.79 MIN: 4.72 / MAX: 6.18 MIN: 4.71 / MAX: 6.85 MIN: 4.72 / MAX: 5.99 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
InfluxDB This is a benchmark of the InfluxDB open-source time-series database optimized for fast, high-availability storage for IoT and other use-cases. The InfluxDB test profile makes use of InfluxDB Inch for facilitating the benchmarks. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org val/sec, More Is Better InfluxDB 1.8.2 Concurrent Streams: 64 - Batch Size: 10000 - Tags: 2,5000,1 - Points Per Series: 10000 2 3 1 300K 600K 900K 1200K 1500K SE +/- 1343.37, N = 3 SE +/- 2565.01, N = 3 SE +/- 3035.14, N = 3 1447435.3 1445347.6 1444861.4
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.org ms, Fewer Is Better NCNN 20200916 Target: Vulkan GPU - Model: googlenet 1 3 2 2 4 6 8 10 SE +/- 0.04, N = 6 SE +/- 0.05, N = 3 SE +/- 0.05, N = 3 6.32 6.32 6.33 MIN: 6.09 / MAX: 18.95 MIN: 6.11 / MAX: 20.01 MIN: 6.1 / MAX: 21.22 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20200916 Target: Vulkan GPU - Model: alexnet 1 2 3 1.2555 2.511 3.7665 5.022 6.2775 SE +/- 0.02, N = 7 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 5.58 5.58 5.58 MIN: 5.46 / MAX: 15.53 MIN: 5.48 / MAX: 9.56 MIN: 5.48 / MAX: 10.49 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenCV This is a benchmark of the OpenCV (Computer Vision) library's built-in performance tests. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OpenCV 4.4 Test: DNN - Deep Neural Network 2 1 3 1000 2000 3000 4000 5000 SE +/- 147.14, N = 12 SE +/- 189.01, N = 15 SE +/- 135.75, N = 15 4354 4384 4443 1. (CXX) g++ options: -fPIC -fsigned-char -pthread -fomit-frame-pointer -ffunction-sections -fdata-sections -msse -msse2 -msse3 -fvisibility=hidden -O3 -shared
OpenBenchmarking.org ms, Fewer Is Better OpenCV 4.4 Test: Features 2D 2 3 1 30K 60K 90K 120K 150K SE +/- 1818.45, N = 12 SE +/- 2214.44, N = 12 SE +/- 2677.01, N = 12 135808 142960 143206 1. (CXX) g++ options: -fPIC -fsigned-char -pthread -fomit-frame-pointer -ffunction-sections -fdata-sections -msse -msse2 -msse3 -fvisibility=hidden -O3 -shared
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.org ms, Fewer Is Better NCNN 20200916 Target: CPU - Model: resnet50 2 3 1 20 40 60 80 100 SE +/- 0.08, N = 3 SE +/- 0.22, N = 3 SE +/- 53.67, N = 3 27.24 27.69 81.33 MIN: 26.91 / MAX: 29.33 MIN: 27.07 / MAX: 28.69 MIN: 27.13 / MAX: 4085.39 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OSBench OSBench is a collection of micro-benchmarks for measuring operating system primitives like time to create threads/processes, launching programs, creating files, and memory allocation. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org us Per Event, Fewer Is Better OSBench Test: Create Threads 3 2 1 4 8 12 16 20 SE +/- 0.38, N = 15 SE +/- 0.37, N = 15 SE +/- 0.09, N = 3 12.51 12.55 14.06 1. (CC) gcc options: -lm
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 -vProcessor Notes: Scaling Governor: acpi-cpufreq ondemand - CPU Microcode: 0x8701021Graphics Notes: GLAMORSecurity Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional STIBP: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 25 September 2020 08:03 by user pts.
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 -vProcessor Notes: Scaling Governor: acpi-cpufreq ondemand - CPU Microcode: 0x8701021Graphics Notes: GLAMORSecurity Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional STIBP: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 25 September 2020 10:27 by user pts.
3 Processor: AMD Ryzen 9 3900X 12-Core @ 3.80GHz (12 Cores / 24 Threads), Motherboard: ASUS TUF GAMING X570-PLUS (WI-FI) (2203 BIOS), Chipset: AMD Starship/Matisse, Memory: 16GB, Disk: Samsung SSD 970 EVO Plus 250GB, Graphics: MSI AMD Radeon RX 470/480/570/570X/580/580X/590 8GB (1366/2000MHz), Audio: AMD Ellesmere HDMI Audio, Monitor: G237HL, Network: Realtek RTL8111/8168/8411 + Intel-AC 9260
OS: Ubuntu 20.04, Kernel: 5.9.0-050900rc6daily20200922-generic (x86_64) 20200921, Desktop: GNOME Shell 3.36.4, Display Server: X Server 1.20.8, Display Driver: modesetting 1.20.8, OpenGL: 4.6 Mesa 20.2.0-devel (git-64cdc13 2020-07-02 focal-oibaf-ppa) (LLVM 10.0.0), Vulkan: 1.2.131, 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 -vProcessor Notes: Scaling Governor: acpi-cpufreq ondemand - CPU Microcode: 0x8701021Graphics Notes: GLAMORSecurity Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional STIBP: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 25 September 2020 12:11 by user pts.