Ryzen 9 5950X Clang 12 vs. GCC 11 Benchmarks

GCC 11.1 versus LLVM Clang 12 on AMD Ryzen 9 5950X. Benchmarks by Michael Larabel for a future article.

HTML result view exported from: https://openbenchmarking.org/result/2105198-IB-11900KCOM08&grt&sor.

ProcessorMotherboardChipsetMemoryDiskGraphicsAudioMonitorNetworkOSKernelDesktopDisplay ServerOpenGLCompilerFile-SystemScreen ResolutionGCC 11.1Clang 12 -O2 -O3 -march=native -O3 -march=native -flto -O2 -O3 -march=native -O3 -march=native -fltoAMD Ryzen 9 5950X 16-Core @ 3.40GHz (16 Cores / 32 Threads)ASUS ROG CROSSHAIR VIII HERO (WI-FI) (3302 BIOS)AMD Starship/Matisse32GB500GB Western Digital WDS500G3X0C-00SJG0AMD Radeon RX 5600 OEM/5600 XT / 5700/5700 8GB (2100/875MHz)AMD Navi 10 HDMI AudioASUS MG28URealtek RTL8125 2.5GbE + Intel I211 + Intel Wi-Fi 6 AX200Fedora 345.11.20-300.fc34.x86_64 (x86_64)GNOME Shell 40.1X Server + Wayland4.6 Mesa 21.0.3 (LLVM 12.0.0)GCC 11.1.1 20210428 + Clang 12.0.0btrfs3840x2160Clang 12.0.0OpenBenchmarking.orgKernel Details- Transparent Huge Pages: madviseEnvironment Details- GCC 11.1: -O2: CXXFLAGS=-O2 CFLAGS=-O2- GCC 11.1: -O3 -march=native: CXXFLAGS="-O3 -march=native" CFLAGS="-O3 -march=native"- GCC 11.1: -O3 -march=native -flto: CXXFLAGS="-O3 -march=native -flto" CFLAGS="-O3 -march=native -flto"- Clang 12: -O2: CXXFLAGS=-O2 CFLAGS=-O2- Clang 12: -O3 -march=native: CXXFLAGS="-O3 -march=native" CFLAGS="-O3 -march=native"- Clang 12: -O3 -march=native -flto: CXXFLAGS="-O3 -march=native -flto" CFLAGS="-O3 -march=native -flto"Compiler Details- GCC 11.1: -O2, GCC 11.1: -O3 -march=native, GCC 11.1: -O3 -march=native -flto: --build=x86_64-redhat-linux --disable-libunwind-exceptions --enable-__cxa_atexit --enable-bootstrap --enable-cet --enable-checking=release --enable-gnu-indirect-function --enable-gnu-unique-object --enable-initfini-array --enable-languages=c,c++,fortran,objc,obj-c++,ada,go,d,lto --enable-multilib --enable-offload-targets=nvptx-none --enable-plugin --enable-shared --enable-threads=posix --mandir=/usr/share/man --with-arch_32=i686 --with-gcc-major-version-only --with-linker-hash-style=gnu --with-tune=generic --without-cuda-driver Processor Details- Scaling Governor: acpi-cpufreq schedutil (Boost: Enabled) - CPU Microcode: 0xa201009Security Details- SELinux + 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 IBRS_FW STIBP: always-on RSB filling + srbds: Not affected + tsx_async_abort: Not affected

aobench: 2048 x 2048 - Total Timeastcenc: Mediumastcenc: Thoroughastcenc: Exhaustivec-ray: Total Time - 4K, 16 Rays Per Pixelcoremark: CoreMark Size 666 - Iterations Per Secondencode-flac: WAV To FLACgraphics-magick: Rotategraphics-magick: Sharpengraphics-magick: Enhancedgraphics-magick: Resizinghimeno: Poisson Pressure Solverencode-mp3: WAV To MP3lammps: Rhodopsin Proteintjbench: Decompression Throughputliquid-dsp: 8 - 256 - 57liquid-dsp: 16 - 256 - 57ncnn: CPU - mobilenetncnn: CPU-v2-v2 - mobilenet-v2ncnn: CPU-v3-v3 - mobilenet-v3ncnn: CPU - mnasnetncnn: CPU - efficientnet-b0ncnn: CPU - googlenetncnn: CPU - vgg16ncnn: CPU - resnet18ncnn: CPU - alexnetncnn: CPU - resnet50ncnn: CPU - squeezenet_ssdncnn: CPU - regnety_400mncnn: CPU - shufflenet-v2ncnn: CPU - blazefacencnn: CPU - yolov4-tinyencode-opus: WAV To Opus Encodepjsip: INVITEpjsip: OPTIONS, Statefulpjsip: OPTIONS, Statelesssqlite-speedtest: Timed Time - Size 1,000svt-hevc: 7 - Bosphorus 1080psvt-hevc: 10 - Bosphorus 1080psvt-vp9: VMAF Optimized - Bosphorus 1080psvt-vp9: PSNR/SSIM Optimized - Bosphorus 1080psvt-vp9: Visual Quality Optimized - Bosphorus 1080phmmer: Pfam Database Searchmrbayes: Primate Phylogeny Analysistnn: CPU - MobileNet v2tnn: CPU - SqueezeNet v1.1webp: Quality 100, Losslesswebp: Quality 100, Highest Compressionwebp: Quality 100, Lossless, Highest Compressionx265: Bosphorus 4KGCC 11.1Clang 12 -O2 -O3 -march=native -O3 -march=native -flto -O2 -O3 -march=native -O3 -march=native -flto31.7324.62967.684157.144560.931830811.5321165.87498522642218035120.9402626.80212.855277.496802594850000104120000012.954.544.153.975.4113.1556.8214.5011.1626.1814.6117.244.331.8521.066.5154815786023059947.404217.32365.79230.53235.27222.6399.72296.785219.130208.78013.1815.35527.42426.2425.4794.62777.571856.641125.405808580.8626506.237103637044921205314.7339045.50312.749268.113456611560000108500000012.704.433.843.955.3712.8557.6114.3310.9925.3313.9317.184.391.8021.005.3874671798222175946.917221.71369.54231.16235.59223.1197.74492.973227.768214.27713.5505.20028.32725.9126.3124.69957.529956.190025.482849671.9594426.19596438244920855445.6352215.40312.813270.909150619393333109326666713.744.463.863.925.3613.2657.8414.8011.1325.8414.5018.536.012.5523.945.4754616791622258146.479227.04368.55230.43234.70222.3397.71493.337215.341202.57213.2545.18427.94926.1733.7783.33679.395052.204849.275712758.3578267.68792120641116934877.3064106.56513.087267.553068576660000105640000011.583.563.363.264.5211.8857.2614.0511.0922.9912.6112.933.811.5421.975.8744575794222110747.792225.15368.55231.35235.64225.5796.09293.529271.028208.55513.6474.77829.11927.7730.2343.31799.354951.823045.010722694.0062755.66698123545217675192.1900426.09813.299272.513559576050000106190000011.463.573.073.264.5111.7756.9313.9011.1623.2512.3512.593.741.5221.735.5084708795822257247.630229.27375.95234.71239.88227.7395.21490.785355.925215.78114.1054.70328.64027.7529.7303.37289.443251.897444.742714933.8589835.75696123745117625289.9175715.81013.342266.229829587800000107346666711.053.422.933.084.3011.6256.1113.8611.2522.9012.4312.143.691.4621.585.55848.584236.05375.08232.66238.13227.6693.31192.405261.836207.92513.6364.72227.40028.47OpenBenchmarking.org

AOBench

Size: 2048 x 2048 - Total Time

GCC 11.1Clang 12OpenBenchmarking.orgSeconds, Fewer Is BetterAOBenchSize: 2048 x 2048 - Total Time-O3 -march=native-O3 -march=native -flto-O2816243240SE +/- 0.02, N = 3SE +/- 0.01, N = 3SE +/- 0.14, N = 3SE +/- 0.09, N = 3SE +/- 0.43, N = 3SE +/- 0.04, N = 325.4826.3131.7329.7330.2333.781. (CC) gcc options: -lm -O3

ASTC Encoder

Preset: Medium

Clang 12GCC 11.1OpenBenchmarking.orgSeconds, Fewer Is BetterASTC Encoder 2.4Preset: Medium-O3 -march=native-O2-O3 -march=native -flto1.05742.11483.17224.22965.287SE +/- 0.0243, N = 3SE +/- 0.0156, N = 3SE +/- 0.0302, N = 3SE +/- 0.0292, N = 3SE +/- 0.0337, N = 3SE +/- 0.0045, N = 33.31793.33673.37284.62774.62964.69951. (CXX) g++ options: -O2 -flto -pthread

ASTC Encoder

Preset: Thorough

GCC 11.1Clang 12OpenBenchmarking.orgSeconds, Fewer Is BetterASTC Encoder 2.4Preset: Thorough-O3 -march=native -flto-O3 -march=native-O23691215SE +/- 0.0109, N = 3SE +/- 0.0351, N = 3SE +/- 0.0109, N = 3SE +/- 0.0407, N = 3SE +/- 0.0279, N = 3SE +/- 0.0114, N = 37.52997.57187.68419.35499.39509.44321. (CXX) g++ options: -flto -O2 -pthread

ASTC Encoder

Preset: Exhaustive

Clang 12GCC 11.1OpenBenchmarking.orgSeconds, Fewer Is BetterASTC Encoder 2.4Preset: Exhaustive-O3 -march=native-O3 -march=native -flto-O21326395265SE +/- 0.11, N = 3SE +/- 0.07, N = 3SE +/- 0.11, N = 3SE +/- 0.10, N = 3SE +/- 0.08, N = 3SE +/- 0.09, N = 351.8251.9052.2056.1956.6457.141. (CXX) g++ options: -O2 -flto -pthread

C-Ray

Total Time - 4K, 16 Rays Per Pixel

GCC 11.1Clang 12OpenBenchmarking.orgSeconds, Fewer Is BetterC-Ray 1.1Total Time - 4K, 16 Rays Per Pixel-O3 -march=native-O3 -march=native -flto-O21428425670SE +/- 0.08, N = 3SE +/- 0.08, N = 3SE +/- 0.18, N = 3SE +/- 0.17, N = 3SE +/- 0.10, N = 3SE +/- 0.16, N = 325.4125.4860.9344.7445.0149.281. (CC) gcc options: -lm -lpthread -O3

Coremark

CoreMark Size 666 - Iterations Per Second

GCC 11.1Clang 12OpenBenchmarking.orgIterations/Sec, More Is BetterCoremark 1.0CoreMark Size 666 - Iterations Per Second-O3 -march=native -flto-O2-O3 -march=native200K400K600K800K1000KSE +/- 836.68, N = 3SE +/- 989.68, N = 3SE +/- 2230.96, N = 3SE +/- 845.83, N = 3SE +/- 581.76, N = 3SE +/- 789.22, N = 3849671.96830811.53808580.86722694.01714933.86712758.361. (CC) gcc options: -O2 -lrt" -lrt

FLAC Audio Encoding

WAV To FLAC

Clang 12GCC 11.1OpenBenchmarking.orgSeconds, Fewer Is BetterFLAC Audio Encoding 1.3.2WAV To FLAC-O3 -march=native-O3 -march=native -flto-O2246810SE +/- 0.005, N = 5SE +/- 0.030, N = 5SE +/- 0.024, N = 5SE +/- 0.045, N = 5SE +/- 0.036, N = 5SE +/- 0.020, N = 55.6665.7567.6875.8746.1956.2371. (CXX) g++ options: -logg -lm

GraphicsMagick

Operation: Rotate

GCC 11.1Clang 12OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.33Operation: Rotate-O3 -march=native-O2-O3 -march=native -flto2004006008001000SE +/- 5.57, N = 3SE +/- 11.70, N = 3SE +/- 3.48, N = 3SE +/- 0.58, N = 3SE +/- 4.37, N = 3SE +/- 5.55, N = 310369859649819619211. (CC) gcc options: -fopenmp -pthread -ljpeg -lz -lm -lpthread

GraphicsMagick

Operation: Sharpen

GCC 11.1Clang 12OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.33Operation: Sharpen-O3 -march=native -flto-O3 -march=native-O280160240320400SE +/- 0.88, N = 3SE +/- 1.20, N = 3SE +/- 0.33, N = 3SE +/- 0.67, N = 3SE +/- 0.33, N = 3SE +/- 0.58, N = 33823702262372352061. (CC) gcc options: -fopenmp -pthread -ljpeg -lz -lm -lpthread

GraphicsMagick

Operation: Enhanced

Clang 12GCC 11.1OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.33Operation: Enhanced-O3 -march=native-O3 -march=native -flto-O2100200300400500SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.33, N = 34524514114494494221. (CC) gcc options: -fopenmp -pthread -ljpeg -lz -lm -lpthread

GraphicsMagick

Operation: Resizing

GCC 11.1Clang 12OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.33Operation: Resizing-O3 -march=native-O3 -march=native -flto-O25001000150020002500SE +/- 1.53, N = 3SE +/- 3.71, N = 3SE +/- 0.88, N = 3SE +/- 3.61, N = 3SE +/- 2.85, N = 3SE +/- 4.48, N = 32120208518031767176216931. (CC) gcc options: -fopenmp -pthread -ljpeg -lz -lm -lpthread

Himeno Benchmark

Poisson Pressure Solver

GCC 11.1Clang 12OpenBenchmarking.orgMFLOPS, More Is BetterHimeno Benchmark 3.0Poisson Pressure Solver-O3 -march=native -flto-O3 -march=native-O212002400360048006000SE +/- 68.97, N = 15SE +/- 90.91, N = 15SE +/- 83.12, N = 15SE +/- 96.57, N = 13SE +/- 66.92, N = 15SE +/- 62.44, N = 155445.645314.735120.945289.925192.194877.311. (CC) gcc options: -O3 -mavx2

LAME MP3 Encoding

WAV To MP3

GCC 11.1Clang 12OpenBenchmarking.orgSeconds, Fewer Is BetterLAME MP3 Encoding 3.100WAV To MP3-O3 -march=native -flto-O3 -march=native-O2246810SE +/- 0.010, N = 3SE +/- 0.071, N = 3SE +/- 0.084, N = 3SE +/- 0.056, N = 3SE +/- 0.045, N = 3SE +/- 0.068, N = 45.4035.5036.8025.8106.0986.5651. (CC) gcc options: -O3 -pipe -lm

LAMMPS Molecular Dynamics Simulator

Model: Rhodopsin Protein

Clang 12GCC 11.1OpenBenchmarking.orgns/day, More Is BetterLAMMPS Molecular Dynamics Simulator 29Oct2020Model: Rhodopsin Protein-O3 -march=native -flto-O3 -march=native-O23691215SE +/- 0.01, N = 3SE +/- 0.02, N = 3SE +/- 0.08, N = 3SE +/- 0.11, N = 8SE +/- 0.09, N = 15SE +/- 0.14, N = 413.3413.3013.0912.8612.8112.751. (CXX) g++ options: -O2 -pthread -lm

libjpeg-turbo tjbench

Test: Decompression Throughput

GCC 11.1Clang 12OpenBenchmarking.orgMegapixels/sec, More Is Betterlibjpeg-turbo tjbench 2.1.0Test: Decompression Throughput-O2-O3 -march=native-O3 -march=native -flto60120180240300SE +/- 1.92, N = 3SE +/- 0.45, N = 3SE +/- 1.27, N = 3SE +/- 0.59, N = 3SE +/- 1.65, N = 3SE +/- 0.54, N = 3277.50270.91268.11272.51267.55266.231. (CC) gcc options: -O3 -rdynamic

Liquid-DSP

Threads: 8 - Buffer Length: 256 - Filter Length: 57

GCC 11.1Clang 12OpenBenchmarking.orgsamples/s, More Is BetterLiquid-DSP 2021.01.31Threads: 8 - Buffer Length: 256 - Filter Length: 57-O3 -march=native -flto-O3 -march=native-O2130M260M390M520M650MSE +/- 1343755.10, N = 3SE +/- 4455270.29, N = 3SE +/- 185202.59, N = 3SE +/- 1855020.22, N = 3SE +/- 1824399.44, N = 3SE +/- 4072558.57, N = 36193933336115600005948500005878000005766600005760500001. (CC) gcc options: -O3 -pthread -lm -lc -lliquid

Liquid-DSP

Threads: 16 - Buffer Length: 256 - Filter Length: 57

GCC 11.1Clang 12OpenBenchmarking.orgsamples/s, More Is BetterLiquid-DSP 2021.01.31Threads: 16 - Buffer Length: 256 - Filter Length: 57-O3 -march=native -flto-O3 -march=native-O2200M400M600M800M1000MSE +/- 783865.07, N = 3SE +/- 1252996.41, N = 3SE +/- 4650806.38, N = 3SE +/- 3219903.38, N = 3SE +/- 493288.29, N = 3SE +/- 2690724.81, N = 31093266667108500000010412000001073466667106190000010564000001. (CC) gcc options: -O3 -pthread -lm -lc -lliquid

NCNN

Target: CPU - Model: mobilenet

Clang 12GCC 11.1OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: mobilenet-O3 -march=native -flto-O3 -march=native-O248121620SE +/- 0.12, N = 3SE +/- 0.11, N = 3SE +/- 0.03, N = 3SE +/- 0.13, N = 15SE +/- 0.11, N = 15SE +/- 0.17, N = 311.0511.4611.5812.7012.9513.741. (CXX) g++ options: -O2 -rdynamic -lgomp -lpthread

NCNN

Target: CPU-v2-v2 - Model: mobilenet-v2

Clang 12GCC 11.1OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU-v2-v2 - Model: mobilenet-v2-O3 -march=native -flto-O2-O3 -march=native1.02152.0433.06454.0865.1075SE +/- 0.08, N = 3SE +/- 0.11, N = 3SE +/- 0.04, N = 3SE +/- 0.01, N = 15SE +/- 0.01, N = 3SE +/- 0.04, N = 153.423.563.574.434.464.541. (CXX) g++ options: -O2 -rdynamic -lgomp -lpthread

NCNN

Target: CPU-v3-v3 - Model: mobilenet-v3

Clang 12GCC 11.1OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU-v3-v3 - Model: mobilenet-v3-O3 -march=native -flto-O3 -march=native-O20.93381.86762.80143.73524.669SE +/- 0.01, N = 3SE +/- 0.05, N = 3SE +/- 0.07, N = 3SE +/- 0.02, N = 15SE +/- 0.03, N = 3SE +/- 0.03, N = 152.933.073.363.843.864.151. (CXX) g++ options: -O2 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: mnasnet

Clang 12GCC 11.1OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: mnasnet-O3 -march=native -flto-O2-O3 -march=native0.89331.78662.67993.57324.4665SE +/- 0.04, N = 3SE +/- 0.08, N = 3SE +/- 0.05, N = 3SE +/- 0.02, N = 3SE +/- 0.04, N = 15SE +/- 0.04, N = 153.083.263.263.923.953.971. (CXX) g++ options: -O2 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: efficientnet-b0

Clang 12GCC 11.1OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: efficientnet-b0-O3 -march=native -flto-O3 -march=native-O21.21732.43463.65194.86926.0865SE +/- 0.03, N = 3SE +/- 0.01, N = 3SE +/- 0.06, N = 3SE +/- 0.02, N = 3SE +/- 0.03, N = 15SE +/- 0.05, N = 154.304.514.525.365.375.411. (CXX) g++ options: -O2 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: googlenet

Clang 12GCC 11.1OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: googlenet-O3 -march=native -flto-O3 -march=native-O23691215SE +/- 0.09, N = 3SE +/- 0.16, N = 3SE +/- 0.13, N = 3SE +/- 0.12, N = 15SE +/- 0.22, N = 15SE +/- 0.44, N = 311.6211.7711.8812.8513.1513.261. (CXX) g++ options: -O2 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: vgg16

Clang 12GCC 11.1OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: vgg16-O3 -march=native -flto-O2-O3 -march=native1326395265SE +/- 0.45, N = 3SE +/- 0.53, N = 3SE +/- 0.08, N = 3SE +/- 0.08, N = 15SE +/- 0.08, N = 15SE +/- 0.23, N = 356.1156.9357.2656.8257.6157.841. (CXX) g++ options: -O2 -rdynamic -lpthread

NCNN

Target: CPU - Model: resnet18

Clang 12GCC 11.1OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: resnet18-O3 -march=native -flto-O3 -march=native-O248121620SE +/- 0.13, N = 3SE +/- 0.13, N = 3SE +/- 0.23, N = 3SE +/- 0.09, N = 15SE +/- 0.11, N = 15SE +/- 0.22, N = 313.8613.9014.0514.3314.5014.801. (CXX) g++ options: -O2 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: alexnet

GCC 11.1Clang 12OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: alexnet-O3 -march=native-O2-O3 -march=native -flto3691215SE +/- 0.04, N = 15SE +/- 0.06, N = 3SE +/- 0.06, N = 15SE +/- 0.03, N = 3SE +/- 0.05, N = 3SE +/- 0.10, N = 310.9911.1311.1611.0911.1611.251. (CXX) g++ options: -O2 -rdynamic -lpthread

NCNN

Target: CPU - Model: resnet50

Clang 12GCC 11.1OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: resnet50-O3 -march=native -flto-O2-O3 -march=native612182430SE +/- 0.09, N = 3SE +/- 0.11, N = 3SE +/- 0.06, N = 3SE +/- 0.10, N = 15SE +/- 0.21, N = 3SE +/- 0.18, N = 1522.9022.9923.2525.3325.8426.181. (CXX) g++ options: -O2 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: squeezenet_ssd

Clang 12GCC 11.1OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: squeezenet_ssd-O3 -march=native-O3 -march=native -flto-O248121620SE +/- 0.20, N = 3SE +/- 0.20, N = 3SE +/- 0.31, N = 3SE +/- 0.09, N = 15SE +/- 0.06, N = 3SE +/- 0.10, N = 1512.3512.4312.6113.9314.5014.611. (CXX) g++ options: -O2 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: regnety_400m

Clang 12GCC 11.1OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: regnety_400m-O3 -march=native -flto-O3 -march=native-O2510152025SE +/- 0.14, N = 3SE +/- 0.16, N = 3SE +/- 0.20, N = 3SE +/- 0.05, N = 15SE +/- 0.06, N = 15SE +/- 0.09, N = 312.1412.5912.9317.1817.2418.531. (CXX) g++ options: -O2 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: shufflenet-v2

Clang 12GCC 11.1OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: shufflenet-v2-O3 -march=native -flto-O3 -march=native-O2246810SE +/- 0.02, N = 3SE +/- 0.05, N = 3SE +/- 0.05, N = 3SE +/- 0.01, N = 14SE +/- 0.15, N = 15SE +/- 0.02, N = 33.693.743.814.334.396.011. (CXX) g++ options: -O2 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: blazeface

Clang 12GCC 11.1OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: blazeface-O3 -march=native -flto-O3 -march=native-O20.57381.14761.72142.29522.869SE +/- 0.05, N = 3SE +/- 0.04, N = 3SE +/- 0.04, N = 3SE +/- 0.01, N = 15SE +/- 0.01, N = 15SE +/- 0.01, N = 31.461.521.541.801.852.551. (CXX) g++ options: -O2 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: yolov4-tiny

GCC 11.1Clang 12OpenBenchmarking.orgms, Fewer Is BetterNCNN 20201218Target: CPU - Model: yolov4-tiny-O3 -march=native-O2-O3 -march=native -flto612182430SE +/- 0.14, N = 15SE +/- 0.11, N = 15SE +/- 0.41, N = 3SE +/- 0.10, N = 3SE +/- 0.12, N = 3SE +/- 0.18, N = 321.0021.0623.9421.5821.7321.971. (CXX) g++ options: -O2 -rdynamic -lpthread

Opus Codec Encoding

WAV To Opus Encode

GCC 11.1Clang 12OpenBenchmarking.orgSeconds, Fewer Is BetterOpus Codec Encoding 1.3.1WAV To Opus Encode-O3 -march=native-O3 -march=native -flto-O2246810SE +/- 0.048, N = 5SE +/- 0.023, N = 5SE +/- 0.035, N = 5SE +/- 0.045, N = 5SE +/- 0.016, N = 5SE +/- 0.052, N = 55.3875.4756.5155.5085.5585.8741. (CXX) g++ options: -logg -lm

PJSIP

Method: INVITE

GCC 11.1Clang 12OpenBenchmarking.orgResponses Per Second, More Is BetterPJSIP 2.11Method: INVITE-O2-O3 -march=native-O3 -march=native -flto10002000300040005000SE +/- 16.33, N = 3SE +/- 32.67, N = 15SE +/- 65.77, N = 3SE +/- 66.36, N = 3SE +/- 42.78, N = 3481546714616470845751. (CC) gcc options: -lstdc++ -lssl -lcrypto -lm -lrt -lpthread

PJSIP

Method: OPTIONS, Stateful

GCC 11.1Clang 12OpenBenchmarking.orgResponses Per Second, More Is BetterPJSIP 2.11Method: OPTIONS, Stateful-O3 -march=native-O2-O3 -march=native -flto2K4K6K8K10KSE +/- 21.79, N = 3SE +/- 56.75, N = 3SE +/- 72.34, N = 3SE +/- 47.54, N = 3SE +/- 27.78, N = 3798279167860795879421. (CC) gcc options: -lstdc++ -lssl -lcrypto -lm -lrt -lpthread

PJSIP

Method: OPTIONS, Stateless

GCC 11.1Clang 12OpenBenchmarking.orgResponses Per Second, More Is BetterPJSIP 2.11Method: OPTIONS, Stateless-O2-O3 -march=native -flto-O3 -march=native50K100K150K200K250KSE +/- 586.86, N = 3SE +/- 705.50, N = 3SE +/- 1749.48, N = 3SE +/- 962.04, N = 3SE +/- 1827.48, N = 32305992225812217592225722211071. (CC) gcc options: -lstdc++ -lssl -lcrypto -lm -lrt -lpthread

SQLite Speedtest

Timed Time - Size 1,000

GCC 11.1Clang 12OpenBenchmarking.orgSeconds, Fewer Is BetterSQLite Speedtest 3.30Timed Time - Size 1,000-O3 -march=native -flto-O3 -march=native-O21122334455SE +/- 0.54, N = 3SE +/- 0.16, N = 3SE +/- 0.02, N = 3SE +/- 0.50, N = 3SE +/- 0.13, N = 3SE +/- 0.18, N = 346.4846.9247.4047.6347.7948.581. (CC) gcc options: -ldl -lz -lpthread

SVT-HEVC

Tuning: 7 - Input: Bosphorus 1080p

Clang 12GCC 11.1OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-HEVC 1.5.0Tuning: 7 - Input: Bosphorus 1080p-O3 -march=native -flto-O3 -march=native-O250100150200250SE +/- 1.15, N = 3SE +/- 0.32, N = 3SE +/- 0.95, N = 3SE +/- 0.03, N = 3SE +/- 0.83, N = 3SE +/- 0.49, N = 3236.05229.27225.15227.04221.71217.321. (CC) gcc options: -O3 -fPIE -fPIC -O2 -pie -rdynamic -lpthread -lrt

SVT-HEVC

Tuning: 10 - Input: Bosphorus 1080p

Clang 12GCC 11.1OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-HEVC 1.5.0Tuning: 10 - Input: Bosphorus 1080p-O3 -march=native-O3 -march=native -flto-O280160240320400SE +/- 1.66, N = 3SE +/- 0.90, N = 3SE +/- 0.23, N = 3SE +/- 1.07, N = 3SE +/- 0.57, N = 3SE +/- 1.24, N = 3375.95375.08368.55369.54368.55365.791. (CC) gcc options: -O3 -fPIE -fPIC -O2 -pie -rdynamic -lpthread -lrt

SVT-VP9

Tuning: VMAF Optimized - Input: Bosphorus 1080p

Clang 12GCC 11.1OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-VP9 0.3Tuning: VMAF Optimized - Input: Bosphorus 1080p-O3 -march=native-O3 -march=native -flto-O250100150200250SE +/- 0.32, N = 3SE +/- 0.70, N = 3SE +/- 0.33, N = 3SE +/- 0.46, N = 3SE +/- 0.09, N = 3SE +/- 0.18, N = 3234.71232.66231.35231.16230.53230.431. (CC) gcc options: -O3 -fcommon -fPIE -fPIC -fvisibility=hidden -O2 -pie -rdynamic -lpthread -lrt -lm

SVT-VP9

Tuning: PSNR/SSIM Optimized - Input: Bosphorus 1080p

Clang 12GCC 11.1OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-VP9 0.3Tuning: PSNR/SSIM Optimized - Input: Bosphorus 1080p-O3 -march=native-O3 -march=native -flto-O250100150200250SE +/- 0.40, N = 3SE +/- 0.48, N = 3SE +/- 0.18, N = 3SE +/- 0.23, N = 3SE +/- 0.37, N = 3SE +/- 0.33, N = 3239.88238.13235.64235.59235.27234.701. (CC) gcc options: -O3 -fcommon -fPIE -fPIC -fvisibility=hidden -O2 -pie -rdynamic -lpthread -lrt -lm

SVT-VP9

Tuning: Visual Quality Optimized - Input: Bosphorus 1080p

Clang 12GCC 11.1OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-VP9 0.3Tuning: Visual Quality Optimized - Input: Bosphorus 1080p-O3 -march=native-O3 -march=native -flto-O250100150200250SE +/- 0.75, N = 3SE +/- 0.89, N = 3SE +/- 1.28, N = 3SE +/- 0.87, N = 3SE +/- 0.82, N = 3SE +/- 0.97, N = 3227.73227.66225.57223.11222.63222.331. (CC) gcc options: -O3 -fcommon -fPIE -fPIC -fvisibility=hidden -O2 -pie -rdynamic -lpthread -lrt -lm

Timed HMMer Search

Pfam Database Search

Clang 12GCC 11.1OpenBenchmarking.orgSeconds, Fewer Is BetterTimed HMMer Search 3.3.2Pfam Database Search-O3 -march=native -flto-O3 -march=native-O220406080100SE +/- 0.08, N = 3SE +/- 0.11, N = 3SE +/- 0.08, N = 3SE +/- 0.08, N = 3SE +/- 0.04, N = 3SE +/- 0.04, N = 393.3195.2196.0997.7197.7499.721. (CC) gcc options: -pthread -lhmmer -leasel -lm -lmpi

Timed MrBayes Analysis

Primate Phylogeny Analysis

Clang 12GCC 11.1OpenBenchmarking.orgSeconds, Fewer Is BetterTimed MrBayes Analysis 3.2.7Primate Phylogeny Analysis-O3 -march=native-O3 -march=native -flto-O220406080100SE +/- 0.18, N = 3SE +/- 0.19, N = 3SE +/- 0.21, N = 3SE +/- 0.29, N = 3SE +/- 0.13, N = 3SE +/- 0.13, N = 390.7992.4193.5392.9793.3496.791. (CC) gcc options: -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4a -msha -maes -mavx -mfma -mavx2 -mrdrnd -mbmi -mbmi2 -madx -mabm -O3 -std=c99 -pedantic -lm

TNN

Target: CPU - Model: MobileNet v2

GCC 11.1Clang 12OpenBenchmarking.orgms, Fewer Is BetterTNN 0.2.3Target: CPU - Model: MobileNet v2-O3 -march=native -flto-O2-O3 -march=native80160240320400SE +/- 1.27, N = 3SE +/- 1.57, N = 3SE +/- 0.86, N = 3SE +/- 0.55, N = 3SE +/- 0.41, N = 3SE +/- 0.48, N = 3215.34219.13227.77261.84271.03355.931. (CXX) g++ options: -fopenmp=libomp -pthread -fvisibility=hidden -O2 -rdynamic -ldl

TNN

Target: CPU - Model: SqueezeNet v1.1

GCC 11.1Clang 12OpenBenchmarking.orgms, Fewer Is BetterTNN 0.2.3Target: CPU - Model: SqueezeNet v1.1-O3 -march=native -flto-O2-O3 -march=native50100150200250SE +/- 0.11, N = 3SE +/- 0.08, N = 3SE +/- 0.50, N = 3SE +/- 0.41, N = 3SE +/- 0.23, N = 3SE +/- 0.69, N = 3202.57208.78214.28207.93208.56215.781. (CXX) g++ options: -pthread -fvisibility=hidden -O2 -rdynamic -ldl

WebP Image Encode

Encode Settings: Quality 100, Lossless

GCC 11.1Clang 12OpenBenchmarking.orgEncode Time - Seconds, Fewer Is BetterWebP Image Encode 1.1Encode Settings: Quality 100, Lossless-O2-O3 -march=native -flto-O3 -march=native48121620SE +/- 0.04, N = 3SE +/- 0.17, N = 3SE +/- 0.06, N = 3SE +/- 0.15, N = 4SE +/- 0.19, N = 3SE +/- 0.05, N = 313.1813.2513.5513.6413.6514.111. (CC) gcc options: -fvisibility=hidden -pthread -lm -ljpeg

WebP Image Encode

Encode Settings: Quality 100, Highest Compression

Clang 12GCC 11.1OpenBenchmarking.orgEncode Time - Seconds, Fewer Is BetterWebP Image Encode 1.1Encode Settings: Quality 100, Highest Compression-O3 -march=native-O3 -march=native -flto-O21.20492.40983.61474.81966.0245SE +/- 0.026, N = 3SE +/- 0.027, N = 3SE +/- 0.059, N = 4SE +/- 0.036, N = 3SE +/- 0.023, N = 3SE +/- 0.021, N = 34.7034.7224.7785.1845.2005.3551. (CC) gcc options: -fvisibility=hidden -pthread -lm -ljpeg

WebP Image Encode

Encode Settings: Quality 100, Lossless, Highest Compression

Clang 12GCC 11.1OpenBenchmarking.orgEncode Time - Seconds, Fewer Is BetterWebP Image Encode 1.1Encode Settings: Quality 100, Lossless, Highest Compression-O3 -march=native -flto-O2-O3 -march=native714212835SE +/- 0.02, N = 3SE +/- 0.05, N = 3SE +/- 0.13, N = 3SE +/- 0.03, N = 3SE +/- 0.07, N = 3SE +/- 0.02, N = 327.4028.6429.1227.4227.9528.331. (CC) gcc options: -fvisibility=hidden -pthread -lm -ljpeg

x265

Video Input: Bosphorus 4K

Clang 12GCC 11.1OpenBenchmarking.orgFrames Per Second, More Is Betterx265 3.4Video Input: Bosphorus 4K-O3 -march=native -flto-O2-O3 -march=native714212835SE +/- 0.10, N = 3SE +/- 0.12, N = 3SE +/- 0.04, N = 3SE +/- 0.06, N = 3SE +/- 0.09, N = 3SE +/- 0.04, N = 328.4727.7727.7526.2426.1725.911. (CXX) g++ options: -O2 -rdynamic -lpthread -lrt -ldl


Phoronix Test Suite v10.8.5