mnn ncnn apple m1

Apple M1 testing with a Apple Mac mini (M1 2020) and llvmpipe on Arch Linux ARM via the Phoronix Test Suite.

HTML result view exported from: https://openbenchmarking.org/result/2208136-NE-MNNNCNNAP23&sro&grs.

mnn ncnn apple m1ProcessorMotherboardMemoryDiskGraphicsNetworkOSKernelDesktopDisplay ServerOpenGLCompilerFile-SystemScreen ResolutionABCDApple M1 @ 2.06GHz (4 Cores / 8 Threads)Apple Mac mini (M1 2020)8GB251GB APPLE SSD AP0256Q + 2 x 0GB APPLE SSD AP0256QllvmpipeBroadcom NetXtreme BCM57762 PCIe + Broadcom BRCM4378 + Broadcom Device 5f69Arch Linux ARM5.19.0-rc7-asahi-2-1-ARCH (aarch64)KDE Plasma 5.25.4X Server 1.21.1.44.5 Mesa 22.1.4 (LLVM 14.0.6 128 bits)GCC 12.1.0 + Clang 14.0.6ext41920x1080OpenBenchmarking.orgCompiler Details- --build=aarch64-unknown-linux-gnu --disable-libssp --disable-libstdcxx-pch --disable-multilib --disable-werror --enable-__cxa_atexit --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-default-ssp --enable-fix-cortex-a53-835769 --enable-fix-cortex-a53-843419 --enable-gnu-indirect-function --enable-gnu-unique-object --enable-languages=c,c++,fortran,go,lto,objc,obj-c++ --enable-lto --enable-plugin --enable-shared --enable-threads=posix --host=aarch64-unknown-linux-gnu --mandir=/usr/share/man --with-arch=armv8-a --with-linker-hash-style=gnu Processor Details- Scaling Governor: apple-cpufreq schedutilSecurity Details- itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Not affected + srbds: Not affected + tsx_async_abort: Not affected

mnn ncnn apple m1ncnn: CPU-v3-v3 - mobilenet-v3ncnn: CPU - squeezenet_ssdncnn: CPU - resnet18ncnn: CPU - googlenetncnn: CPU - shufflenet-v2ncnn: CPU - mobilenetncnn: CPU - blazefacencnn: CPU - regnety_400mncnn: CPU - alexnetncnn: CPU - resnet50mnn: inception-v3ncnn: CPU - efficientnet-b0ncnn: CPU-v2-v2 - mobilenet-v2ncnn: CPU - yolov4-tinymnn: SqueezeNetV1.0mnn: resnet-v2-50ncnn: CPU - vgg16mnn: squeezenetv1.1ncnn: CPU - vision_transformermnn: mobilenet-v1-1.0ncnn: CPU - FastestDetncnn: CPU - mnasnetmnn: MobileNetV2_224mnn: mobilenetV3ABCD2.5510.946.6314.382.1314.512.835.8113.4317.6944.4763.852.4916.937.20231.26534.134.120556.155.7521.912.634.6782.0122.5211.136.8214.772.1814.672.795.8213.6217.5443.8843.862.4617.097.22531.51234.074.139557.835.7412.152.455.0252.1222.4311.386.5614.322.1914.892.815.8613.7317.5644.0703.902.4617.047.16631.34633.914.115558.925.7352.012.564.5111.9292.5611.106.7814.282.1614.622.865.9513.6017.8143.9273.902.4817.137.16531.33333.944.119557.345.7492.002.434.5361.976OpenBenchmarking.org

NCNN

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

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20220729Target: CPU-v3-v3 - Model: mobilenet-v3ABCD0.5761.1521.7282.3042.88SE +/- 0.02, N = 2SE +/- 0.03, N = 3SE +/- 0.08, N = 3SE +/- 0.01, N = 32.552.522.432.56MIN: 2.5 / MAX: 4.94MIN: 2.46 / MAX: 2.78MIN: 2.27 / MAX: 2.59MIN: 2.51 / MAX: 5.271. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: squeezenet_ssd

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20220729Target: CPU - Model: squeezenet_ssdABCD3691215SE +/- 0.17, N = 3SE +/- 0.19, N = 3SE +/- 0.06, N = 3SE +/- 0.20, N = 310.9411.1311.3811.10MIN: 8.03 / MAX: 20.45MIN: 8.08 / MAX: 21.53MIN: 8.02 / MAX: 21.68MIN: 8.04 / MAX: 21.851. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: resnet18

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20220729Target: CPU - Model: resnet18ABCD246810SE +/- 0.02, N = 3SE +/- 0.04, N = 3SE +/- 0.04, N = 3SE +/- 0.05, N = 36.636.826.566.78MIN: 5.77 / MAX: 15.63MIN: 5.74 / MAX: 16.88MIN: 5.75 / MAX: 20.92MIN: 5.74 / MAX: 16.121. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: googlenet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20220729Target: CPU - Model: googlenetABCD48121620SE +/- 0.10, N = 3SE +/- 0.19, N = 2SE +/- 0.25, N = 3SE +/- 0.26, N = 314.3814.7714.3214.28MIN: 9.51 / MAX: 22.97MIN: 9.79 / MAX: 24.43MIN: 9.41 / MAX: 23.61MIN: 9.29 / MAX: 23.481. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: shufflenet-v2

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20220729Target: CPU - Model: shufflenet-v2ABCD0.49280.98561.47841.97122.464SE +/- 0.02, N = 3SE +/- 0.04, N = 3SE +/- 0.02, N = 3SE +/- 0.01, N = 32.132.182.192.16MIN: 2.09 / MAX: 2.44MIN: 2.12 / MAX: 7.27MIN: 2.11 / MAX: 20.5MIN: 2.12 / MAX: 3.031. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: mobilenet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20220729Target: CPU - Model: mobilenetABCD48121620SE +/- 0.14, N = 3SE +/- 0.14, N = 3SE +/- 0.04, N = 3SE +/- 0.09, N = 314.5114.6714.8914.62MIN: 9.47 / MAX: 25.78MIN: 9.45 / MAX: 26.22MIN: 9.41 / MAX: 25.28MIN: 9.37 / MAX: 25.121. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: blazeface

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20220729Target: CPU - Model: blazefaceABCD0.64351.2871.93052.5743.2175SE +/- 0.07, N = 2SE +/- 0.03, N = 3SE +/- 0.08, N = 3SE +/- 0.03, N = 32.832.792.812.86MIN: 1.26 / MAX: 9.38MIN: 1.38 / MAX: 10.17MIN: 1.17 / MAX: 9.42MIN: 1.14 / MAX: 9.271. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: regnety_400m

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20220729Target: CPU - Model: regnety_400mABCD1.33882.67764.01645.35526.694SE +/- 0.01, N = 3SE +/- 0.02, N = 3SE +/- 0.04, N = 3SE +/- 0.07, N = 35.815.825.865.95MIN: 5.78 / MAX: 6.09MIN: 5.78 / MAX: 10.73MIN: 5.79 / MAX: 7.87MIN: 5.79 / MAX: 8.561. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: alexnet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20220729Target: CPU - Model: alexnetABCD48121620SE +/- 0.10, N = 3SE +/- 0.09, N = 3SE +/- 0.07, N = 3SE +/- 0.03, N = 313.4313.6213.7313.60MIN: 9.68 / MAX: 21.9MIN: 9.81 / MAX: 22.21MIN: 9.79 / MAX: 22.09MIN: 9.79 / MAX: 21.971. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: resnet50

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20220729Target: CPU - Model: resnet50ABCD48121620SE +/- 0.06, N = 3SE +/- 0.11, N = 3SE +/- 0.11, N = 3SE +/- 0.02, N = 317.6917.5417.5617.81MIN: 15.3 / MAX: 27.57MIN: 15.32 / MAX: 28.27MIN: 15.37 / MAX: 27.48MIN: 15.42 / MAX: 30.471. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

Mobile Neural Network

Model: inception-v3

OpenBenchmarking.orgms, Fewer Is BetterMobile Neural Network 2.0Model: inception-v3ABCD1020304050SE +/- 0.27, N = 12SE +/- 0.81, N = 3SE +/- 0.23, N = 12SE +/- 0.23, N = 1244.4843.8844.0743.93MIN: 43.4 / MAX: 71.04MIN: 42.56 / MAX: 52.14MIN: 43.4 / MAX: 51.58MIN: 42.55 / MAX: 50.761. (CXX) g++ options: -std=c++11 -O3 -fvisibility=hidden -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -ffast-math -fno-rtti -fno-exceptions -rdynamic -pthread -ldl

NCNN

Target: CPU - Model: efficientnet-b0

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20220729Target: CPU - Model: efficientnet-b0ABCD0.87751.7552.63253.514.3875SE +/- 0.02, N = 3SE +/- 0.02, N = 3SE +/- 0.01, N = 3SE +/- 0.02, N = 33.853.863.903.90MIN: 3.81 / MAX: 6.92MIN: 3.81 / MAX: 14.93MIN: 3.81 / MAX: 15.96MIN: 3.81 / MAX: 13.031. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

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

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20220729Target: CPU-v2-v2 - Model: mobilenet-v2ABCD0.56031.12061.68092.24122.8015SE +/- 0.02, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 32.492.462.462.48MIN: 2.44 / MAX: 13.25MIN: 2.43 / MAX: 2.9MIN: 2.44 / MAX: 4.62MIN: 2.44 / MAX: 11.81. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: yolov4-tiny

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20220729Target: CPU - Model: yolov4-tinyABCD48121620SE +/- 0.11, N = 3SE +/- 0.07, N = 3SE +/- 0.13, N = 3SE +/- 0.11, N = 316.9317.0917.0417.13MIN: 14.21 / MAX: 26.86MIN: 13.92 / MAX: 30.17MIN: 14.15 / MAX: 25.65MIN: 14.16 / MAX: 27.371. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

Mobile Neural Network

Model: SqueezeNetV1.0

OpenBenchmarking.orgms, Fewer Is BetterMobile Neural Network 2.0Model: SqueezeNetV1.0ABCD246810SE +/- 0.051, N = 12SE +/- 0.102, N = 3SE +/- 0.051, N = 12SE +/- 0.049, N = 127.2027.2257.1667.165MIN: 6.94 / MAX: 10.14MIN: 6.99 / MAX: 7.66MIN: 6.93 / MAX: 11.58MIN: 6.94 / MAX: 8.771. (CXX) g++ options: -std=c++11 -O3 -fvisibility=hidden -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -ffast-math -fno-rtti -fno-exceptions -rdynamic -pthread -ldl

Mobile Neural Network

Model: resnet-v2-50

OpenBenchmarking.orgms, Fewer Is BetterMobile Neural Network 2.0Model: resnet-v2-50ABCD714212835SE +/- 0.06, N = 12SE +/- 0.12, N = 3SE +/- 0.07, N = 12SE +/- 0.05, N = 1231.2731.5131.3531.33MIN: 30.89 / MAX: 36.52MIN: 30.98 / MAX: 35.98MIN: 30.87 / MAX: 35.51MIN: 31.02 / MAX: 36.841. (CXX) g++ options: -std=c++11 -O3 -fvisibility=hidden -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -ffast-math -fno-rtti -fno-exceptions -rdynamic -pthread -ldl

NCNN

Target: CPU - Model: vgg16

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20220729Target: CPU - Model: vgg16ABCD816243240SE +/- 0.02, N = 3SE +/- 0.08, N = 3SE +/- 0.08, N = 3SE +/- 0.15, N = 334.1334.0733.9133.94MIN: 30.19 / MAX: 46.59MIN: 30.33 / MAX: 45.17MIN: 30.14 / MAX: 47.42MIN: 30.26 / MAX: 46.131. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

Mobile Neural Network

Model: squeezenetv1.1

OpenBenchmarking.orgms, Fewer Is BetterMobile Neural Network 2.0Model: squeezenetv1.1ABCD0.93131.86262.79393.72524.6565SE +/- 0.012, N = 12SE +/- 0.014, N = 3SE +/- 0.010, N = 12SE +/- 0.009, N = 124.1204.1394.1154.119MIN: 4.03 / MAX: 4.54MIN: 4.07 / MAX: 8.54MIN: 4.04 / MAX: 7.45MIN: 4.04 / MAX: 9.151. (CXX) g++ options: -std=c++11 -O3 -fvisibility=hidden -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -ffast-math -fno-rtti -fno-exceptions -rdynamic -pthread -ldl

NCNN

Target: CPU - Model: vision_transformer

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20220729Target: CPU - Model: vision_transformerABCD120240360480600SE +/- 0.36, N = 3SE +/- 0.64, N = 3SE +/- 0.98, N = 3SE +/- 1.03, N = 3556.15557.83558.92557.34MIN: 516.47 / MAX: 589.99MIN: 520.8 / MAX: 597.49MIN: 522.41 / MAX: 596.29MIN: 523.45 / MAX: 591.961. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

Mobile Neural Network

Model: mobilenet-v1-1.0

OpenBenchmarking.orgms, Fewer Is BetterMobile Neural Network 2.0Model: mobilenet-v1-1.0ABCD1.29422.58843.88265.17686.471SE +/- 0.008, N = 12SE +/- 0.031, N = 3SE +/- 0.008, N = 12SE +/- 0.010, N = 125.7525.7415.7355.749MIN: 5.68 / MAX: 9.4MIN: 5.66 / MAX: 8.59MIN: 5.68 / MAX: 11.55MIN: 5.65 / MAX: 11.691. (CXX) g++ options: -std=c++11 -O3 -fvisibility=hidden -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -ffast-math -fno-rtti -fno-exceptions -rdynamic -pthread -ldl

NCNN

Target: CPU - Model: FastestDet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20220729Target: CPU - Model: FastestDetABCD0.48380.96761.45141.93522.419SE +/- 0.01, N = 2SE +/- 0.02, N = 3SE +/- 0.07, N = 3SE +/- 0.08, N = 31.912.152.012.00MIN: 1.88 / MAX: 3.36MIN: 2.08 / MAX: 9.79MIN: 1.88 / MAX: 8.67MIN: 1.89 / MAX: 2.181. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: mnasnet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20220729Target: CPU - Model: mnasnetABCD0.59181.18361.77542.36722.959SE +/- 0.09, N = 3SE +/- 0.01, N = 3SE +/- 0.06, N = 3SE +/- 0.00, N = 32.632.452.562.43MIN: 2.41 / MAX: 12.5MIN: 2.41 / MAX: 4.31MIN: 2.41 / MAX: 4.91MIN: 2.41 / MAX: 3.111. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

Mobile Neural Network

Model: MobileNetV2_224

OpenBenchmarking.orgms, Fewer Is BetterMobile Neural Network 2.0Model: MobileNetV2_224ABCD1.13062.26123.39184.52245.653SE +/- 0.117, N = 12SE +/- 0.031, N = 3SE +/- 0.116, N = 12SE +/- 0.110, N = 124.6785.0254.5114.536MIN: 4.14 / MAX: 8.31MIN: 4.97 / MAX: 5.27MIN: 4.09 / MAX: 10MIN: 4.16 / MAX: 10.911. (CXX) g++ options: -std=c++11 -O3 -fvisibility=hidden -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -ffast-math -fno-rtti -fno-exceptions -rdynamic -pthread -ldl

Mobile Neural Network

Model: mobilenetV3

OpenBenchmarking.orgms, Fewer Is BetterMobile Neural Network 2.0Model: mobilenetV3ABCD0.47750.9551.43251.912.3875SE +/- 0.042, N = 12SE +/- 0.020, N = 3SE +/- 0.034, N = 12SE +/- 0.043, N = 122.0122.1221.9291.976MIN: 1.79 / MAX: 5.19MIN: 2.07 / MAX: 2.37MIN: 1.78 / MAX: 2.9MIN: 1.78 / MAX: 2.461. (CXX) g++ options: -std=c++11 -O3 -fvisibility=hidden -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -ffast-math -fno-rtti -fno-exceptions -rdynamic -pthread -ldl

Phoronix Test Suite v10.8.5