llama ncnn 9950X

AMD Ryzen 9 9950X 16-Core testing with a ASRock X870E Taichi (3.12.AS02 BIOS) and AMD Radeon RX 7800 XT 16GB on Ubuntu 24.04 via the Phoronix Test Suite.

HTML result view exported from: https://openbenchmarking.org/result/2412293-SYST-LLAMANC79.

llama ncnn 9950XProcessorMotherboardChipsetMemoryDiskGraphicsAudioMonitorNetworkOSKernelDesktopDisplay ServerOpenGLCompilerFile-SystemScreen ResolutionabcdAMD Ryzen 9 9950X 16-Core @ 5.75GHz (16 Cores / 32 Threads)ASRock X870E Taichi (3.12.AS02 BIOS)AMD Device 14d82 x 16GB DDR5-6000MT/s F5-6000J2836G16GWestern Digital WD_BLACK SN850X 2000GB + 32GB Flash DriveAMD Radeon RX 7800 XT 16GBAMD Navi 31 HDMI/DPDELL U2723QERealtek Device 8126 + MEDIATEK Device 0717Ubuntu 24.046.12.3-061203-generic (x86_64)GNOME Shell 46.0X Server 1.21.1.11 + Wayland4.6 Mesa 24.2.0-devel (LLVM 18.1.7 DRM 3.59)GCC 13.3.0ext43840x2160OpenBenchmarking.orgKernel Details- Transparent Huge Pages: madviseCompiler Details- --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-bootstrap --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-backtrace --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-link-serialization=2 --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-13-fG75Ri/gcc-13-13.3.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-fG75Ri/gcc-13-13.3.0/debian/tmp-gcn/usr --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-build-config=bootstrap-lto-lean --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: amd-pstate-epp powersave (Boost: Enabled EPP: balance_performance) - CPU Microcode: 0xb404023 Security Details- gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + reg_file_data_sampling: Not affected + retbleed: Not affected + spec_rstack_overflow: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced / Automatic IBRS; IBPB: conditional; STIBP: always-on; RSB filling; PBRSB-eIBRS: Not affected; BHI: Not affected + srbds: Not affected + tsx_async_abort: Not affected

llama ncnn 9950Xncnn: CPU - mobilenetncnn: CPU-v2-v2 - mobilenet-v2ncnn: CPU-v3-v3 - mobilenet-v3ncnn: CPU - shufflenet-v2ncnn: CPU - mnasnetncnn: CPU - efficientnet-b0ncnn: CPU - blazefacencnn: CPU - googlenetncnn: CPU - vgg16ncnn: CPU - resnet18ncnn: CPU - alexnetncnn: CPU - resnet50ncnn: CPUv2-yolov3v2-yolov3 - mobilenetv2-yolov3ncnn: CPU - yolov4-tinyncnn: CPU - squeezenet_ssdncnn: CPU - regnety_400mncnn: CPU - vision_transformerncnn: CPU - FastestDetncnn: Vulkan GPU - mobilenetncnn: Vulkan GPU-v2-v2 - mobilenet-v2ncnn: Vulkan GPU-v3-v3 - mobilenet-v3ncnn: Vulkan GPU - shufflenet-v2ncnn: Vulkan GPU - mnasnetncnn: Vulkan GPU - efficientnet-b0ncnn: Vulkan GPU - blazefacencnn: Vulkan GPU - googlenetncnn: Vulkan GPU - vgg16ncnn: Vulkan GPU - resnet18ncnn: Vulkan GPU - alexnetncnn: Vulkan GPU - resnet50ncnn: Vulkan GPUv2-yolov3v2-yolov3 - mobilenetv2-yolov3ncnn: Vulkan GPU - yolov4-tinyncnn: Vulkan GPU - squeezenet_ssdncnn: Vulkan GPU - regnety_400mncnn: Vulkan GPU - vision_transformerncnn: Vulkan GPU - FastestDetllama-cpp: CPU BLAS - Llama-3.1-Tulu-3-8B-Q8_0 - Text Generation 128llama-cpp: CPU BLAS - Llama-3.1-Tulu-3-8B-Q8_0 - Prompt Processing 512llama-cpp: CPU BLAS - Llama-3.1-Tulu-3-8B-Q8_0 - Prompt Processing 1024llama-cpp: CPU BLAS - Llama-3.1-Tulu-3-8B-Q8_0 - Prompt Processing 2048llama-cpp: CPU BLAS - Mistral-7B-Instruct-v0.3-Q8_0 - Text Generation 128llama-cpp: CPU BLAS - Mistral-7B-Instruct-v0.3-Q8_0 - Prompt Processing 512llama-cpp: CPU BLAS - Mistral-7B-Instruct-v0.3-Q8_0 - Prompt Processing 1024llama-cpp: CPU BLAS - Mistral-7B-Instruct-v0.3-Q8_0 - Prompt Processing 2048llama-cpp: CPU BLAS - granite-3.0-3b-a800m-instruct-Q8_0 - Text Generation 128llama-cpp: CPU BLAS - granite-3.0-3b-a800m-instruct-Q8_0 - Prompt Processing 512llama-cpp: CPU BLAS - granite-3.0-3b-a800m-instruct-Q8_0 - Prompt Processing 1024llama-cpp: CPU BLAS - granite-3.0-3b-a800m-instruct-Q8_0 - Prompt Processing 2048abcd6.842.452.372.192.203.090.925.7823.654.053.309.016.8410.465.486.3426.572.466.922.562.402.272.423.190.976.0124.013.993.208.976.9210.885.536.5127.892.749.2590.4791.8688.949.7489.9691.6388.1965.97412.57396.01372.756.772.472.342.202.323.130.945.8423.524.003.269.016.7710.545.496.3426.822.666.792.532.402.252.403.210.955.9123.624.043.299.016.7910.595.476.4426.592.489.2591.7390.5989.409.7589.9292.0190.5565.92418.20395.63372.046.942.562.42.232.433.210.965.8524.773.993.29.16.9410.415.766.6226.561.86.412.572.422.262.433.20.985.9423.983.983.198.916.4110.255.136.4626.852.079.2487.4494.6389.69.890.0991.7390.4466.05414.72391.87371.896.792.592.452.242.533.230.995.9523.644.053.39.026.7910.735.676.6426.492.646.812.612.412.272.453.30.975.9724.1543.338.946.819.735.586.5226.311.819.2692.8292.690.59.7592.0893.4989.7165.86410.23397.2371.23OpenBenchmarking.org

NCNN

Target: CPU - Model: mobilenet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPU - Model: mobilenetabcd246810SE +/- 0.06, N = 3SE +/- 0.05, N = 96.846.776.946.79MIN: 6.66 / MAX: 13.62MIN: 6.49 / MAX: 28.5MIN: 6.9 / MAX: 7.15MIN: 6.74 / MAX: 6.921. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

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

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPU-v2-v2 - Model: mobilenet-v2abcd0.58281.16561.74842.33122.914SE +/- 0.11, N = 3SE +/- 0.06, N = 92.452.472.562.59MIN: 2.22 / MAX: 3.55MIN: 2.12 / MAX: 4.06MIN: 2.54 / MAX: 3.26MIN: 2.57 / MAX: 3.321. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

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

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPU-v3-v3 - Model: mobilenet-v3abcd0.55131.10261.65392.20522.7565SE +/- 0.04, N = 3SE +/- 0.04, N = 92.372.342.402.45MIN: 2.26 / MAX: 3.84MIN: 2.07 / MAX: 3.58MIN: 2.37 / MAX: 3.08MIN: 2.43 / MAX: 3.251. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: shufflenet-v2

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPU - Model: shufflenet-v2abcd0.5041.0081.5122.0162.52SE +/- 0.04, N = 3SE +/- 0.02, N = 92.192.202.232.24MIN: 2.08 / MAX: 3.1MIN: 2.06 / MAX: 6.27MIN: 2.21 / MAX: 2.91MIN: 2.21 / MAX: 2.511. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: mnasnet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPU - Model: mnasnetabcd0.56931.13861.70792.27722.8465SE +/- 0.11, N = 3SE +/- 0.06, N = 92.202.322.432.53MIN: 1.99 / MAX: 3.11MIN: 1.95 / MAX: 3.84MIN: 2.4 / MAX: 3.29MIN: 2.47 / MAX: 9.551. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: efficientnet-b0

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPU - Model: efficientnet-b0abcd0.72681.45362.18042.90723.634SE +/- 0.10, N = 3SE +/- 0.04, N = 93.093.133.213.23MIN: 2.93 / MAX: 10.74MIN: 2.84 / MAX: 4.74MIN: 3.17 / MAX: 4.38MIN: 3.2 / MAX: 4.391. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: blazeface

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPU - Model: blazefaceabcd0.22280.44560.66840.89121.114SE +/- 0.04, N = 3SE +/- 0.02, N = 90.920.940.960.99MIN: 0.85 / MAX: 1.88MIN: 0.82 / MAX: 1.84MIN: 0.95 / MAX: 1.07MIN: 0.98 / MAX: 1.231. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: googlenet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPU - Model: googlenetabcd1.33882.67764.01645.35526.694SE +/- 0.09, N = 3SE +/- 0.04, N = 95.785.845.855.95MIN: 5.59 / MAX: 8.98MIN: 5.54 / MAX: 9.4MIN: 5.8 / MAX: 6.04MIN: 5.83 / MAX: 6.111. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: vgg16

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPU - Model: vgg16abcd612182430SE +/- 0.46, N = 3SE +/- 0.18, N = 923.6523.5224.7723.64MIN: 22.18 / MAX: 73.49MIN: 22.14 / MAX: 92.6MIN: 22.23 / MAX: 104.18MIN: 22.18 / MAX: 78.611. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: resnet18

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPU - Model: resnet18abcd0.91131.82262.73393.64524.5565SE +/- 0.02, N = 3SE +/- 0.01, N = 94.054.003.994.05MIN: 3.94 / MAX: 7.83MIN: 3.91 / MAX: 4.2MIN: 3.91 / MAX: 4.15MIN: 3.97 / MAX: 4.171. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: alexnet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPU - Model: alexnetabcd0.74251.4852.22752.973.7125SE +/- 0.05, N = 3SE +/- 0.03, N = 93.303.263.203.30MIN: 3.18 / MAX: 3.81MIN: 3.18 / MAX: 3.61MIN: 3.17 / MAX: 3.32MIN: 3.27 / MAX: 3.411. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: resnet50

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPU - Model: resnet50abcd3691215SE +/- 0.06, N = 3SE +/- 0.02, N = 99.019.019.109.02MIN: 8.86 / MAX: 16.7MIN: 8.81 / MAX: 16.39MIN: 9 / MAX: 9.35MIN: 8.97 / MAX: 9.171. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPUv2-yolov3v2-yolov3 - Model: mobilenetv2-yolov3

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPUv2-yolov3v2-yolov3 - Model: mobilenetv2-yolov3abcd246810SE +/- 0.06, N = 3SE +/- 0.05, N = 96.846.776.946.79MIN: 6.66 / MAX: 13.62MIN: 6.49 / MAX: 28.5MIN: 6.9 / MAX: 7.15MIN: 6.74 / MAX: 6.921. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: yolov4-tiny

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPU - Model: yolov4-tinyabcd3691215SE +/- 0.27, N = 3SE +/- 0.05, N = 910.4610.5410.4110.73MIN: 9.62 / MAX: 56.72MIN: 10.21 / MAX: 26.19MIN: 9.46 / MAX: 19.13MIN: 10.49 / MAX: 15.561. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: squeezenet_ssd

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPU - Model: squeezenet_ssdabcd1.2962.5923.8885.1846.48SE +/- 0.08, N = 3SE +/- 0.07, N = 95.485.495.765.67MIN: 5.26 / MAX: 9.58MIN: 5.09 / MAX: 5.87MIN: 5.72 / MAX: 5.91MIN: 5.6 / MAX: 5.831. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: regnety_400m

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPU - Model: regnety_400mabcd246810SE +/- 0.11, N = 3SE +/- 0.06, N = 96.346.346.626.64MIN: 6.05 / MAX: 12.6MIN: 5.91 / MAX: 8.04MIN: 6.29 / MAX: 34.38MIN: 6.58 / MAX: 6.781. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: vision_transformer

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPU - Model: vision_transformerabcd612182430SE +/- 0.03, N = 3SE +/- 0.12, N = 926.5726.8226.5626.49MIN: 25.83 / MAX: 50.83MIN: 26.28 / MAX: 68.9MIN: 26.46 / MAX: 26.89MIN: 26.38 / MAX: 26.841. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: CPU - Model: FastestDet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: CPU - Model: FastestDetabcd0.59851.1971.79552.3942.9925SE +/- 0.33, N = 3SE +/- 0.04, N = 92.462.661.802.64MIN: 1.77 / MAX: 4.58MIN: 2.45 / MAX: 2.99MIN: 1.77 / MAX: 1.88MIN: 2.61 / MAX: 2.751. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: mobilenet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: mobilenetabcd246810SE +/- 0.08, N = 3SE +/- 0.06, N = 156.926.796.416.81MIN: 6.72 / MAX: 58.33MIN: 6.33 / MAX: 38.29MIN: 6.38 / MAX: 6.51MIN: 6.76 / MAX: 6.981. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU-v2-v2 - Model: mobilenet-v2

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU-v2-v2 - Model: mobilenet-v2abcd0.58731.17461.76192.34922.9365SE +/- 0.01, N = 3SE +/- 0.03, N = 152.562.532.572.61MIN: 2.52 / MAX: 3.79MIN: 2.2 / MAX: 3.58MIN: 2.55 / MAX: 3.69MIN: 2.59 / MAX: 3.31. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU-v3-v3 - Model: mobilenet-v3

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU-v3-v3 - Model: mobilenet-v3abcd0.54451.0891.63352.1782.7225SE +/- 0.01, N = 3SE +/- 0.02, N = 152.402.402.422.41MIN: 2.36 / MAX: 3.96MIN: 2.2 / MAX: 3.83MIN: 2.39 / MAX: 3.74MIN: 2.38 / MAX: 3.271. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: shufflenet-v2

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: shufflenet-v2abcd0.51081.02161.53242.04322.554SE +/- 0.01, N = 3SE +/- 0.01, N = 152.272.252.262.27MIN: 2.22 / MAX: 7.29MIN: 2.13 / MAX: 4.56MIN: 2.24 / MAX: 2.74MIN: 2.24 / MAX: 3.531. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: mnasnet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: mnasnetabcd0.55131.10261.65392.20522.7565SE +/- 0.01, N = 3SE +/- 0.03, N = 152.422.402.432.45MIN: 2.39 / MAX: 4.45MIN: 2.1 / MAX: 9.89MIN: 2.4 / MAX: 3.58MIN: 2.43 / MAX: 2.91. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: efficientnet-b0

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: efficientnet-b0abcd0.74251.4852.22752.973.7125SE +/- 0.01, N = 3SE +/- 0.02, N = 153.193.213.203.30MIN: 3.14 / MAX: 4.7MIN: 2.98 / MAX: 12.56MIN: 3.17 / MAX: 4.07MIN: 3.25 / MAX: 5.071. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: blazeface

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: blazefaceabcd0.22050.4410.66150.8821.1025SE +/- 0.01, N = 3SE +/- 0.01, N = 150.970.950.980.97MIN: 0.95 / MAX: 1.59MIN: 0.84 / MAX: 1.96MIN: 0.97 / MAX: 1MIN: 0.96 / MAX: 1.081. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: googlenet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: googlenetabcd246810SE +/- 0.10, N = 3SE +/- 0.04, N = 156.015.915.945.97MIN: 5.82 / MAX: 52.97MIN: 5.57 / MAX: 7.74MIN: 5.88 / MAX: 6.31MIN: 5.91 / MAX: 6.161. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: vgg16

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: vgg16abcd612182430SE +/- 0.41, N = 3SE +/- 0.11, N = 1524.0123.6223.9824.15MIN: 22.2 / MAX: 71.88MIN: 22.14 / MAX: 85.86MIN: 22.14 / MAX: 76.35MIN: 22.23 / MAX: 116.391. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: resnet18

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: resnet18abcd0.9091.8182.7273.6364.545SE +/- 0.03, N = 3SE +/- 0.02, N = 153.994.043.984.00MIN: 3.91 / MAX: 9.9MIN: 3.88 / MAX: 28.94MIN: 3.92 / MAX: 4.16MIN: 3.94 / MAX: 4.141. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: alexnet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: alexnetabcd0.74931.49862.24792.99723.7465SE +/- 0.00, N = 3SE +/- 0.02, N = 153.203.293.193.33MIN: 3.17 / MAX: 3.59MIN: 3.16 / MAX: 4.4MIN: 3.17 / MAX: 3.38MIN: 3.3 / MAX: 3.581. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: resnet50

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: resnet50abcd3691215SE +/- 0.03, N = 3SE +/- 0.03, N = 158.979.018.918.94MIN: 8.83 / MAX: 34.28MIN: 8.53 / MAX: 29.89MIN: 8.83 / MAX: 9.33MIN: 8.87 / MAX: 9.181. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPUv2-yolov3v2-yolov3 - Model: mobilenetv2-yolov3

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPUv2-yolov3v2-yolov3 - Model: mobilenetv2-yolov3abcd246810SE +/- 0.08, N = 3SE +/- 0.06, N = 156.926.796.416.81MIN: 6.72 / MAX: 58.33MIN: 6.33 / MAX: 38.29MIN: 6.38 / MAX: 6.51MIN: 6.76 / MAX: 6.981. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: yolov4-tiny

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: yolov4-tinyabcd3691215SE +/- 0.21, N = 3SE +/- 0.06, N = 1510.8810.5910.259.73MIN: 10.45 / MAX: 67.19MIN: 9.55 / MAX: 64.48MIN: 9.54 / MAX: 84.53MIN: 9.47 / MAX: 16.011. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: squeezenet_ssd

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: squeezenet_ssdabcd1.25552.5113.76655.0226.2775SE +/- 0.03, N = 3SE +/- 0.06, N = 155.535.475.135.58MIN: 5.43 / MAX: 5.83MIN: 5.09 / MAX: 5.97MIN: 5.08 / MAX: 5.62MIN: 5.53 / MAX: 5.711. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: regnety_400m

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: regnety_400mabcd246810SE +/- 0.04, N = 3SE +/- 0.03, N = 156.516.446.466.52MIN: 6.33 / MAX: 29.36MIN: 6.07 / MAX: 23.61MIN: 6.4 / MAX: 6.61MIN: 6.49 / MAX: 6.691. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: vision_transformer

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: vision_transformerabcd714212835SE +/- 1.18, N = 3SE +/- 0.09, N = 1527.8926.5926.8526.31MIN: 26.33 / MAX: 415.2MIN: 25.99 / MAX: 55.19MIN: 26.56 / MAX: 37.88MIN: 25.93 / MAX: 42.011. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: FastestDet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: FastestDetabcd0.61651.2331.84952.4663.0825SE +/- 0.06, N = 3SE +/- 0.09, N = 152.742.482.071.81MIN: 2.59 / MAX: 2.9MIN: 1.79 / MAX: 95.79MIN: 2.05 / MAX: 2.32MIN: 1.79 / MAX: 1.891. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

Llama.cpp

Backend: CPU BLAS - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Text Generation 128

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: CPU BLAS - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Text Generation 128abcd3691215SE +/- 0.00, N = 3SE +/- 0.00, N = 39.259.259.249.261. (CXX) g++ options: -O3

Llama.cpp

Backend: CPU BLAS - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Prompt Processing 512

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: CPU BLAS - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Prompt Processing 512abcd20406080100SE +/- 0.87, N = 6SE +/- 0.93, N = 390.4791.7387.4492.821. (CXX) g++ options: -O3

Llama.cpp

Backend: CPU BLAS - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Prompt Processing 1024

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: CPU BLAS - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Prompt Processing 1024abcd20406080100SE +/- 0.48, N = 3SE +/- 0.58, N = 391.8690.5994.6392.601. (CXX) g++ options: -O3

Llama.cpp

Backend: CPU BLAS - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Prompt Processing 2048

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: CPU BLAS - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Prompt Processing 2048abcd20406080100SE +/- 0.42, N = 3SE +/- 0.55, N = 388.9489.4089.6090.501. (CXX) g++ options: -O3

Llama.cpp

Backend: CPU BLAS - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Text Generation 128

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: CPU BLAS - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Text Generation 128abcd3691215SE +/- 0.02, N = 3SE +/- 0.01, N = 39.749.759.809.751. (CXX) g++ options: -O3

Llama.cpp

Backend: CPU BLAS - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Prompt Processing 512

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: CPU BLAS - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Prompt Processing 512abcd20406080100SE +/- 0.46, N = 3SE +/- 0.79, N = 889.9689.9290.0992.081. (CXX) g++ options: -O3

Llama.cpp

Backend: CPU BLAS - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Prompt Processing 1024

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: CPU BLAS - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Prompt Processing 1024abcd20406080100SE +/- 1.24, N = 3SE +/- 1.02, N = 391.6392.0191.7393.491. (CXX) g++ options: -O3

Llama.cpp

Backend: CPU BLAS - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Prompt Processing 2048

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: CPU BLAS - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Prompt Processing 2048abcd20406080100SE +/- 0.49, N = 3SE +/- 0.47, N = 388.1990.5590.4489.711. (CXX) g++ options: -O3

Llama.cpp

Backend: CPU BLAS - Model: granite-3.0-3b-a800m-instruct-Q8_0 - Test: Text Generation 128

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: CPU BLAS - Model: granite-3.0-3b-a800m-instruct-Q8_0 - Test: Text Generation 128abcd1530456075SE +/- 0.06, N = 3SE +/- 0.03, N = 365.9765.9266.0565.861. (CXX) g++ options: -O3

Llama.cpp

Backend: CPU BLAS - Model: granite-3.0-3b-a800m-instruct-Q8_0 - Test: Prompt Processing 512

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: CPU BLAS - Model: granite-3.0-3b-a800m-instruct-Q8_0 - Test: Prompt Processing 512abcd90180270360450SE +/- 3.32, N = 3SE +/- 3.52, N = 3412.57418.20414.72410.231. (CXX) g++ options: -O3

Llama.cpp

Backend: CPU BLAS - Model: granite-3.0-3b-a800m-instruct-Q8_0 - Test: Prompt Processing 1024

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: CPU BLAS - Model: granite-3.0-3b-a800m-instruct-Q8_0 - Test: Prompt Processing 1024abcd90180270360450SE +/- 2.61, N = 3SE +/- 2.77, N = 3396.01395.63391.87397.201. (CXX) g++ options: -O3

Llama.cpp

Backend: CPU BLAS - Model: granite-3.0-3b-a800m-instruct-Q8_0 - Test: Prompt Processing 2048

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: CPU BLAS - Model: granite-3.0-3b-a800m-instruct-Q8_0 - Test: Prompt Processing 2048abcd80160240320400SE +/- 0.84, N = 3SE +/- 0.66, N = 3372.75372.04371.89371.231. (CXX) g++ options: -O3


Phoronix Test Suite v10.8.5