CUDA 8.0 + cuDNN Caffe deep learning benchmarks with many different GPUs. Tests by Michael Larabel.
Compare your own system(s) to this result file with the
Phoronix Test Suite by running the command:
phoronix-test-suite benchmark 1702028-TA-1611066TA46 CUDA Caffe NVIDIA Comparison - Phoronix Test Suite CUDA Caffe NVIDIA Comparison CUDA 8.0 + cuDNN Caffe deep learning benchmarks with many different GPUs. Tests by Michael Larabel.
HTML result view exported from: https://openbenchmarking.org/result/1702028-TA-1611066TA46&grr&sor .
CUDA Caffe NVIDIA Comparison Processor Motherboard Chipset Memory Disk Graphics Audio Network OS Kernel Desktop Display Server Display Driver OpenGL Vulkan Compiler File-System Screen Resolution GTX 680 GTX 760 GTX 780 Ti GTX 950 GTX 960 GTX 970 GTX 980 GTX 980 Ti GTX 1050 GTX 1050 Ti GTX 1060 GTX 1070 GTX 1080 ubu_ml2 ubu_ml_375.26 Intel Xeon E3-1280 v5 @ 4.00GHz (8 Cores) MSI C236A WORKSTATION (MS-7998) v1.0 Intel Sky Lake 16384MB 256GB INTEL SSDPEKKW256G7 NVIDIA GeForce GTX 680 2048MB (1006/3004MHz) Realtek ALC1150 Intel Connection Ubuntu 16.04 4.8.4-040804-generic (x86_64) Unity 7.4.0 X Server 1.18.4 NVIDIA 375.10 4.5.0 1.0.8 GCC 5.4.0 20160609 + LLVM 3.8.0 + CUDA 8.0 ext4 3840x2160 NVIDIA GeForce GTX 760 2048MB (980/3004MHz) NVIDIA GeForce GTX 780 Ti 3072MB (875/3500MHz) eVGA NVIDIA GeForce GTX 950 2048MB (1201/3304MHz) eVGA NVIDIA GeForce GTX 960 2048MB (1277/3505MHz) eVGA NVIDIA GeForce GTX 970 4096MB (1163/3505MHz) NVIDIA GeForce GTX 980 4096MB (135/324MHz) NVIDIA GeForce GTX 980 Ti 6144MB (999/3505MHz) Zotac NVIDIA GeForce GTX 1050 2048MB (1316/3504MHz) eVGA NVIDIA GeForce GTX 1050 Ti 4096MB (1341/3504MHz) NVIDIA GeForce GTX 1060 6GB 6144MB (1506/4006MHz) NVIDIA GeForce GTX 1070 8192MB (1505/4006MHz) NVIDIA GeForce GTX 1080 8192MB (1615/5005MHz) 2 x Intel 0000 @ 3.00GHz (48 Cores) Supermicro X10DRi-LN4+ v1.01 Intel Xeon E7 v4/Xeon 64512MB 1000GB My Passport 0820 NVIDIA Device 10f0 Intel I350 Gigabit Connection 4.4.0-59-generic (x86_64) GCC 5.4.0 20160609 + CUDA 8.0 1024x768 1.0.24 OpenBenchmarking.org Compiler Details - --build=x86_64-linux-gnu --disable-browser-plugin --disable-vtable-verify --disable-werror --enable-checking=release --enable-clocale=gnu --enable-gnu-unique-object --enable-gtk-cairo --enable-java-awt=gtk --enable-java-home --enable-languages=c,ada,c++,java,go,d,fortran,objc,obj-c++ --enable-libmpx --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-arch-directory=amd64 --with-default-libstdcxx-abi=new --with-multilib-list=m32,m64,mx32 --with-tune=generic -v Processor Details - GTX 680: Scaling Governor: intel_pstate performance - GTX 760: Scaling Governor: intel_pstate performance - GTX 780 Ti: Scaling Governor: intel_pstate performance - GTX 950: Scaling Governor: intel_pstate performance - GTX 960: Scaling Governor: intel_pstate performance - GTX 970: Scaling Governor: intel_pstate performance - GTX 980: Scaling Governor: intel_pstate performance - GTX 980 Ti: Scaling Governor: intel_pstate performance - GTX 1050: Scaling Governor: intel_pstate performance - GTX 1050 Ti: Scaling Governor: intel_pstate performance - GTX 1060: Scaling Governor: intel_pstate performance - GTX 1070: Scaling Governor: intel_pstate performance - GTX 1080: Scaling Governor: intel_pstate performance - ubu_ml2: Scaling Governor: intel_pstate powersave - ubu_ml_375.26: Scaling Governor: intel_pstate powersave
CUDA Caffe NVIDIA Comparison caffe: CUDA Googlenet caffe: CUDA AlexNet GTX 680 GTX 760 GTX 780 Ti GTX 950 GTX 960 GTX 970 GTX 980 GTX 980 Ti GTX 1050 GTX 1050 Ti GTX 1060 GTX 1070 GTX 1080 ubu_ml2 ubu_ml_375.26 138342 54520.37 164643 66711.23 65152.00 28177.90 69528.20 30783.13 60318.23 27512.80 40193.77 16987.30 36217.23 15013.43 31349.83 11652.17 70347.70 30970.00 61541.57 27452.53 37604.73 16184.70 27661.90 11438.00 24019.97 9630.36 25975.07 9452.12 25979.53 9781.05 OpenBenchmarking.org
System Power Consumption Monitor Phoronix Test Suite System Monitoring OpenBenchmarking.org Watts System Power Consumption Monitor Phoronix Test Suite System Monitoring GTX 1050 Ti GTX 1050 GTX 1060 GTX 950 GTX 960 GTX 1070 GTX 970 GTX 1080 GTX 980 GTX 680 GTX 980 Ti GTX 760 GTX 780 Ti 50 100 150 200 250 Min: 36.1 / Avg: 90.28 / Max: 99.3 Min: 36.5 / Avg: 96.75 / Max: 105.9 Min: 38.4 / Avg: 126.82 / Max: 149.3 Min: 42.3 / Avg: 127.39 / Max: 140.5 Min: 43.4 / Avg: 136.19 / Max: 153.2 Min: 40.1 / Avg: 146.82 / Max: 185.9 Min: 44.2 / Avg: 161.86 / Max: 191.1 Min: 41 / Avg: 165.2 / Max: 202.8 Min: 45.9 / Avg: 177.58 / Max: 211.6 Min: 51.1 / Avg: 193.35 / Max: 204.5 Min: 48.6 / Avg: 194.02 / Max: 242.9 Min: 49.9 / Avg: 196.96 / Max: 210.5 Min: 48.5 / Avg: 241.86 / Max: 284.5
GPU Temperature Monitor Phoronix Test Suite System Monitoring OpenBenchmarking.org Celsius GPU Temperature Monitor Phoronix Test Suite System Monitoring GTX 1050 GTX 1060 GTX 970 GTX 1080 GTX 1050 Ti GTX 1070 GTX 980 GTX 960 GTX 950 GTX 980 Ti GTX 780 Ti GTX 680 GTX 760 15 30 45 60 75 Min: 28 / Avg: 46.2 / Max: 52 Min: 30 / Avg: 47.83 / Max: 58 Min: 30 / Avg: 50.28 / Max: 59 Min: 31 / Avg: 50.95 / Max: 64 Min: 30 / Avg: 51.15 / Max: 59 Min: 32 / Avg: 52.08 / Max: 65 Min: 36 / Avg: 61.1 / Max: 75 Min: 31 / Avg: 62.53 / Max: 71 Min: 33 / Avg: 63.23 / Max: 70 Min: 38 / Avg: 64.35 / Max: 79 Min: 37 / Avg: 70.44 / Max: 81 Min: 40 / Avg: 73.59 / Max: 79 Min: 38 / Avg: 76.14 / Max: 81
Caffe AlexNet System Power Consumption Monitor OpenBenchmarking.org Watts, Fewer Is Better Caffe AlexNet 2016-06-11 System Power Consumption Monitor GTX 1050 Ti GTX 1050 GTX 950 GTX 1060 GTX 960 GTX 1070 GTX 970 GTX 1080 GTX 980 GTX 680 GTX 760 GTX 980 Ti GTX 780 Ti 50 100 150 200 250 Min: 36.1 / Avg: 95.11 / Max: 99.3 Min: 36.5 / Avg: 101.99 / Max: 105.9 Min: 42.3 / Avg: 137.23 / Max: 140.5 Min: 100.9 / Avg: 144.81 / Max: 149.3 Min: 113.7 / Avg: 150.46 / Max: 153.2 Min: 40.1 / Avg: 164.62 / Max: 185.9 Min: 44.2 / Avg: 181.07 / Max: 191.1 Min: 41 / Avg: 182.39 / Max: 202.8 Min: 45.9 / Avg: 197.86 / Max: 211.6 Min: 51.1 / Avg: 200.58 / Max: 204.5 Min: 49.9 / Avg: 204.65 / Max: 210.5 Min: 48.6 / Avg: 225.77 / Max: 242.9 Min: 57.5 / Avg: 263.36 / Max: 284.5
Caffe AlexNet GPU Temperature Monitor OpenBenchmarking.org Celsius, Fewer Is Better Caffe AlexNet 2016-06-11 GPU Temperature Monitor GTX 1050 GTX 1060 GTX 970 GTX 1050 Ti GTX 1080 GTX 1070 GTX 950 GTX 980 GTX 960 GTX 980 Ti GTX 780 Ti GTX 680 GTX 760 15 30 45 60 75 Min: 43 / Avg: 49.79 / Max: 52 Min: 37 / Avg: 51.89 / Max: 58 Min: 46 / Avg: 55.05 / Max: 59 Min: 48 / Avg: 55.69 / Max: 59 Min: 43 / Avg: 56.27 / Max: 64 Min: 42 / Avg: 57.69 / Max: 65 Min: 64 / Avg: 68.12 / Max: 70 Min: 57 / Avg: 68.25 / Max: 75 Min: 51 / Avg: 68.64 / Max: 71 Min: 59 / Avg: 72.07 / Max: 79 Min: 59 / Avg: 75.4 / Max: 81 Min: 69 / Avg: 76.94 / Max: 79 Min: 63 / Avg: 78.92 / Max: 81
Caffe AlexNet Build: CUDA Googlenet OpenBenchmarking.org Milli-Seconds, Fewer Is Better Caffe AlexNet 2016-06-11 Build: CUDA Googlenet GTX 1080 ubu_ml2 ubu_ml_375.26 GTX 1070 GTX 980 Ti GTX 980 GTX 1060 GTX 970 GTX 960 GTX 1050 Ti GTX 780 Ti GTX 950 GTX 1050 GTX 680 GTX 760 40K 80K 120K 160K 200K SE +/- 36.77, N = 3 SE +/- 44.07, N = 3 SE +/- 58.61, N = 3 SE +/- 47.22, N = 3 SE +/- 85.73, N = 3 SE +/- 79.51, N = 3 SE +/- 55.94, N = 3 SE +/- 28.41, N = 3 SE +/- 23.55, N = 3 SE +/- 59.49, N = 3 SE +/- 56.68, N = 3 SE +/- 39.05, N = 3 SE +/- 20.52, N = 3 SE +/- 26.87, N = 3 SE +/- 99.04, N = 3 24019.97 25975.07 25979.53 27661.90 31349.83 36217.23 37604.73 40193.77 60318.23 61541.57 65152.00 69528.20 70347.70 138342.00 164643.00 1. (CXX) g++ options: -pthread -fPIC -O2 -lcaffe -lglog -lgflags -lprotobuf -lboost_system -lboost_filesystem -lm -lhdf5_hl -lhdf5 -lleveldb -lsnappy -llmdb -lopencv_core -lopencv_highgui -lopencv_imgproc -lboost_thread -lstdc++ -lcblas -latlas
Caffe AlexNet System Power Consumption Monitor OpenBenchmarking.org Watts, Fewer Is Better Caffe AlexNet 2016-06-11 System Power Consumption Monitor GTX 1050 Ti GTX 1050 GTX 950 GTX 960 GTX 1060 GTX 1070 GTX 970 GTX 980 GTX 1080 GTX 680 GTX 760 GTX 980 Ti GTX 780 Ti 50 100 150 200 250 Min: 59.3 / Avg: 94.01 / Max: 97.4 Min: 93.6 / Avg: 98.84 / Max: 99.6 Min: 124.9 / Avg: 129.66 / Max: 133.1 Min: 79.6 / Avg: 129.82 / Max: 136.7 Min: 133 / Avg: 134.46 / Max: 136.2 Min: 81.2 / Avg: 149.37 / Max: 170.9 Min: 90.8 / Avg: 162.51 / Max: 175 Min: 160.8 / Avg: 186.13 / Max: 192 Min: 186.1 / Avg: 188.6 / Max: 192.6 Min: 182.5 / Avg: 190.05 / Max: 195 Min: 172.9 / Avg: 190.92 / Max: 199.3 Min: 122 / Avg: 210.6 / Max: 234.5 Min: 196.1 / Avg: 256.14 / Max: 266.6
Caffe AlexNet GPU Temperature Monitor OpenBenchmarking.org Celsius, Fewer Is Better Caffe AlexNet 2016-06-11 GPU Temperature Monitor GTX 1060 GTX 1050 GTX 1080 GTX 1070 GTX 1050 Ti GTX 970 GTX 980 GTX 980 Ti GTX 960 GTX 950 GTX 780 Ti GTX 680 GTX 760 15 30 45 60 75 Min: 36 / Avg: 41.29 / Max: 46 Min: 37 / Avg: 42 / Max: 46 Min: 38 / Avg: 42.6 / Max: 47 Min: 39 / Avg: 43 / Max: 48 Min: 39 / Avg: 45.31 / Max: 50 Min: 41 / Avg: 45.86 / Max: 50 Min: 45 / Avg: 51.43 / Max: 58 Min: 46 / Avg: 52.67 / Max: 60 Min: 43 / Avg: 53.55 / Max: 63 Min: 46 / Avg: 57.46 / Max: 67 Min: 45 / Avg: 61.67 / Max: 72 Min: 52 / Avg: 68.04 / Max: 75 Min: 48 / Avg: 72.38 / Max: 81
Caffe AlexNet Build: CUDA AlexNet OpenBenchmarking.org Milli-Seconds, Fewer Is Better Caffe AlexNet 2016-06-11 Build: CUDA AlexNet ubu_ml2 GTX 1080 ubu_ml_375.26 GTX 1070 GTX 980 Ti GTX 980 GTX 1060 GTX 970 GTX 1050 Ti GTX 960 GTX 780 Ti GTX 950 GTX 1050 GTX 680 GTX 760 14K 28K 42K 56K 70K SE +/- 18.42, N = 3 SE +/- 8.72, N = 3 SE +/- 29.84, N = 3 SE +/- 17.68, N = 3 SE +/- 5.86, N = 3 SE +/- 8.54, N = 3 SE +/- 5.77, N = 3 SE +/- 6.55, N = 3 SE +/- 8.51, N = 3 SE +/- 24.09, N = 3 SE +/- 16.03, N = 3 SE +/- 31.65, N = 3 SE +/- 22.56, N = 3 SE +/- 123.58, N = 3 SE +/- 18.99, N = 3 9452.12 9630.36 9781.05 11438.00 11652.17 15013.43 16184.70 16987.30 27452.53 27512.80 28177.90 30783.13 30970.00 54520.37 66711.23 1. (CXX) g++ options: -pthread -fPIC -O2 -lcaffe -lglog -lgflags -lprotobuf -lboost_system -lboost_filesystem -lm -lhdf5_hl -lhdf5 -lleveldb -lsnappy -llmdb -lopencv_core -lopencv_highgui -lopencv_imgproc -lboost_thread -lstdc++ -lcblas -latlas
Phoronix Test Suite v10.8.4