AMD Radeon VII OpenCL GPU Compute Benchmarks

Benchmarks for Radeon VII launch day by Michael Larabel for a future article. Radeon tests with ROCm 2.0 on Linux 5.0 kernel.

HTML result view exported from: https://openbenchmarking.org/result/1902066-SP-VIICOMPUT16&rdt.

AMD Radeon VII OpenCL GPU Compute BenchmarksProcessorMotherboardChipsetMemoryDiskGraphicsAudioMonitorNetworkOSKernelDesktopDisplay ServerDisplay DriverOpenGLOpenCLVulkanCompilerFile-SystemScreen ResolutionGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VIIIntel Core i9-9900K @ 5.00GHz (8 Cores / 16 Threads)ASUS PRIME Z390-A (0602 BIOS)Intel Cannon Lake PCH Shared SRAM16384MBSamsung SSD 970 EVO 250GB + 2000GB SABRENTNVIDIA GeForce GTX 1080 8GB (1607/5005MHz)Realtek ALC1220Acer B286HKIntel I219-VUbuntu 18.104.20.3-042003-generic (x86_64)GNOME Shell 3.30.1X Server 1.20.1NVIDIA 415.274.6.0OpenCL 1.2 CUDA 10.0.1321.1.84GCC 8.2.0ext43840x2160NVIDIA GeForce GTX 1070 8GB (1506/4006MHz)Zotac NVIDIA GeForce GTX 1070 Ti 8GB (1607/4006MHz)NVIDIA GeForce GTX 1080 Ti 11GB (1480/5508MHz)NVIDIA TITAN RTX 24GB (1350/7000MHz)NVIDIA GeForce RTX 2080 Ti 11GB (1350/7000MHz)Zotac NVIDIA GeForce RTX 2080 8GB (1515/7000MHz)NVIDIA GeForce RTX 2060 6GB (1365/7000MHz)ASUS NVIDIA GeForce RTX 2070 8GB (1410/7000MHz)NVIDIA GeForce GTX 1060 6GB (1506/4006MHz)AMD Radeon RX 64 8GB (1630/945MHz)5.0.0-050000rc4-generic (x86_64) 201901274.5 Mesa 19.0.0-devel padoka PPA (LLVM 9.0.0)OpenCL 2.1 AMD-APP (2783.0)1.1.90AMD Radeon RX 64 8GB (1590/800MHz)MSI AMD Radeon RX 470/480/570/570X/580/580X 8GB (1366/2000MHz)Sapphire AMD Radeon RX 470/480/570/570X/580/580X 8GB (1560/2100MHz)AMD Vega 20 16GB (1801/1000MHz)OpenBenchmarking.orgCompiler 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++ --enable-libmpx --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-targets=nvptx-none --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 --with-tune=generic --without-cuda-driver -vProcessor Details- Scaling Governor: intel_pstate performanceOpenCL Details- GTX 1080: GPU Compute Cores: 2560- GTX 1070: GPU Compute Cores: 1920- GTX 1070 Ti: GPU Compute Cores: 2432- GTX 1080 Ti: GPU Compute Cores: 3584- TITAN RTX: GPU Compute Cores: 4608- RTX 2080 Ti: GPU Compute Cores: 4352- RTX 2080: GPU Compute Cores: 2944- RTX 2060: GPU Compute Cores: 1920- RTX 2070: GPU Compute Cores: 2304- GTX 1060: GPU Compute Cores: 1280Python Details- Python 2.7.15+ + Python 3.6.7Security Details- __user pointer sanitization + Full generic retpoline IBPB: conditional IBRS_FW STIBP: conditional RSB filling + SSB disabled via prctl and seccomp

AMD Radeon VII OpenCL GPU Compute Benchmarkslczero: OpenCLplaidml: Yes - Inference - Mobilenet - OpenCLplaidml: Yes - Inference - Inception V3 - OpenCLplaidml: Yes - Inference - ResNet 50 - OpenCLplaidml: Yes - Inference - VGG16 - OpenCLplaidml: No - Training - IMDB LSTM - OpenCLplaidml: No - Inference - Mobilenet - OpenCLplaidml: No - Inference - ResNet 50 - OpenCLplaidml: No - Inference - VGG16 - OpenCLluxmark: GPU - Luxball HDRclpeak: Global Memory Bandwidthclpeak: Single-Precision FloatGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII1750686132264139167801292143138032227934143066411723712615676425712917287196587814786451202371281537552621311677619767762333774167334194187883339189216823291085332759772154742642441336556289460225251638631079712114582502461291537273427875051535222598931683541922161045400203291463681027014847791342821481888633091502135627666301729830157333179198100837618929532365796889955288.8518394.6213459219494.08122561454239856587120.03213119.99184835270137.623256536212469758575109.03195106.66178720228120.44310793171035764056376.5912860.9813041413365.1815271206620268359982.7113764.0414445314268.691722121770691391904166.30353197.632001233403224.775557080713453OpenBenchmarking.org

LeelaChessZero

Backend: OpenCL

OpenBenchmarking.orgNodes Per Second, More Is BetterLeelaChessZero 0.20.1Backend: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII7001400210028003500SE +/- 32.06, N = 3SE +/- 9.63, N = 3SE +/- 13.26, N = 3SE +/- 25.14, N = 3SE +/- 28.14, N = 3SE +/- 20.98, N = 3SE +/- 22.18, N = 3SE +/- 13.06, N = 3SE +/- 20.57, N = 3SE +/- 6.27, N = 3SE +/- 5.51, N = 3SE +/- 2.66, N = 3SE +/- 2.68, N = 3SE +/- 2.57, N = 3SE +/- 8.64, N = 317501430147823333275310722591484172989985675864068313911. (CXX) g++ options: -lpthread

LeelaChessZero

Backend: OpenCL

OpenBenchmarking.orgNodes Per Second Per Watt, More Is BetterLeelaChessZero 0.20.1Backend: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII51015202511.817.779.458.8022.1115.1010.137.579.275.972.734.383.743.629.38

LeelaChessZero

System Power Consumption Monitor

OpenBenchmarking.orgWatts, Fewer Is BetterLeelaChessZero 0.20.1System Power Consumption MonitorGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII60120180240300Min: 40.2 / Avg: 148.14 / Max: 249.5Min: 42.6 / Avg: 184.04 / Max: 213.3Min: 80 / Avg: 156.39 / Max: 189.7Min: 96.9 / Avg: 265.17 / Max: 323.1Min: 97.7 / Avg: 148.16 / Max: 357.7Min: 48.2 / Avg: 205.76 / Max: 349.4Min: 88.3 / Avg: 222.99 / Max: 288.3Min: 69.4 / Avg: 196.01 / Max: 230Min: 75.1 / Avg: 186.61 / Max: 251.7Min: 41.9 / Avg: 150.59 / Max: 180Min: 50.2 / Avg: 313.27 / Max: 342.7Min: 47.2 / Avg: 173.2 / Max: 262.2Min: 69.3 / Avg: 171.21 / Max: 229.5Min: 75.2 / Avg: 188.72 / Max: 253.3Min: 56 / Avg: 148.33 / Max: 269

PlaidML

FP16: Yes - Mode: Inference - Network: Mobilenet - Device: OpenCL

OpenBenchmarking.orgExamples Per Second, More Is BetterPlaidMLFP16: Yes - Mode: Inference - Network: Mobilenet - Device: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII2004006008001000SE +/- 5.02, N = 3SE +/- 4.06, N = 3SE +/- 3.15, N = 3SE +/- 3.70, N = 3SE +/- 3.45, N = 3SE +/- 2.70, N = 3SE +/- 1.63, N = 3SE +/- 1.27, N = 3SE +/- 1.07, N = 3SE +/- 0.87, N = 3SE +/- 9.53, N = 3SE +/- 6.73, N = 3SE +/- 1.03, N = 3SE +/- 1.00, N = 3SE +/- 4.52, N = 3686664645774977971893779830552587575563599904

PlaidML

FP16: Yes - Mode: Inference - Network: Mobilenet - Device: OpenCL

OpenBenchmarking.orgExamples Per Second Per Watt, More Is BetterPlaidMLFP16: Yes - Mode: Inference - Network: Mobilenet - Device: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII36912157.347.226.575.198.8210.286.826.418.875.128.486.235.144.7411.54

PlaidML

System Power Consumption Monitor

OpenBenchmarking.orgWatts, Fewer Is BetterPlaidMLSystem Power Consumption MonitorGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII4080120160200Min: 43.7 / Avg: 93.48 / Max: 127.3Min: 42.8 / Avg: 91.98 / Max: 117.4Min: 48.7 / Avg: 98.18 / Max: 132.6Min: 91.6 / Avg: 149.17 / Max: 203.2Min: 52.9 / Avg: 110.8 / Max: 140.2Min: 48.4 / Avg: 94.48 / Max: 113.3Min: 56.3 / Avg: 130.87 / Max: 184.7Min: 45.5 / Avg: 121.65 / Max: 166.1Min: 46 / Avg: 93.68 / Max: 129.3Min: 42.1 / Avg: 107.84 / Max: 141.6Min: 54 / Avg: 69.27 / Max: 82.4Min: 49.6 / Avg: 92.35 / Max: 157.7Min: 75.6 / Avg: 109.4 / Max: 179Min: 82.3 / Avg: 126.56 / Max: 194.9Min: 58.1 / Avg: 78.39 / Max: 92.6

PlaidML

FP16: Yes - Mode: Inference - Network: Inception V3 - Device: OpenCL

OpenBenchmarking.orgExamples Per Second, More Is BetterPlaidMLFP16: Yes - Mode: Inference - Network: Inception V3 - Device: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII50100150200250SE +/- 0.07, N = 3SE +/- 0.07, N = 3SE +/- 0.30, N = 3SE +/- 0.58, N = 3SE +/- 1.30, N = 3SE +/- 0.48, N = 3SE +/- 0.40, N = 3SE +/- 0.44, N = 3SE +/- 0.18, N = 3SE +/- 0.05, N = 3SE +/- 0.02, N = 3SE +/- 0.06, N = 3SE +/- 0.04, N = 3SE +/- 0.11, N = 3SE +/- 0.20, N = 3132.00117.00120.00167.00215.00211.00168.00134.00157.0088.85120.03109.0376.5982.71166.30

PlaidML

FP16: Yes - Mode: Inference - Network: Inception V3 - Device: OpenCL

OpenBenchmarking.orgExamples Per Second Per Watt, More Is BetterPlaidMLFP16: Yes - Mode: Inference - Network: Inception V3 - Device: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII0.2880.5760.8641.1521.440.780.760.900.821.281.121.080.921.070.670.660.740.520.511.24

PlaidML

System Power Consumption Monitor

OpenBenchmarking.orgWatts, Fewer Is BetterPlaidMLSystem Power Consumption MonitorGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII50100150200250Min: 89.5 / Avg: 168.75 / Max: 202.8Min: 81.8 / Avg: 153.87 / Max: 183Min: 56.1 / Avg: 132.15 / Max: 155Min: 120.7 / Avg: 202.56 / Max: 256.6Min: 51.5 / Avg: 168.92 / Max: 275.4Min: 48.4 / Avg: 188.73 / Max: 275.6Min: 47.5 / Avg: 155.97 / Max: 224.6Min: 45 / Avg: 144.95 / Max: 190.7Min: 45.8 / Avg: 146.81 / Max: 206.6Min: 87.5 / Avg: 132.19 / Max: 156.7Min: 50.6 / Avg: 181.52 / Max: 296Min: 58.9 / Avg: 147.82 / Max: 249.5Min: 73.1 / Avg: 146.45 / Max: 195.4Min: 91.8 / Avg: 163.54 / Max: 221.5Min: 69.7 / Avg: 134.1 / Max: 228.1

PlaidML

FP16: Yes - Mode: Inference - Network: ResNet 50 - Device: OpenCL

OpenBenchmarking.orgExamples Per Second, More Is BetterPlaidMLFP16: Yes - Mode: Inference - Network: ResNet 50 - Device: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII100200300400500SE +/- 0.24, N = 3SE +/- 0.38, N = 3SE +/- 0.16, N = 3SE +/- 0.34, N = 3SE +/- 1.16, N = 3SE +/- 1.03, N = 3SE +/- 0.40, N = 3SE +/- 0.26, N = 3SE +/- 0.76, N = 3SE +/- 0.17, N = 3SE +/- 0.13, N = 3SE +/- 0.07, N = 3SE +/- 0.01, N = 3SE +/- 0.04, N = 3SE +/- 0.96, N = 3264237237334474458354282333183213195128137353

PlaidML

FP16: Yes - Mode: Inference - Network: ResNet 50 - Device: OpenCL

OpenBenchmarking.orgExamples Per Second Per Watt, More Is BetterPlaidMLFP16: Yes - Mode: Inference - Network: ResNet 50 - Device: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII0.66381.32761.99142.65523.3191.621.762.081.782.922.952.391.962.511.491.241.380.910.892.75

PlaidML

System Power Consumption Monitor

OpenBenchmarking.orgWatts, Fewer Is BetterPlaidMLSystem Power Consumption MonitorGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII60120180240300Min: 91.6 / Avg: 162.7 / Max: 211.1Min: 83.8 / Avg: 134.85 / Max: 189.4Min: 48.5 / Avg: 113.89 / Max: 160.8Min: 122.3 / Avg: 187.68 / Max: 270.3Min: 52.6 / Avg: 162.76 / Max: 301.7Min: 48.4 / Avg: 155.12 / Max: 292.5Min: 47.4 / Avg: 148.19 / Max: 240.2Min: 86.5 / Avg: 143.68 / Max: 202.9Min: 78.4 / Avg: 132.98 / Max: 222.4Min: 60.4 / Avg: 122.52 / Max: 164.5Min: 82.4 / Avg: 172.03 / Max: 313.6Min: 50.5 / Avg: 141.19 / Max: 254.3Min: 73.6 / Avg: 141.5 / Max: 187.8Min: 81.1 / Avg: 153.59 / Max: 208.6Min: 70.1 / Avg: 128.33 / Max: 243.5

PlaidML

FP16: Yes - Mode: Inference - Network: VGG16 - Device: OpenCL

OpenBenchmarking.orgExamples Per Second, More Is BetterPlaidMLFP16: Yes - Mode: Inference - Network: VGG16 - Device: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII60120180240300SE +/- 0.09, N = 3SE +/- 0.12, N = 3SE +/- 0.06, N = 3SE +/- 0.32, N = 3SE +/- 0.43, N = 3SE +/- 0.20, N = 3SE +/- 0.32, N = 3SE +/- 0.15, N = 3SE +/- 0.26, N = 3SE +/- 0.03, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.02, N = 3SE +/- 0.01, N = 3SE +/- 0.03, N = 3139.00126.00128.00194.00264.00250.00192.00148.00179.0094.62119.99106.6660.9864.04197.63

PlaidML

FP16: Yes - Mode: Inference - Network: VGG16 - Device: OpenCL

OpenBenchmarking.orgExamples Per Second Per Watt, More Is BetterPlaidMLFP16: Yes - Mode: Inference - Network: VGG16 - Device: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII0.34880.69761.04641.39521.7440.800.921.021.081.501.381.211.081.270.720.550.700.410.391.55

PlaidML

System Power Consumption Monitor

OpenBenchmarking.orgWatts, Fewer Is BetterPlaidMLSystem Power Consumption MonitorGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII60120180240300Min: 70.3 / Avg: 172.91 / Max: 216.6Min: 43.1 / Avg: 136.92 / Max: 191.2Min: 48.4 / Avg: 125.32 / Max: 169.8Min: 65.2 / Avg: 179.46 / Max: 287.9Min: 58.3 / Avg: 176.01 / Max: 322.5Min: 47.7 / Avg: 180.78 / Max: 312Min: 49.8 / Avg: 158.89 / Max: 252.8Min: 62.3 / Avg: 137.76 / Max: 208Min: 47.2 / Avg: 141.31 / Max: 234Min: 58.2 / Avg: 131.14 / Max: 167.2Min: 53.4 / Avg: 217.55 / Max: 319.8Min: 50.2 / Avg: 151.48 / Max: 258.1Min: 73.2 / Avg: 147.27 / Max: 187.1Min: 79.7 / Avg: 162.83 / Max: 210.1Min: 57.5 / Avg: 127.89 / Max: 264.7

PlaidML

FP16: No - Mode: Training - Network: IMDB LSTM - Device: OpenCL

OpenBenchmarking.orgExamples Per Second, More Is BetterPlaidMLFP16: No - Mode: Training - Network: IMDB LSTM - Device: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII50100150200250SE +/- 0.60, N = 3SE +/- 0.50, N = 3SE +/- 0.03, N = 3SE +/- 0.49, N = 3SE +/- 0.96, N = 3SE +/- 0.72, N = 3SE +/- 0.27, N = 3SE +/- 0.38, N = 3SE +/- 0.82, N = 3SE +/- 0.17, N = 3SE +/- 0.19, N = 3SE +/- 0.24, N = 3SE +/- 0.25, N = 3SE +/- 0.19, N = 3SE +/- 0.32, N = 3167156153187244246216188198134184178130144200

PlaidML

FP16: No - Mode: Training - Network: IMDB LSTM - Device: OpenCL

OpenBenchmarking.orgExamples Per Second Per Watt, More Is BetterPlaidMLFP16: No - Mode: Training - Network: IMDB LSTM - Device: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII0.36450.7291.09351.4581.82251.141.141.241.141.601.541.561.621.561.081.111.070.890.951.45

PlaidML

System Power Consumption Monitor

OpenBenchmarking.orgWatts, Fewer Is BetterPlaidMLSystem Power Consumption MonitorGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII50100150200250Min: 90.8 / Avg: 146.71 / Max: 176.8Min: 82.2 / Avg: 136.83 / Max: 164.9Min: 104.8 / Avg: 123.57 / Max: 135.6Min: 50.4 / Avg: 165.05 / Max: 214.8Min: 57.5 / Avg: 152.07 / Max: 231.9Min: 48.6 / Avg: 159.87 / Max: 223.7Min: 46.4 / Avg: 138.03 / Max: 185.4Min: 44.3 / Avg: 116.14 / Max: 162.7Min: 47.1 / Avg: 127.27 / Max: 169.5Min: 70.1 / Avg: 124.21 / Max: 145Min: 52.2 / Avg: 165.89 / Max: 305.6Min: 74.8 / Avg: 165.69 / Max: 271.3Min: 71.3 / Avg: 146.11 / Max: 185.3Min: 80.2 / Avg: 152.62 / Max: 212.4Min: 87.3 / Avg: 137.45 / Max: 192.4

PlaidML

FP16: No - Mode: Inference - Network: Mobilenet - Device: OpenCL

OpenBenchmarking.orgExamples Per Second, More Is BetterPlaidMLFP16: No - Mode: Inference - Network: Mobilenet - Device: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII30060090012001500SE +/- 0.61, N = 3SE +/- 1.02, N = 3SE +/- 0.31, N = 3SE +/- 4.51, N = 3SE +/- 3.01, N = 3SE +/- 1.39, N = 3SE +/- 0.31, N = 3SE +/- 1.86, N = 3SE +/- 1.43, N = 3SE +/- 3.04, N = 3SE +/- 7.40, N = 3SE +/- 9.70, N = 3SE +/- 0.26, N = 3SE +/- 0.24, N = 3SE +/- 3.74, N = 380176475588313361291104586310085928357204144531233

PlaidML

FP16: No - Mode: Inference - Network: Mobilenet - Device: OpenCL

OpenBenchmarking.orgExamples Per Second Per Watt, More Is BetterPlaidMLFP16: No - Mode: Inference - Network: Mobilenet - Device: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII36912155.427.128.067.026.678.4510.679.008.707.727.475.823.623.299.34

PlaidML

System Power Consumption Monitor

OpenBenchmarking.orgWatts, Fewer Is BetterPlaidMLSystem Power Consumption MonitorGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII50100150200250Min: 44.3 / Avg: 147.75 / Max: 195.1Min: 45.9 / Avg: 107.3 / Max: 155.1Min: 49.7 / Avg: 93.67 / Max: 125.7Min: 53.3 / Avg: 125.73 / Max: 165.2Min: 54 / Avg: 200.17 / Max: 274.2Min: 48.2 / Avg: 152.87 / Max: 274.6Min: 52 / Avg: 97.93 / Max: 121Min: 43.1 / Avg: 95.9 / Max: 118.8Min: 115.1 / Avg: 115.83 / Max: 116.5Min: 41.8 / Avg: 76.7 / Max: 94.4Min: 52.5 / Avg: 111.85 / Max: 223.9Min: 72.5 / Avg: 123.7 / Max: 259.6Min: 71.7 / Avg: 114.33 / Max: 178.6Min: 82.4 / Avg: 137.7 / Max: 205.5Min: 92.8 / Avg: 132.03 / Max: 241.7

PlaidML

FP16: No - Mode: Inference - Network: ResNet 50 - Device: OpenCL

OpenBenchmarking.orgExamples Per Second, More Is BetterPlaidMLFP16: No - Mode: Inference - Network: ResNet 50 - Device: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII120240360480600SE +/- 0.41, N = 3SE +/- 0.45, N = 3SE +/- 0.24, N = 3SE +/- 0.55, N = 3SE +/- 1.67, N = 3SE +/- 0.31, N = 3SE +/- 0.45, N = 3SE +/- 0.26, N = 3SE +/- 0.33, N = 3SE +/- 0.11, N = 3SE +/- 1.40, N = 3SE +/- 2.52, N = 3SE +/- 0.04, N = 3SE +/- 0.07, N = 3SE +/- 0.30, N = 3292257262339556537400309376194270228133142403

PlaidML

FP16: No - Mode: Inference - Network: ResNet 50 - Device: OpenCL

OpenBenchmarking.orgExamples Per Second Per Watt, More Is BetterPlaidMLFP16: No - Mode: Inference - Network: ResNet 50 - Device: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII0.69981.39962.09942.79923.4991.921.771.941.822.863.112.512.302.441.591.521.510.920.882.97

PlaidML

System Power Consumption Monitor

OpenBenchmarking.orgWatts, Fewer Is BetterPlaidMLSystem Power Consumption MonitorGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII60120180240300Min: 91.9 / Avg: 152.34 / Max: 225.3Min: 45.2 / Avg: 145.44 / Max: 201.2Min: 48.6 / Avg: 134.71 / Max: 168.9Min: 61.2 / Avg: 185.7 / Max: 280.5Min: 102.3 / Avg: 194.3 / Max: 332.7Min: 100.3 / Avg: 172.57 / Max: 323.3Min: 48.4 / Avg: 158.95 / Max: 257.5Min: 65 / Avg: 134.41 / Max: 208.9Min: 49.8 / Avg: 153.92 / Max: 242.3Min: 71.4 / Avg: 121.65 / Max: 172.6Min: 81.5 / Avg: 177.38 / Max: 338.2Min: 50.1 / Avg: 150.39 / Max: 271.1Min: 73 / Avg: 145.33 / Max: 194.5Min: 81 / Avg: 162.42 / Max: 220Min: 58.8 / Avg: 135.42 / Max: 277

PlaidML

FP16: No - Mode: Inference - Network: VGG16 - Device: OpenCL

OpenBenchmarking.orgExamples Per Second, More Is BetterPlaidMLFP16: No - Mode: Inference - Network: VGG16 - Device: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII60120180240300SE +/- 0.06, N = 3SE +/- 0.04, N = 3SE +/- 0.01, N = 3SE +/- 0.12, N = 3SE +/- 0.72, N = 3SE +/- 0.36, N = 3SE +/- 0.17, N = 3SE +/- 0.16, N = 3SE +/- 0.27, N = 3SE +/- 0.04, N = 3SE +/- 0.11, N = 3SE +/- 0.18, N = 3SE +/- 0.05, N = 3SE +/- 0.01, N = 3SE +/- 0.38, N = 3143.00129.00131.00189.00289.00273.00203.00150.00189.0094.08137.62120.4465.1868.69224.77

PlaidML

FP16: No - Mode: Inference - Network: VGG16 - Device: OpenCL

OpenBenchmarking.orgExamples Per Second Per Watt, More Is BetterPlaidMLFP16: No - Mode: Inference - Network: VGG16 - Device: OpenCLGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII0.29250.5850.87751.171.46250.790.790.890.821.291.281.080.901.030.650.640.640.400.371.30

PlaidML

System Power Consumption Monitor

OpenBenchmarking.orgWatts, Fewer Is BetterPlaidMLSystem Power Consumption MonitorGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII60120180240300Min: 45.1 / Avg: 180.6 / Max: 225Min: 43.1 / Avg: 162.78 / Max: 197.9Min: 47.9 / Avg: 146.58 / Max: 171.7Min: 127.7 / Avg: 231.45 / Max: 296.9Min: 55.4 / Avg: 223.68 / Max: 345.8Min: 47.6 / Avg: 212.98 / Max: 331.1Min: 50.9 / Avg: 188.45 / Max: 269.3Min: 44 / Avg: 167.51 / Max: 215.1Min: 68.7 / Avg: 183.65 / Max: 245.5Min: 72.7 / Avg: 144.61 / Max: 170.9Min: 52.7 / Avg: 216.48 / Max: 338.1Min: 78.5 / Avg: 188.36 / Max: 262.2Min: 73 / Avg: 163.55 / Max: 194.8Min: 86.9 / Avg: 184.98 / Max: 216.4Min: 58.2 / Avg: 172.91 / Max: 296.6

LuxMark

OpenCL Device: GPU - Scene: Luxball HDR

OpenBenchmarking.orgScore, More Is BetterLuxMark 3.1OpenCL Device: GPU - Scene: Luxball HDRGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII12K24K36K48K60KSE +/- 42.00, N = 3SE +/- 1.33, N = 3SE +/- 14.90, N = 3SE +/- 8.33, N = 3SE +/- 103.04, N = 3SE +/- 155.67, N = 3SE +/- 25.12, N = 3SE +/- 31.58, N = 3SE +/- 107.00, N = 3SE +/- 46.58, N = 3SE +/- 583.87, N = 3SE +/- 65.67, N = 3SE +/- 17.84, N = 3SE +/- 74.83, N = 3SE +/- 258.00, N = 3138031728716776216824602242787291462135629532122563256531079152711722155570

LuxMark

OpenCL Device: GPU - Scene: Luxball HDR

OpenBenchmarking.orgScore Per Watt, More Is BetterLuxMark 3.1OpenCL Device: GPU - Scene: Luxball HDRGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII408012016020079.4298.48114.1689.34137.54133.42123.99110.80131.0985.5398.89120.5880.5378.90188.36

LuxMark

System Power Consumption Monitor

OpenBenchmarking.orgWatts, Fewer Is BetterLuxMark 3.1System Power Consumption MonitorGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII60120180240300Min: 89.3 / Avg: 173.79 / Max: 175.8Min: 82.6 / Avg: 175.54 / Max: 179.2Min: 47.6 / Avg: 146.95 / Max: 149.7Min: 161.6 / Avg: 242.69 / Max: 248.1Min: 50.6 / Avg: 334.61 / Max: 351.7Min: 47.7 / Avg: 320.7 / Max: 332Min: 48.9 / Avg: 235.08 / Max: 244.1Min: 44 / Avg: 192.74 / Max: 197.4Min: 46 / Avg: 225.28 / Max: 233.2Min: 111.3 / Avg: 143.3 / Max: 145.4Min: 49.5 / Avg: 329.31 / Max: 338.6Min: 60.2 / Avg: 257.74 / Max: 269.1Min: 71.5 / Avg: 189.63 / Max: 196.7Min: 82.1 / Avg: 218.27 / Max: 228.9Min: 87.3 / Avg: 295.02 / Max: 302.2

clpeak

OpenCL Test: Global Memory Bandwidth

OpenBenchmarking.orgGBPS, More Is BetterclpeakOpenCL Test: Global Memory BandwidthGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII2004006008001000SE +/- 0.05, N = 3SE +/- 0.02, N = 3SE +/- 0.22, N = 3SE +/- 0.68, N = 3SE +/- 1.70, N = 3SE +/- 0.48, N = 3SE +/- 0.37, N = 3SE +/- 0.19, N = 3SE +/- 2.69, N = 3SE +/- 0.95, N = 3SE +/- 0.05, N = 3SE +/- 0.01, N = 3SE +/- 1.46, N = 3SE +/- 1.82, N = 3SE +/- 0.10, N = 32221961973295255053682763651453623172062178071. (CXX) g++ options: -O3 -rdynamic -lOpenCL

clpeak

OpenCL Test: Single-Precision Float

OpenBenchmarking.orgGFLOPS, More Is BetterclpeakOpenCL Test: Single-Precision FloatGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070GTX 1060RX Vega 64RX Vega 56RX 580RX 590Radeon VII4K8K12K16K20KSE +/- 396.02, N = 3SE +/- 393.27, N = 3SE +/- 14.43, N = 3SE +/- 782.31, N = 3SE +/- 688.69, N = 3SE +/- 1065.18, N = 3SE +/- 732.21, N = 3SE +/- 657.16, N = 3SE +/- 495.99, N = 3SE +/- 3.56, N = 3SE +/- 41.13, N = 3SE +/- 38.50, N = 3SE +/- 1.21, N = 3SE +/- 15.75, N = 3SE +/- 1.83, N = 379345878677610853163861535210270663079684239124691035762027069134531. (CXX) g++ options: -O3 -rdynamic -lOpenCL

GPU Temperature Monitor

Phoronix Test Suite System Monitoring

OpenBenchmarking.orgCelsiusGPU Temperature MonitorPhoronix Test Suite System MonitoringGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070RX Vega 64RX Vega 56RX 580RX 590Radeon VII20406080100Min: 31 / Avg: 60.14 / Max: 74Min: 32 / Avg: 61.66 / Max: 76Min: 38 / Avg: 45.88 / Max: 53Min: 39 / Avg: 66.36 / Max: 82Min: 33 / Avg: 62.56 / Max: 79Min: 43 / Avg: 61.88 / Max: 77Min: 48 / Avg: 67.06 / Max: 83Min: 37 / Avg: 55.81 / Max: 70Min: 32 / Avg: 56.83 / Max: 74Min: 45 / Avg: 52.93 / Max: 65Min: 28 / Avg: 47.45 / Max: 74Min: 26 / Avg: 39.71 / Max: 61Min: 35 / Avg: 46.18 / Max: 73Min: 36 / Avg: 52.37 / Max: 101

System Power Consumption Monitor

Phoronix Test Suite System Monitoring

OpenBenchmarking.orgWattsSystem Power Consumption MonitorPhoronix Test Suite System MonitoringGTX 1080GTX 1070GTX 1070 TiGTX 1080 TiTITAN RTXRTX 2080 TiRTX 2080RTX 2060RTX 2070RX Vega 64RX Vega 56RX 580RX 590Radeon VII70140210280350Min: 40.2 / Avg: 143.79 / Max: 318.2Min: 42.3 / Avg: 142.19 / Max: 280.5Min: 47.2 / Avg: 123.43 / Max: 255.8Min: 47.2 / Avg: 198.85 / Max: 370.3Min: 49.4 / Avg: 228.34 / Max: 363.1Min: 47.2 / Avg: 220.83 / Max: 367.7Min: 46 / Avg: 176.62 / Max: 334.5Min: 42.5 / Avg: 144.37 / Max: 300.1Min: 43.9 / Avg: 160.13 / Max: 317.6Min: 49.7 / Avg: 129.09 / Max: 305.4Min: 47.1 / Avg: 125.63 / Max: 271.3Min: 69 / Avg: 125.65 / Max: 229.5Min: 75 / Avg: 138.2 / Max: 253.3Min: 55.7 / Avg: 125.32 / Max: 337.5

Phoronix Test Suite v10.8.4