NCNN Vulkan - AMD vs. NVIDIA

NCNN Vulkan benchmarks 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 2009260-PTS-NCNNVULK08
Jump To Table - Results

View

Do Not Show Noisy Results
Do Not Show Results With Incomplete Data
Do Not Show Results With Little Change/Spread
List Notable Results
Show Result Confidence Charts
Allow Limiting Results To Certain Suite(s)

Statistics

Show Overall Harmonic Mean(s)
Show Overall Geometric Mean
Show Wins / Losses Counts (Pie Chart)
Normalize Results
Remove Outliers Before Calculating Averages

Graph Settings

Force Line Graphs Where Applicable
Convert To Scalar Where Applicable
Disable Color Branding
Prefer Vertical Bar Graphs
No Box Plots
On Line Graphs With Missing Data, Connect The Line Gaps

Multi-Way Comparison

Condense Multi-Option Tests Into Single Result Graphs
Condense Test Profiles With Multiple Version Results Into Single Result Graphs

Table

Show Detailed System Result Table

Run Management

Highlight
Result
Toggle/Hide
Result
Result
Identifier
Performance Per
Dollar
Date
Run
  Test
  Duration
GTX 1060
September 25 2020
  34 Minutes
GTX 1070
September 26 2020
  1 Hour, 20 Minutes
GTX 1080
September 25 2020
  23 Minutes
GTX 1650
September 25 2020
  38 Minutes
GTX 1650 SUPER
September 26 2020
  29 Minutes
GTX 1660
September 26 2020
  41 Minutes
GTX 1660 SUPER
September 26 2020
  23 Minutes
GTX 1660 Ti
September 26 2020
  1 Hour, 15 Minutes
RTX 2060
September 25 2020
  1 Hour, 7 Minutes
RTX 2060 SUPER
September 25 2020
  19 Minutes
RTX 2070
September 26 2020
  19 Minutes
RTX 2070 SUPER
September 25 2020
  1 Hour, 1 Minute
RTX 2080
September 26 2020
  1 Hour
RTX 2080 SUPER
September 26 2020
  58 Minutes
RTX 2080 Ti
September 26 2020
  55 Minutes
RX Vega 56
September 25 2020
  20 Minutes
RX 5600 XT
September 25 2020
  19 Minutes
RX 5700
September 25 2020
  18 Minutes
RX 5700 XT
September 25 2020
  1 Hour, 5 Minutes
Radeon VII
September 25 2020
  25 Minutes
Invert Behavior (Only Show Selected Data)
  42 Minutes

Only show results where is faster than
Only show results matching title/arguments (delimit multiple options with a comma):
Do not show results matching title/arguments (delimit multiple options with a comma):


NCNN Vulkan - AMD vs. NVIDIAProcessorMotherboardChipsetMemoryDiskGraphicsAudioMonitorNetworkOSKernelDesktopDisplay ServerDisplay DriverOpenGLOpenCLVulkanCompilerFile-SystemScreen ResolutionGTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX Vega 56RX 5600 XTRX 5700RX 5700 XTRadeon VIIAMD Ryzen 9 3950X 16-Core @ 3.50GHz (16 Cores / 32 Threads)ASUS ROG CROSSHAIR VIII HERO (WI-FI) (1302 BIOS)AMD Starship/Matisse16GB2000GB Corsair Force MP600 + 2000GBNVIDIA GeForce GTX 1060 6GB (1506/4006MHz)NVIDIA GP106 HD AudioDELL P2415QRealtek RTL8125 2.5GbE + Intel I211 + Intel Wi-Fi 6 AX200Ubuntu 20.045.4.0-48-generic (x86_64)GNOME Shell 3.36.4X Server 1.20.8NVIDIA 450.664.6.0OpenCL 2.0 AMD-APP (3182.0) + OpenCL 1.2 CUDA 11.0.2281.2.133GCC 9.3.0 + CUDA 11.0ext43840x2160NVIDIA GeForce GTX 1070 8GB (1506/4006MHz)NVIDIA GP104 HD AudioNVIDIA GeForce GTX 1080 8GB (1607/5005MHz)ASUS NVIDIA GeForce GTX 1650 4GB (1485/4001MHz)NVIDIA Device 10faASUS NVIDIA GeForce GTX 1650 SUPER 4GB (375/810MHz)NVIDIA TU116 HD AudioASUS NVIDIA GeForce GTX 1660 6GB (1530/4001MHz)eVGA NVIDIA GeForce GTX 1660 SUPER 6GB (1530/7000MHz)eVGA NVIDIA GeForce GTX 1660 Ti 6GB (1500/6000MHz)NVIDIA GeForce RTX 2060 6GB (1365/7000MHz)NVIDIA TU106 HD AudioNVIDIA GeForce RTX 2060 SUPER 8GB (1470/7000MHz)ASUS NVIDIA GeForce RTX 2070 8GB (420/405MHz)NVIDIA GeForce RTX 2070 SUPER 8GB (375/405MHz)NVIDIA TU104 HD AudioZotac NVIDIA GeForce RTX 2080 8GB (1515/7000MHz)NVIDIA GeForce RTX 2080 SUPER 8GB (405/405MHz)NVIDIA GeForce RTX 2080 Ti 11GB (1350/7000MHz)NVIDIA TU102 HD AudioAMD Radeon RX 56/64 8GB (1590/800MHz)AMD Vega 10 HDMI Audio5.9.0-050900rc6daily20200925-generic (x86_64) 202009244.6 Mesa 20.3.0-devel (git-3173367 2020-09-25 focal-oibaf-ppa) (LLVM 10.0.1)OpenCL 2.0 AMD-APP (3182.0)1.2.145Sapphire AMD Radeon RX 5600 OEM/5600 XT / 5700/5700 6GB (1780/875MHz)AMD Navi 10 HDMI AudioAMD Radeon RX 5600 OEM/5600 XT / 5700/5700 8GB (1750/875MHz)AMD Radeon RX 5600 OEM/5600 XT / 5700/5700 8GB (2100/875MHz)AMD Radeon VII 16GB (1801/1000MHz)AMD Vega 20 HDMI AudioOpenBenchmarking.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++,gm2 --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-targets=nvptx-none,hsa --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=auto --with-tune=generic --without-cuda-driver -vProcessor Details- Scaling Governor: acpi-cpufreq ondemand - CPU Microcode: 0x8701013Security Details- 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 STIBP: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected

GTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX Vega 56RX 5600 XTRX 5700RX 5700 XTRadeon VIIResult OverviewPhoronix Test Suite100%206%311%417%522%RealSR-NCNNNCNNNCNNRealSR-NCNNNCNNNCNNNCNNNCNNNCNNNCNNNCNNNCNNNCNNNCNNNCNNNCNN4x - YesVulkan GPU - vgg16Vulkan GPU - efficientnet-b04x - NoVulkan GPU - alexnetVulkan GPU-v3-v3 - mobilenet-v3Vulkan GPU - resnet50Vulkan GPU - resnet18Vulkan GPU - googlenetVulkan GPU - squeezenetVulkan GPU - mnasnetVulkan GPU-v2-v2 - mobilenet-v2Vulkan GPU - mobilenetVulkan GPU - shufflenet-v2Vulkan GPU - yolov4-tinyVulkan GPU - blazeface

NCNN Vulkan - AMD vs. NVIDIAncnn: Vulkan GPU - squeezenetncnn: 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 GPU - yolov4-tinyrealsr-ncnn: 4x - Yesrealsr-ncnn: 4x - NoGTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX Vega 56RX 5600 XTRX 5700RX 5700 XTRadeon VII7.775.991.972.421.602.073.650.705.4520.203.135.408.0010.52143.94920.5436.525.471.852.321.551.963.490.714.5814.982.393.845.919.15102.78715.3745.835.262.232.251.471.933.400.714.1315.042.013.784.918.84116.45217.2146.836.291.942.291.572.173.500.705.3124.153.373.748.4211.95215.51829.3725.805.421.601.881.401.633.210.644.0921.552.612.955.8310.67186.17925.6585.525.571.772.141.512.003.190.674.3316.982.373.055.759.92138.72119.8935.205.091.481.981.331.792.870.633.9415.42.382.755.029.41120.14817.5385.245.131.611.831.461.632.840.663.7815.332.262.745.049.45117.67417.1504.724.901.541.791.421.642.840.693.5611.082.002.434.268.7486.37213.4004.604.621.451.931.321.682.610.623.3610.311.722.223.848.5476.37312.1294.674.831.451.701.321.772.890.623.719.641.732.243.888.4275.43712.0074.334.651.441.751.361.542.690.663.238.471.682.063.758.3063.86310.6154.284.651.491.761.441.512.760.713.338.051.631.983.578.2260.36010.0614.304.451.481.621.391.512.710.633.237.631.511.933.538.1254.8409.4634.014.531.441.731.361.482.730.673.055.601.421.613.147.7744.9868.2604.896.312.954.652.343.0910.540.875.7411.562.204.386.547.6766.10810.2543.586.542.153.111.842.337.250.764.2415.241.735.245.028.5380.26012.0793.586.602.183.131.862.257.850.774.2714.911.715.385.248.4958.9299.3583.366.582.052.961.732.086.630.733.8713.701.574.814.468.3255.6788.9825.039.022.844.132.603.1511.070.927.2210.422.043.786.6210.0041.2887.177OpenBenchmarking.org

GPU Power Consumption Monitor

OpenBenchmarking.orgWattsGPU Power Consumption MonitorPhoronix Test Suite System MonitoringGTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII50100150200250Min: 5.95 / Avg: 86 / Max: 131.08Min: 6.75 / Avg: 92.52 / Max: 162.74Min: 6.57 / Avg: 115.01 / Max: 188.22Min: 5.43 / Avg: 53.96 / Max: 68.86Min: 6.51 / Avg: 71.29 / Max: 96.43Min: 6.69 / Avg: 72.09 / Max: 109.23Min: 10.75 / Avg: 90.37 / Max: 129.22Min: 7.41 / Avg: 78.56 / Max: 133.19Min: 9.49 / Avg: 91.04 / Max: 164.41Min: 10.34 / Avg: 108.02 / Max: 179.76Min: 7.53 / Avg: 105.98 / Max: 181.01Min: 14.42 / Avg: 107.67 / Max: 220.78Min: 13.31 / Avg: 108.43 / Max: 221.85Min: 8.89 / Avg: 110.83 / Max: 255.3Min: 7.79 / Avg: 115.51 / Max: 274.43Min: 22 / Avg: 98.3 / Max: 171Min: 31 / Avg: 89.52 / Max: 165Min: 29 / Avg: 88.95 / Max: 225Min: 11 / Avg: 96.87 / Max: 167Min: 21 / Avg: 104.49 / Max: 267

GPU Temperature Monitor

OpenBenchmarking.orgCelsiusGPU Temperature MonitorPhoronix Test Suite System MonitoringGTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII1632486480Min: 27 / Avg: 55.8 / Max: 69Min: 29 / Avg: 60.05 / Max: 78Min: 29 / Avg: 62.17 / Max: 79Min: 31 / Avg: 56.9 / Max: 66Min: 30 / Avg: 55.56 / Max: 66Min: 29 / Avg: 60.56 / Max: 79Min: 33 / Avg: 59.37 / Max: 70Min: 29 / Avg: 52.99 / Max: 69Min: 28 / Avg: 50.87 / Max: 70Min: 29 / Avg: 52.23 / Max: 71Min: 29 / Avg: 54.07 / Max: 74Min: 26 / Avg: 46.54 / Max: 68Min: 29 / Avg: 58.81 / Max: 82Min: 28 / Avg: 48.45 / Max: 72Min: 32 / Avg: 49.95 / Max: 71Min: 44 / Avg: 57.67 / Max: 66Min: 38 / Avg: 52.05 / Max: 67Min: 35 / Avg: 52.14 / Max: 74Min: 29 / Avg: 51.18 / Max: 65Min: 29 / Avg: 44.95 / Max: 66

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20200916Target: Vulkan GPU - Model: squeezenetGTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII246810SE +/- 0.11, N = 4SE +/- 0.21, N = 15SE +/- 0.02, N = 3SE +/- 0.09, N = 4SE +/- 0.04, N = 3SE +/- 0.06, N = 7SE +/- 0.03, N = 3SE +/- 0.17, N = 15SE +/- 0.05, N = 15SE +/- 0.05, N = 3SE +/- 0.07, N = 3SE +/- 0.05, N = 15SE +/- 0.04, N = 15SE +/- 0.03, N = 15SE +/- 0.05, N = 15SE +/- 0.02, N = 3SE +/- 0.01, N = 3SE +/- 0.03, N = 14SE +/- 0.00, N = 3SE +/- 0.07, N = 47.776.525.836.835.805.525.205.244.724.604.674.334.284.304.013.583.583.364.895.031. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20200916Target: Vulkan GPU - Model: mobilenetGTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII3691215SE +/- 0.08, N = 4SE +/- 0.02, N = 15SE +/- 0.03, N = 3SE +/- 0.02, N = 4SE +/- 0.04, N = 3SE +/- 0.01, N = 7SE +/- 0.12, N = 3SE +/- 0.04, N = 15SE +/- 0.05, N = 15SE +/- 0.10, N = 3SE +/- 0.03, N = 3SE +/- 0.04, N = 15SE +/- 0.04, N = 15SE +/- 0.05, N = 15SE +/- 0.03, N = 14SE +/- 0.03, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 14SE +/- 0.06, N = 3SE +/- 0.06, N = 45.995.475.266.295.425.575.095.134.904.624.834.654.654.454.536.546.606.586.319.021. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20200916Target: Vulkan GPU-v2-v2 - Model: mobilenet-v2GTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII0.66381.32761.99142.65523.319SE +/- 0.02, N = 4SE +/- 0.02, N = 15SE +/- 0.16, N = 3SE +/- 0.00, N = 4SE +/- 0.02, N = 3SE +/- 0.06, N = 7SE +/- 0.01, N = 3SE +/- 0.04, N = 15SE +/- 0.04, N = 15SE +/- 0.04, N = 3SE +/- 0.02, N = 3SE +/- 0.03, N = 15SE +/- 0.06, N = 15SE +/- 0.05, N = 15SE +/- 0.04, N = 15SE +/- 0.00, N = 3SE +/- 0.03, N = 3SE +/- 0.07, N = 14SE +/- 0.36, N = 3SE +/- 0.01, N = 41.971.852.231.941.601.771.481.611.541.451.451.441.491.481.442.152.182.052.952.841. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20200916Target: Vulkan GPU-v3-v3 - Model: mobilenet-v3GTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII1.04632.09263.13894.18525.2315SE +/- 0.09, N = 4SE +/- 0.05, N = 15SE +/- 0.11, N = 3SE +/- 0.12, N = 4SE +/- 0.12, N = 3SE +/- 0.06, N = 7SE +/- 0.15, N = 3SE +/- 0.04, N = 15SE +/- 0.04, N = 15SE +/- 0.11, N = 3SE +/- 0.02, N = 3SE +/- 0.04, N = 15SE +/- 0.06, N = 15SE +/- 0.03, N = 15SE +/- 0.05, N = 15SE +/- 0.01, N = 3SE +/- 0.00, N = 3SE +/- 0.12, N = 14SE +/- 0.93, N = 3SE +/- 0.12, N = 42.422.322.252.291.882.141.981.831.791.931.701.751.761.621.733.113.132.964.654.131. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20200916Target: Vulkan GPU - Model: shufflenet-v2GTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII0.5851.171.7552.342.925SE +/- 0.01, N = 4SE +/- 0.02, N = 15SE +/- 0.03, N = 3SE +/- 0.00, N = 3SE +/- 0.02, N = 3SE +/- 0.05, N = 7SE +/- 0.01, N = 3SE +/- 0.05, N = 15SE +/- 0.04, N = 15SE +/- 0.03, N = 3SE +/- 0.02, N = 3SE +/- 0.04, N = 15SE +/- 0.08, N = 15SE +/- 0.05, N = 15SE +/- 0.04, N = 15SE +/- 0.01, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 14SE +/- 0.01, N = 3SE +/- 0.00, N = 41.601.551.471.571.401.511.331.461.421.321.321.361.441.391.361.841.861.732.342.601. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20200916Target: Vulkan GPU - Model: mnasnetGTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII0.70881.41762.12642.83523.544SE +/- 0.03, N = 4SE +/- 0.03, N = 15SE +/- 0.12, N = 3SE +/- 0.10, N = 4SE +/- 0.01, N = 3SE +/- 0.07, N = 7SE +/- 0.13, N = 3SE +/- 0.04, N = 15SE +/- 0.05, N = 15SE +/- 0.17, N = 2SE +/- 0.14, N = 3SE +/- 0.04, N = 15SE +/- 0.03, N = 15SE +/- 0.05, N = 15SE +/- 0.04, N = 14SE +/- 0.08, N = 3SE +/- 0.00, N = 3SE +/- 0.02, N = 14SE +/- 0.31, N = 3SE +/- 0.01, N = 42.071.961.932.171.632.001.791.631.641.681.771.541.511.511.482.332.252.083.093.151. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20200916Target: Vulkan GPU - Model: efficientnet-b0GTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII3691215SE +/- 0.10, N = 4SE +/- 0.04, N = 15SE +/- 0.10, N = 3SE +/- 0.08, N = 4SE +/- 0.25, N = 3SE +/- 0.05, N = 7SE +/- 0.12, N = 3SE +/- 0.04, N = 15SE +/- 0.04, N = 15SE +/- 0.01, N = 3SE +/- 0.11, N = 3SE +/- 0.04, N = 15SE +/- 0.04, N = 15SE +/- 0.05, N = 15SE +/- 0.08, N = 15SE +/- 0.12, N = 3SE +/- 0.05, N = 3SE +/- 0.05, N = 14SE +/- 0.19, N = 3SE +/- 0.06, N = 43.653.493.403.503.213.192.872.842.842.612.892.692.762.712.737.257.856.6310.5411.071. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20200916Target: Vulkan GPU - Model: blazefaceGTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII0.2070.4140.6210.8281.035SE +/- 0.01, N = 4SE +/- 0.03, N = 15SE +/- 0.05, N = 3SE +/- 0.01, N = 4SE +/- 0.01, N = 3SE +/- 0.00, N = 7SE +/- 0.00, N = 3SE +/- 0.03, N = 15SE +/- 0.04, N = 15SE +/- 0.00, N = 3SE +/- 0.01, N = 3SE +/- 0.03, N = 15SE +/- 0.04, N = 15SE +/- 0.02, N = 15SE +/- 0.05, N = 15SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.00, N = 14SE +/- 0.01, N = 3SE +/- 0.00, N = 40.700.710.710.700.640.670.630.660.690.620.620.660.710.630.670.760.770.730.870.921. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20200916Target: Vulkan GPU - Model: googlenetGTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII246810SE +/- 0.01, N = 4SE +/- 0.04, N = 15SE +/- 0.06, N = 3SE +/- 0.01, N = 4SE +/- 0.16, N = 3SE +/- 0.06, N = 7SE +/- 0.17, N = 3SE +/- 0.04, N = 15SE +/- 0.04, N = 15SE +/- 0.09, N = 3SE +/- 0.22, N = 3SE +/- 0.06, N = 15SE +/- 0.07, N = 15SE +/- 0.05, N = 15SE +/- 0.04, N = 15SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 14SE +/- 0.01, N = 3SE +/- 0.15, N = 45.454.584.135.314.094.333.943.783.563.363.713.233.333.233.054.244.273.875.747.221. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20200916Target: Vulkan GPU - Model: vgg16GTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII612182430SE +/- 0.19, N = 4SE +/- 0.08, N = 15SE +/- 0.14, N = 3SE +/- 0.03, N = 4SE +/- 0.27, N = 3SE +/- 0.12, N = 7SE +/- 0.17, N = 3SE +/- 0.09, N = 15SE +/- 0.06, N = 15SE +/- 0.56, N = 3SE +/- 0.22, N = 3SE +/- 0.03, N = 15SE +/- 0.06, N = 15SE +/- 0.06, N = 15SE +/- 0.03, N = 15SE +/- 0.15, N = 3SE +/- 0.04, N = 3SE +/- 0.08, N = 14SE +/- 0.13, N = 3SE +/- 0.11, N = 420.2014.9815.0424.1521.5516.9815.4015.3311.0810.319.648.478.057.635.6015.2414.9113.7011.5610.421. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20200916Target: Vulkan GPU - Model: resnet18GTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII0.75831.51662.27493.03323.7915SE +/- 0.06, N = 4SE +/- 0.03, N = 15SE +/- 0.01, N = 3SE +/- 0.06, N = 4SE +/- 0.02, N = 3SE +/- 0.01, N = 7SE +/- 0.12, N = 3SE +/- 0.03, N = 15SE +/- 0.04, N = 15SE +/- 0.04, N = 3SE +/- 0.03, N = 3SE +/- 0.04, N = 15SE +/- 0.04, N = 14SE +/- 0.02, N = 15SE +/- 0.06, N = 14SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 14SE +/- 0.01, N = 3SE +/- 0.00, N = 43.132.392.013.372.612.372.382.262.001.721.731.681.631.511.421.731.711.572.202.041. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20200916Target: Vulkan GPU - Model: alexnetGTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII1.2152.433.6454.866.075SE +/- 0.02, N = 4SE +/- 0.04, N = 14SE +/- 0.04, N = 3SE +/- 0.03, N = 4SE +/- 0.03, N = 3SE +/- 0.06, N = 7SE +/- 0.03, N = 3SE +/- 0.04, N = 15SE +/- 0.04, N = 15SE +/- 0.12, N = 3SE +/- 0.10, N = 3SE +/- 0.02, N = 15SE +/- 0.01, N = 14SE +/- 0.03, N = 15SE +/- 0.02, N = 15SE +/- 0.04, N = 3SE +/- 0.02, N = 3SE +/- 0.01, N = 14SE +/- 0.00, N = 3SE +/- 0.00, N = 45.403.843.783.742.953.052.752.742.432.222.242.061.981.931.615.245.384.814.383.781. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20200916Target: Vulkan GPU - Model: resnet50GTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII246810SE +/- 0.09, N = 4SE +/- 0.04, N = 15SE +/- 0.10, N = 3SE +/- 0.18, N = 4SE +/- 0.03, N = 3SE +/- 0.02, N = 7SE +/- 0.13, N = 3SE +/- 0.03, N = 14SE +/- 0.04, N = 15SE +/- 0.08, N = 3SE +/- 0.09, N = 3SE +/- 0.04, N = 15SE +/- 0.04, N = 15SE +/- 0.04, N = 15SE +/- 0.04, N = 15SE +/- 0.04, N = 3SE +/- 0.16, N = 3SE +/- 0.00, N = 14SE +/- 0.07, N = 3SE +/- 0.08, N = 48.005.914.918.425.835.755.025.044.263.843.883.753.573.533.145.025.244.466.546.621. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20200916Target: Vulkan GPU - Model: yolov4-tinyGTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII3691215SE +/- 0.08, N = 4SE +/- 0.04, N = 15SE +/- 0.11, N = 3SE +/- 0.11, N = 4SE +/- 0.08, N = 3SE +/- 0.12, N = 7SE +/- 0.12, N = 3SE +/- 0.04, N = 15SE +/- 0.05, N = 15SE +/- 0.10, N = 3SE +/- 0.07, N = 3SE +/- 0.06, N = 15SE +/- 0.06, N = 15SE +/- 0.08, N = 14SE +/- 0.07, N = 15SE +/- 0.03, N = 3SE +/- 0.02, N = 3SE +/- 0.01, N = 14SE +/- 0.02, N = 3SE +/- 0.01, N = 410.529.158.8411.9510.679.929.419.458.748.548.428.308.228.127.778.538.498.327.6710.001. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

RealSR-NCNN

RealSR-NCNN is an NCNN neural network implementation of the RealSR project and accelerated using the Vulkan API. RealSR is the Real-World Super Resolution via Kernel Estimation and Noise Injection. NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. This test profile times how long it takes to increase the resolution of a sample image by a scale of 4x with Vulkan. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterRealSR-NCNN 20200818Scale: 4x - TAA: YesGTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII50100150200250SE +/- 0.33, N = 3SE +/- 0.36, N = 3SE +/- 0.08, N = 3SE +/- 0.29, N = 3SE +/- 0.09, N = 3SE +/- 0.27, N = 3SE +/- 0.07, N = 3SE +/- 0.19, N = 3SE +/- 0.26, N = 3SE +/- 0.43, N = 3SE +/- 0.56, N = 3SE +/- 0.31, N = 3SE +/- 0.38, N = 3SE +/- 0.19, N = 3SE +/- 0.23, N = 3SE +/- 0.13, N = 3SE +/- 0.04, N = 3SE +/- 0.14, N = 3SE +/- 0.08, N = 3SE +/- 0.12, N = 3143.95102.79116.45215.52186.18138.72120.15117.6786.3776.3775.4463.8660.3654.8444.9980.2658.9355.6866.1141.29

OpenBenchmarking.orgSeconds, Fewer Is BetterRealSR-NCNN 20200818Scale: 4x - TAA: NoGTX 1060GTX 1070GTX 1080GTX 1650GTX 1650 SUPERGTX 1660GTX 1660 SUPERGTX 1660 TiRTX 2060RTX 2060 SUPERRTX 2070RTX 2070 SUPERRTX 2080RTX 2080 SUPERRTX 2080 TiRX 5600 XTRX 5700RX 5700 XTRX Vega 56Radeon VII714212835SE +/- 0.020, N = 3SE +/- 0.035, N = 4SE +/- 0.068, N = 3SE +/- 0.039, N = 3SE +/- 0.050, N = 3SE +/- 0.045, N = 3SE +/- 0.010, N = 3SE +/- 0.030, N = 3SE +/- 0.021, N = 4SE +/- 0.039, N = 4SE +/- 0.039, N = 4SE +/- 0.034, N = 5SE +/- 0.038, N = 5SE +/- 0.041, N = 5SE +/- 0.020, N = 6SE +/- 0.024, N = 4SE +/- 0.035, N = 5SE +/- 0.020, N = 5SE +/- 0.021, N = 5SE +/- 0.018, N = 620.54315.37417.21429.37225.65819.89317.53817.15013.40012.12912.00710.61510.0619.4638.26012.0799.3588.98210.2547.177