Linux GPU Compute

Intel Core i9-12900K testing with a ASUS ROG STRIX Z690-E GAMING WIFI (1601 BIOS) and NVIDIA GeForce RTX 3070 Ti 8GB on Ubuntu 22.04 via the Phoronix Test Suite.

Compare your own system(s) to this result file with the Phoronix Test Suite by running the command: phoronix-test-suite benchmark 2208115-NE-GPUCOMPUT74
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
RTX 3060
July 10 2022
  3 Hours, 51 Minutes
RTX 3060 Ti
July 12 2022
  3 Hours, 39 Minutes
RTX 3070
July 12 2022
  3 Hours, 20 Minutes
RTX 3070 Ti
July 12 2022
  3 Hours, 17 Minutes
RTX 3080
July 12 2022
  3 Hours, 14 Minutes
RTX 3080 Ti
July 12 2022
  3 Hours, 43 Minutes
RTX 3090
July 08 2022
  3 Hours, 39 Minutes
Invert Behavior (Only Show Selected Data)
  3 Hours, 32 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):


Linux GPU ComputeProcessorMotherboardChipsetMemoryDiskGraphicsAudioMonitorNetworkOSKernelDesktopDisplay ServerDisplay DriverOpenGLOpenCLVulkanCompilerFile-SystemScreen ResolutionRTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 3080 TiRTX 3090Intel Core i9-12900K @ 5.20GHz (16 Cores / 24 Threads)ASUS ROG STRIX Z690-E GAMING WIFI (1601 BIOS)Intel Alder Lake-S PCH32GB1000GB Western Digital WDS100T1X0E-00AFY0eVGA NVIDIA GeForce RTX 3060 12GBIntel Alder Lake-S HD AudioASUS VP28UIntel I225-V + Intel Wi-Fi 6 AX210/AX211/AX411Ubuntu 22.045.18.0-051800-generic (x86_64)GNOME Shell 42.2X Server 1.21.1.3NVIDIA 515.49.064.6.0OpenCL 3.0 CUDA 11.7.991.3.217GCC 11.2.0ext43840x21601000GB Western Digital WDS100T1X0E-00AFY0 + 2000GBNVIDIA GeForce RTX 3060 Ti 8GBNVIDIA GeForce RTX 3070 8GBNVIDIA GeForce RTX 3070 Ti 8GBNVIDIA GeForce RTX 3080 10GBNVIDIA GeForce RTX 3080 Ti 12GBIntel Device 7aa7NVIDIA GeForce RTX 3090 24GBIntel Device 7ad0OpenBenchmarking.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,brig,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-link-serialization=2 --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-targets=nvptx-none=/build/gcc-11-gBFGDP/gcc-11-11.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-11-gBFGDP/gcc-11-11.2.0/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: intel_pstate performance (EPP: performance) - CPU Microcode: 0x1f - Thermald 2.4.9Graphics Details- RTX 3060: BAR1 / Visible vRAM Size: 16384 MiB - vBIOS Version: 94.06.14.40.46- RTX 3060 Ti: BAR1 / Visible vRAM Size: 8192 MiB - vBIOS Version: 94.04.25.00.2c- RTX 3070: BAR1 / Visible vRAM Size: 8192 MiB - vBIOS Version: 94.04.25.00.2b- RTX 3070 Ti: BAR1 / Visible vRAM Size: 8192 MiB - vBIOS Version: 94.04.5b.00.02- RTX 3080: BAR1 / Visible vRAM Size: 16384 MiB - vBIOS Version: 94.02.20.00.07- RTX 3080 Ti: BAR1 / Visible vRAM Size: 16384 MiB - vBIOS Version: 94.02.71.00.01- RTX 3090: BAR1 / Visible vRAM Size: 32768 MiB - vBIOS Version: 94.02.27.00.02OpenCL Details- RTX 3060: GPU Compute Cores: 3584- RTX 3060 Ti: GPU Compute Cores: 4864- RTX 3070: GPU Compute Cores: 5888- RTX 3070 Ti: GPU Compute Cores: 6144- RTX 3080: GPU Compute Cores: 8704- RTX 3080 Ti: GPU Compute Cores: 10240- RTX 3090: GPU Compute Cores: 10496Security Details- itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: 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 IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected Python Details- RTX 3090: Python 3.10.4

RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 3080 TiRTX 3090Result OverviewPhoronix Test Suite100%151%202%253%303%vkpeakHashcatVkResampleLuxCoreRenderLeelaChessZeroChaos Group V-RAYMandelGPUOctaneBenchIndigoBenchFinanceBenchRealSR-NCNNcl-memFAHBenchViennaCLclpeakSHOC Scalable HeterOgeneous ComputingDarktableRodiniaWaifu2x-NCNN VulkanJuliaGPUSmallPT GPU

RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 3080 TiRTX 3090Logarithmic Per Watt Result OverviewPhoronix Test SuiteMeta Performance Per WattsSmallPT GPUJuliaGPUSHOC Scalable HeterOgeneous ComputingclpeakFAHBenchHashcatMandelGPUChaos Group V-RAYIndigoBenchLuxCoreRenderLeelaChessZeroOctaneBenchcl-memP.W.G.MP.W.G.MP.W.G.MP.W.G.MP.W.G.MP.W.G.MP.W.G.MP.W.G.MP.W.G.MP.W.G.MP.W.G.MP.W.G.MP.W.G.MP.W.G.M

Linux GPU Computevkpeak: fp64-vec4vkpeak: fp64-scalarvkpeak: fp32-scalarvkpeak: int32-vec4vkpeak: int32-scalarvkpeak: int16-scalarvkpeak: fp16-scalarvkpeak: fp32-vec4vkpeak: fp16-vec4clpeak: Double-Precision Doubleshoc: OpenCL - GEMM SGEMM_Nviennacl: OpenCL BLAS - dGEMM-NNviennacl: OpenCL BLAS - dGEMM-NTviennacl: OpenCL BLAS - dGEMM-TNshoc: OpenCL - MD5 Hashhashcat: SHA1hashcat: SHA-512vkpeak: int16-vec4clpeak: Integer Compute INThashcat: MD5hashcat: TrueCrypt RIPEMD160 + XTSshoc: OpenCL - FFT SPclpeak: Single-Precision Floathashcat: 7-Zipluxcorerender: LuxCore Benchmark - GPUdarktable: Server Room - OpenCLluxcorerender: DLSC - GPUluxcorerender: Danish Mood - GPUfinancebench: Black-Scholes OpenCLrodinia: OpenCL Particle Filterclpeak: Global Memory Bandwidthcl-mem: Readshoc: OpenCL - S3Dv-ray: NVIDIA RTX GPUcl-mem: Writevkresample: 2x - Singlelczero: OpenCLv-ray: NVIDIA CUDA GPUindigobench: OpenCL GPU - Bedroomrealsr-ncnn: 4x - Yesmandelgpu: GPUoctanebench: Total Scoredarktable: Boat - OpenCLluxcorerender: Rainbow Colors and Prism - GPUviennacl: OpenCL BLAS - dAXPYluxcorerender: Orange Juice - GPUindigobench: OpenCL GPU - Supercarviennacl: OpenCL BLAS - dDOTfinancebench: Monte-Carlo OpenCLviennacl: OpenCL BLAS - dCOPYfahbench: realsr-ncnn: 4x - Noshoc: OpenCL - Texture Read Bandwidthviennacl: OpenCL BLAS - sAXPYwaifu2x-ncnn: 2x - 3 - Yesdarktable: Masskrug - OpenCLcl-mem: Copyviennacl: OpenCL BLAS - sCOPYjuliagpu: GPUviennacl: OpenCL BLAS - sDOTshoc: OpenCL - Reductionviennacl: OpenCL BLAS - dGEMV-Tviennacl: OpenCL BLAS - dGEMM-TTclpeak: Kernel Latencyviennacl: OpenCL BLAS - dGEMV-Nrodinia: OpenCL Myocyteshoc: OpenCL - Triadclpeak: Transfer Bandwidth enqueueWriteBuffershoc: OpenCL - Bus Speed Downloadclpeak: Transfer Bandwidth enqueueReadBuffersmallpt-gpu: GPU - Complexsmallpt-gpu: GPU - Cornellsmallpt-gpu: GPU - Caustic3shoc: OpenCL - Bus Speed Readbackdarktable: Server Rack - OpenCLRTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 3080 TiRTX 3090210.82212.356808.216769.026799.664494.326820.269036.9813484.60212.042778.3420420420315.124677354666679949833335955.726347.1424885966667287638758.39712563.044137894.051.8654.293.2616.2444669.521315.43318.0171.9461163299.923.184251498558.50768.692215226018.8294.8620192.57313.503274.8725.928331622.231995316186.724610.4511269.693074.8992.626270.3276378413490.2292307.5673043.7722328.56523.352421.3725.100218.2616576198511657619984165762012626.35220.126304.44304.189750.759663.579705.336413.899725.9012877.5519166.40305.243435.6329129229121.56401098773333314121333338487.358397.0435221800000417944962.41716043.125782505.391.8575.734.3512.8506.927389.15392.9212.4701632385.917.68937359126311.50451.555285067246.1387.1688672.57118.073996.1434.081402465.278429372245.79548.3282137.683584.1902.604297.3297411524637.1330325.4963353.8122427.64223.689021.8024.692718.1716576602611657660394165766053626.35460.122362.02361.3511587.4111509.5811562.597639.1811579.7115337.2622752.71361.763802.5334034134025.326912963333333166498333310095.6310206.34413691571434937751136.2620056.496697446.221.8476.615.1110.3795.956389.39392.9219.1051787382.617.55840342141612.97047.301321865318.9417.9146852.57820.333976.9837.394401465.296714368274.56947.7922150.103594.0262.606296.4285426397082.5328327.4103353.7522228.13923.661721.9224.720518.1016576283591657628493165762863526.35430.126367.88366.4011705.3911651.0411695.997705.9311766.5015490.3423227.90370.114697.6734734834726.213413271300000170406666710244.7510945.90426733714295048781292.0921518.747125446.611.8447.005.429.7305.771529.26539.5285.7871990518.013.53747753150513.79841.326329682482.6461.8634092.57921.085297.2139.128527474.089434482275.12107.0572102.794373.7182.605338.6337427836038.7370381.5593713.7223828.40424.036021.7724.712518.1816576703471657670481165767062426.35460.124539.15539.8417334.5317105.4017172.3711320.9017271.0522809.3634090.93545.586356.2350450650237.534618975633333242603333314594.4515397.27599994857147120781763.5829616.929760229.141.8499.527.327.3524.189662.25673.3338.6892436647.611.27256188178817.71832.853428359837.3574.8658832.57726.546329.0346.999605388.820714558333.58916.0172227.534773.3372.612354.6359459314633.0375390.6263795063.7824327.98024.250621.8824.698118.1716576379191657638055165763819926.35420.126612.78614.2319739.0319459.2719552.2212873.8419662.1026094.9738772.75625.427981.9056156355941.872021421483333272865000016331.7617218.71676895333338000502016.8633795.91108666710.511.87010.988.456.6813.757794.66794.4419.8672840738.89.57061726206420.28829.819461789875.5677.7082812.57930.4870710.2351.903654326.089142591345.59995.6412182.894923.2682.618356.2359472565820.3367383.7123695653.8623228.55824.388721.8924.701218.1316576487381657648873165764901726.35390.128653.38651.7720867.7820669.2220754.4813699.5720781.4127459.9140968.90640.168197.2759259459043.548422100850000282036666716879.0617830.00698872333337981752103.7334509.53110917810.850.70711.288.546.3873.957795.98793.7429.4262869730.29.42261151208420.68329.185501044446.7682.296811.11831.0171710.3452.712644325.266715602353.44835.5872246.054983.2661.921357.8362492011904.2374389.5823745934.0923528.94024.436721.3625.033018.2616573388921657339028165733917226.35270.124OpenBenchmarking.org

Meta Performance Per Watts

OpenBenchmarking.orgPerformance Per Watts, More Is BetterMeta Performance Per WattsPerformance Per WattsRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30602K4K6K8K10K9893.969408.698566.056518.356071.235548.851454.12

vkpeak

Vkpeak is a Vulkan compute benchmark inspired by OpenCL's clpeak. Vkpeak provides Vulkan compute performance measurements for FP16 / FP32 / FP64 / INT16 / INT32 scalar and vec4 performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS, More Is Bettervkpeak 20210424fp64-vec4RTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060140280420560700SE +/- 0.03, N = 3SE +/- 1.42, N = 3SE +/- 2.58, N = 3SE +/- 0.01, N = 3SE +/- 0.42, N = 3SE +/- 0.47, N = 3SE +/- 0.00, N = 2653.38612.78539.15367.88362.02304.44210.82

OpenBenchmarking.orgGFLOPS, More Is Bettervkpeak 20210424fp64-scalarRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060140280420560700SE +/- 1.62, N = 3SE +/- 1.72, N = 3SE +/- 2.71, N = 3SE +/- 0.74, N = 3SE +/- 0.34, N = 3SE +/- 0.72, N = 3SE +/- 0.74, N = 3651.77614.23539.84366.40361.35304.18212.35

OpenBenchmarking.orgGFLOPS, More Is Bettervkpeak 20210424fp32-scalarRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30604K8K12K16K20KSE +/- 73.54, N = 3SE +/- 200.38, N = 3SE +/- 159.53, N = 3SE +/- 22.27, N = 3SE +/- 58.23, N = 3SE +/- 44.58, N = 3SE +/- 41.90, N = 320867.7819739.0317334.5311705.3911587.419750.756808.21

OpenBenchmarking.orgGIOPS, More Is Bettervkpeak 20210424int32-vec4RTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30604K8K12K16K20KSE +/- 0.25, N = 3SE +/- 49.71, N = 3SE +/- 84.25, N = 3SE +/- 0.41, N = 3SE +/- 16.41, N = 3SE +/- 18.74, N = 3SE +/- 0.10, N = 320669.2219459.2717105.4011651.0411509.589663.576769.02

OpenBenchmarking.orgGIOPS, More Is Bettervkpeak 20210424int32-scalarRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30604K8K12K16K20KSE +/- 0.42, N = 3SE +/- 51.14, N = 3SE +/- 86.56, N = 3SE +/- 0.79, N = 3SE +/- 15.80, N = 3SE +/- 18.16, N = 3SE +/- 0.32, N = 320754.4819552.2217172.3711695.9911562.599705.336799.66

OpenBenchmarking.orgGIOPS, More Is Bettervkpeak 20210424int16-scalarRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30603K6K9K12K15KSE +/- 0.10, N = 3SE +/- 0.07, N = 3SE +/- 28.42, N = 3SE +/- 0.41, N = 3SE +/- 11.50, N = 3SE +/- 11.82, N = 3SE +/- 0.20, N = 313699.5712873.8411320.907705.937639.186413.894494.32

OpenBenchmarking.orgGFLOPS, More Is Bettervkpeak 20210424fp16-scalarRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30604K8K12K16K20KSE +/- 0.16, N = 3SE +/- 135.01, N = 3SE +/- 86.43, N = 3SE +/- 43.21, N = 3SE +/- 30.65, N = 3SE +/- 16.37, N = 3SE +/- 16.13, N = 320781.4119662.1017271.0511766.5011579.719725.906820.26

OpenBenchmarking.orgGFLOPS, More Is Bettervkpeak 20210424fp32-vec4RTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30606K12K18K24K30KSE +/- 66.30, N = 3SE +/- 179.42, N = 3SE +/- 50.25, N = 3SE +/- 0.80, N = 3SE +/- 65.69, N = 3SE +/- 45.93, N = 3SE +/- 2.30, N = 327459.9126094.9722809.3615490.3415337.2612877.559036.98

OpenBenchmarking.orgGFLOPS, More Is Bettervkpeak 20210424fp16-vec4RTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30609K18K27K36K45KSE +/- 101.51, N = 3SE +/- 100.05, N = 3SE +/- 170.86, N = 3SE +/- 84.42, N = 3SE +/- 11.98, N = 3SE +/- 30.18, N = 3SE +/- 1.19, N = 340968.9038772.7534090.9323227.9022752.7119166.4013484.60

clpeak

Clpeak is designed to test the peak capabilities of OpenCL devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS, More Is BetterclpeakOpenCL Test: Double-Precision DoubleRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060140280420560700SE +/- 1.62, N = 3SE +/- 2.30, N = 3SE +/- 1.73, N = 3SE +/- 0.96, N = 3SE +/- 1.02, N = 3SE +/- 0.47, N = 3SE +/- 0.85, N = 3640.16625.42545.58370.11361.76305.24212.041. (CXX) g++ options: -O3 -rdynamic -lOpenCL

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: GEMM SGEMM_NRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30602K4K6K8K10KSE +/- 29.85, N = 12SE +/- 24.71, N = 12SE +/- 23.13, N = 11SE +/- 6.05, N = 11SE +/- 16.49, N = 10SE +/- 3.49, N = 10SE +/- 5.73, N = 98197.277981.906356.234697.673802.533435.632778.341. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPs/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dGEMM-NNRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060130260390520650SE +/- 1.00, N = 3SE +/- 1.33, N = 3SE +/- 1.00, N = 3SE +/- 1.00, N = 3SE +/- 0.33, N = 3SE +/- 1.00, N = 3SE +/- 0.67, N = 35925615043473402912041. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

OpenBenchmarking.orgGFLOPs/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dGEMM-NTRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060130260390520650SE +/- 1.00, N = 3SE +/- 1.33, N = 3SE +/- 1.33, N = 3SE +/- 1.00, N = 3SE +/- 0.00, N = 2SE +/- 1.50, N = 2SE +/- 0.00, N = 25945635063483412922041. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

OpenBenchmarking.orgGFLOPs/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dGEMM-TNRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060130260390520650SE +/- 0.88, N = 3SE +/- 1.33, N = 3SE +/- 1.20, N = 3SE +/- 0.50, N = 25905595023483402912041. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGHash/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: MD5 HashRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30601020304050SE +/- 0.03, N = 14SE +/- 0.05, N = 14SE +/- 0.03, N = 14SE +/- 0.00, N = 13SE +/- 0.00, N = 13SE +/- 0.01, N = 13SE +/- 0.00, N = 1543.5541.8737.5326.2125.3321.5615.121. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

Hashcat

Hashcat is an open-source, advanced password recovery tool supporting GPU acceleration with OpenCL, NVIDIA CUDA, and Radeon ROCm. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: SHA1RTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30605000M10000M15000M20000M25000MSE +/- 24085870.69, N = 6SE +/- 21246072.84, N = 6SE +/- 14809381.41, N = 6SE +/- 3912288.33, N = 6SE +/- 11473321.71, N = 6SE +/- 5257481.44, N = 6SE +/- 8351553.42, N = 62210085000021421483333189756333331327130000012963333333109877333337735466667

OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: SHA-512RTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060600M1200M1800M2400M3000MSE +/- 2677519.58, N = 6SE +/- 2958462.44, N = 6SE +/- 3109197.82, N = 6SE +/- 313758.58, N = 6SE +/- 2532115.94, N = 6SE +/- 832532.95, N = 6SE +/- 1211725.12, N = 6282036666727286500002426033333170406666716649833331412133333994983333

vkpeak

Vkpeak is a Vulkan compute benchmark inspired by OpenCL's clpeak. Vkpeak provides Vulkan compute performance measurements for FP16 / FP32 / FP64 / INT16 / INT32 scalar and vec4 performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGIOPS, More Is Bettervkpeak 20210424int16-vec4RTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30604K8K12K16K20KSE +/- 1.27, N = 3SE +/- 2.65, N = 3SE +/- 15.08, N = 3SE +/- 0.41, N = 3SE +/- 24.75, N = 3SE +/- 13.25, N = 3SE +/- 21.63, N = 316879.0616331.7614594.4510244.7510095.638487.355955.72

clpeak

Clpeak is designed to test the peak capabilities of OpenCL devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGIOPS, More Is BetterclpeakOpenCL Test: Integer Compute INTRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30604K8K12K16K20KSE +/- 120.91, N = 13SE +/- 79.28, N = 12SE +/- 119.12, N = 15SE +/- 42.92, N = 13SE +/- 41.26, N = 13SE +/- 65.78, N = 15SE +/- 6.64, N = 1317830.0017218.7115397.2710945.9010206.348397.046347.141. (CXX) g++ options: -O3 -rdynamic -lOpenCL

Hashcat

Hashcat is an open-source, advanced password recovery tool supporting GPU acceleration with OpenCL, NVIDIA CUDA, and Radeon ROCm. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: MD5RTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 306015000M30000M45000M60000M75000MSE +/- 35085900.94, N = 6SE +/- 63727349.79, N = 6SE +/- 47721835.02, N = 7SE +/- 6723155.24, N = 7SE +/- 34107189.98, N = 7SE +/- 35135391.19, N = 7SE +/- 23113714.06, N = 669887233333676895333335999948571442673371429413691571433522180000024885966667

OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: TrueCrypt RIPEMD160 + XTSRTX 3080 TiRTX 3090RTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060200K400K600K800K1000KSE +/- 757.11, N = 8SE +/- 2665.37, N = 8SE +/- 1903.85, N = 9SE +/- 1390.42, N = 9SE +/- 752.08, N = 8SE +/- 1142.14, N = 9SE +/- 2289.57, N = 8800050798175712078504878493775417944287638

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: FFT SPRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30605001000150020002500SE +/- 0.07, N = 12SE +/- 0.14, N = 14SE +/- 0.07, N = 13SE +/- 0.04, N = 12SE +/- 0.29, N = 12SE +/- 0.25, N = 12SE +/- 0.18, N = 122103.732016.861763.581292.091136.26962.42758.401. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

clpeak

Clpeak is designed to test the peak capabilities of OpenCL devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS, More Is BetterclpeakOpenCL Test: Single-Precision FloatRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30607K14K21K28K35KSE +/- 35.71, N = 14SE +/- 46.12, N = 15SE +/- 43.31, N = 13SE +/- 1.58, N = 14SE +/- 12.68, N = 14SE +/- 35.84, N = 14SE +/- 11.99, N = 1434509.5333795.9129616.9221518.7420056.4916043.1212563.041. (CXX) g++ options: -O3 -rdynamic -lOpenCL

Hashcat

Hashcat is an open-source, advanced password recovery tool supporting GPU acceleration with OpenCL, NVIDIA CUDA, and Radeon ROCm. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: 7-ZipRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060200K400K600K800K1000KSE +/- 3011.22, N = 9SE +/- 955.83, N = 9SE +/- 612.10, N = 9SE +/- 348.45, N = 9SE +/- 1441.17, N = 9SE +/- 419.32, N = 10SE +/- 371.72, N = 911091781086667976022712544669744578250413789

LuxCoreRender

LuxCoreRender is an open-source 3D physically based renderer formerly known as LuxRender. LuxCoreRender supports CPU-based rendering as well as GPU acceleration via OpenCL, NVIDIA CUDA, and NVIDIA OptiX interfaces. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgM samples/sec, More Is BetterLuxCoreRender 2.6Scene: LuxCore Benchmark - Acceleration: GPURTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30603691215SE +/- 0.02, N = 3SE +/- 0.01, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 3SE +/- 0.02, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 310.8510.519.146.616.225.394.05MIN: 3.57 / MAX: 12.86MIN: 2.79 / MAX: 12.46MIN: 2.47 / MAX: 10.83MIN: 2.19 / MAX: 7.8MIN: 2.17 / MAX: 7.35MIN: 1.78 / MAX: 6.36MIN: 1.14 / MAX: 4.78

Darktable

Darktable is an open-source photography / workflow application this will use any system-installed Darktable program or on Windows will automatically download the pre-built binary from the project. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterDarktable 3.8.1Test: Server Room - Acceleration: OpenCLRTX 3090RTX 3070 TiRTX 3070RTX 3080RTX 3060 TiRTX 3060RTX 3080 Ti0.42080.84161.26241.68322.104SE +/- 0.002, N = 12SE +/- 0.011, N = 10SE +/- 0.009, N = 10SE +/- 0.007, N = 10SE +/- 0.006, N = 10SE +/- 0.008, N = 10SE +/- 0.006, N = 100.7071.8441.8471.8491.8571.8651.870

LuxCoreRender

LuxCoreRender is an open-source 3D physically based renderer formerly known as LuxRender. LuxCoreRender supports CPU-based rendering as well as GPU acceleration via OpenCL, NVIDIA CUDA, and NVIDIA OptiX interfaces. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgM samples/sec, More Is BetterLuxCoreRender 2.6Scene: DLSC - Acceleration: GPURTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30603691215SE +/- 0.02, N = 3SE +/- 0.03, N = 3SE +/- 0.02, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 311.2810.989.527.006.615.734.29MIN: 10.49 / MAX: 11.58MIN: 10.11 / MAX: 11.41MIN: 8.4 / MAX: 9.74MIN: 6.87 / MAX: 7.16MIN: 6.05 / MAX: 6.91MIN: 5.31 / MAX: 5.98MIN: 4.11 / MAX: 4.53

OpenBenchmarking.orgM samples/sec, More Is BetterLuxCoreRender 2.6Scene: Danish Mood - Acceleration: GPURTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060246810SE +/- 0.06, N = 3SE +/- 0.06, N = 3SE +/- 0.05, N = 12SE +/- 0.01, N = 3SE +/- 0.06, N = 4SE +/- 0.05, N = 3SE +/- 0.02, N = 158.548.457.325.425.114.353.26MIN: 2.14 / MAX: 10.54MIN: 2.17 / MAX: 10.28MIN: 1.64 / MAX: 9.08MIN: 1.62 / MAX: 6.55MIN: 1.22 / MAX: 6.27MIN: 1.49 / MAX: 5.3MIN: 0.87 / MAX: 3.88

FinanceBench

FinanceBench is a collection of financial program benchmarks with support for benchmarking on the GPU via OpenCL and CPU benchmarking with OpenMP. The FinanceBench test cases are focused on Black-Sholes-Merton Process with Analytic European Option engine, QMC (Sobol) Monte-Carlo method (Equity Option Example), Bonds Fixed-rate bond with flat forward curve, and Repo Securities repurchase agreement. FinanceBench was originally written by the Cavazos Lab at University of Delaware. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterFinanceBench 2016-07-25Benchmark: Black-Scholes OpenCLRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 306048121620SE +/- 0.066100, N = 15SE +/- 0.074354, N = 15SE +/- 0.054610, N = 15SE +/- 0.004661, N = 14SE +/- 0.005828, N = 14SE +/- 0.000768, N = 15SE +/- 0.030644, N = 156.3870006.6810007.3520009.73000010.37900012.85000016.2444661. (CXX) g++ options: -O3 -march=native -fopenmp

Rodinia

Rodinia is a suite focused upon accelerating compute-intensive applications with accelerators. CUDA, OpenMP, and OpenCL parallel models are supported by the included applications. This profile utilizes select OpenCL, NVIDIA CUDA and OpenMP test binaries at the moment. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterRodinia 3.1Test: OpenCL Particle FilterRTX 3080 TiRTX 3090RTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30603691215SE +/- 0.009, N = 8SE +/- 0.023, N = 8SE +/- 0.003, N = 8SE +/- 0.002, N = 7SE +/- 0.008, N = 7SE +/- 0.008, N = 6SE +/- 0.028, N = 53.7573.9574.1895.7715.9566.9279.5211. (CXX) g++ options: -O2 -lOpenCL

clpeak

Clpeak is designed to test the peak capabilities of OpenCL devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGBPS, More Is BetterclpeakOpenCL Test: Global Memory BandwidthRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30602004006008001000SE +/- 0.22, N = 15SE +/- 0.04, N = 13SE +/- 0.03, N = 12SE +/- 0.03, N = 13SE +/- 0.13, N = 15SE +/- 0.02, N = 13SE +/- 0.00, N = 11795.98794.66662.25529.26389.39389.15315.431. (CXX) g++ options: -O3 -rdynamic -lOpenCL

cl-mem

A basic OpenCL memory benchmark. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is Bettercl-mem 2017-01-13Benchmark: ReadRTX 3080 TiRTX 3090RTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30602004006008001000SE +/- 0.13, N = 10SE +/- 0.29, N = 9SE +/- 0.08, N = 9SE +/- 0.02, N = 9SE +/- 0.13, N = 8SE +/- 0.11, N = 8SE +/- 0.09, N = 7794.4793.7673.3539.5392.9392.9318.01. (CC) gcc options: -O2 -flto -lOpenCL

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: S3DRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 306090180270360450SE +/- 0.12, N = 13SE +/- 0.09, N = 14SE +/- 0.15, N = 14SE +/- 0.05, N = 14SE +/- 0.07, N = 14SE +/- 0.04, N = 13SE +/- 0.05, N = 13429.43419.87338.69285.79219.11212.47171.951. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

Chaos Group V-RAY

This is a test of Chaos Group's V-RAY benchmark. V-RAY is a commercial renderer that can integrate with various creator software products like SketchUp and 3ds Max. The V-RAY benchmark is standalone and supports CPU and NVIDIA CUDA/RTX based rendering. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgvrays, More Is BetterChaos Group V-RAY 5.02Mode: NVIDIA RTX GPURTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30606001200180024003000SE +/- 4.93, N = 3SE +/- 6.17, N = 3SE +/- 4.91, N = 3SE +/- 1.33, N = 3SE +/- 0.00, N = 3SE +/- 0.67, N = 3SE +/- 1.67, N = 32869284024361990178716321163

cl-mem

A basic OpenCL memory benchmark. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is Bettercl-mem 2017-01-13Benchmark: WriteRTX 3080 TiRTX 3090RTX 3080RTX 3070 TiRTX 3060 TiRTX 3070RTX 3060160320480640800SE +/- 0.10, N = 10SE +/- 0.77, N = 9SE +/- 0.09, N = 9SE +/- 0.04, N = 9SE +/- 0.05, N = 8SE +/- 0.06, N = 8SE +/- 0.04, N = 7738.8730.2647.6518.0385.9382.6299.91. (CC) gcc options: -O2 -flto -lOpenCL

VkResample

VkResample is a Vulkan-based image upscaling library based on VkFFT. The sample input file is upscaling a 4K image to 8K using Vulkan-based GPU acceleration. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterVkResample 1.0Upscale: 2x - Precision: SingleRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060612182430SE +/- 0.009, N = 5SE +/- 0.003, N = 5SE +/- 0.003, N = 5SE +/- 0.002, N = 5SE +/- 0.007, N = 5SE +/- 0.006, N = 5SE +/- 0.006, N = 59.4229.57011.27213.53717.55817.68923.1841. (CXX) g++ options: -O3

LeelaChessZero

LeelaChessZero (lc0 / lczero) is a chess engine automated vian neural networks. This test profile can be used for OpenCL, CUDA + cuDNN, and BLAS (CPU-based) benchmarking. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgNodes Per Second, More Is BetterLeelaChessZero 0.28Backend: OpenCLRTX 3080 TiRTX 3090RTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 306013K26K39K52K65KSE +/- 231.66, N = 3SE +/- 514.58, N = 3SE +/- 199.06, N = 3SE +/- 164.81, N = 3SE +/- 66.16, N = 3SE +/- 136.70, N = 3SE +/- 205.27, N = 3617266115156188477534034237359251491. (CXX) g++ options: -flto -pthread

Chaos Group V-RAY

This is a test of Chaos Group's V-RAY benchmark. V-RAY is a commercial renderer that can integrate with various creator software products like SketchUp and 3ds Max. The V-RAY benchmark is standalone and supports CPU and NVIDIA CUDA/RTX based rendering. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgvpaths, More Is BetterChaos Group V-RAY 5.02Mode: NVIDIA CUDA GPURTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060400800120016002000SE +/- 2.40, N = 3SE +/- 1.53, N = 3SE +/- 3.51, N = 3SE +/- 0.33, N = 3SE +/- 0.88, N = 3SE +/- 0.67, N = 3SE +/- 0.00, N = 3208420641788150514161263855

IndigoBench

This is a test of Indigo Renderer's IndigoBench benchmark. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgM samples/s, More Is BetterIndigoBench 4.4Acceleration: OpenCL GPU - Scene: BedroomRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060510152025SE +/- 0.048, N = 3SE +/- 0.028, N = 3SE +/- 0.022, N = 3SE +/- 0.014, N = 3SE +/- 0.019, N = 3SE +/- 0.009, N = 3SE +/- 0.006, N = 320.68320.28817.71813.79812.97011.5048.507

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: YesRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30601530456075SE +/- 0.04, N = 3SE +/- 0.07, N = 3SE +/- 0.07, N = 3SE +/- 0.06, N = 3SE +/- 0.07, N = 3SE +/- 0.06, N = 3SE +/- 0.08, N = 329.1929.8232.8541.3347.3051.5668.69

MandelGPU

MandelGPU is an OpenCL benchmark and this test runs with the OpenCL rendering float4 kernel with a maximum of 4096 iterations. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSamples/sec, More Is BetterMandelGPU 1.3pts1OpenCL Device: GPURTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060110M220M330M440M550MSE +/- 1780706.00, N = 10SE +/- 1175920.80, N = 10SE +/- 572727.71, N = 10SE +/- 530024.12, N = 9SE +/- 422788.81, N = 9SE +/- 1279497.74, N = 8SE +/- 242876.80, N = 7501044446.7461789875.5428359837.3329682482.6321865318.9285067246.1215226018.81. (CC) gcc options: -O3 -lm -ftree-vectorize -funroll-loops -lglut -lOpenCL -lGL

OctaneBench

OctaneBench is a test of the OctaneRender on the GPU and requires the use of NVIDIA CUDA. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgScore, More Is BetterOctaneBench 2020.1Total ScoreRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060150300450600750682.30677.71574.87461.86417.91387.17294.86

Darktable

Darktable is an open-source photography / workflow application this will use any system-installed Darktable program or on Windows will automatically download the pre-built binary from the project. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterDarktable 3.8.1Test: Boat - Acceleration: OpenCLRTX 3090RTX 3060 TiRTX 3060RTX 3080RTX 3070RTX 3070 TiRTX 3080 Ti0.58031.16061.74092.32122.9015SE +/- 0.006, N = 10SE +/- 0.010, N = 9SE +/- 0.008, N = 9SE +/- 0.011, N = 9SE +/- 0.016, N = 9SE +/- 0.013, N = 9SE +/- 0.009, N = 91.1182.5712.5732.5772.5782.5792.579

LuxCoreRender

LuxCoreRender is an open-source 3D physically based renderer formerly known as LuxRender. LuxCoreRender supports CPU-based rendering as well as GPU acceleration via OpenCL, NVIDIA CUDA, and NVIDIA OptiX interfaces. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgM samples/sec, More Is BetterLuxCoreRender 2.6Scene: Rainbow Colors and Prism - Acceleration: GPURTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060714212835SE +/- 0.32, N = 6SE +/- 0.47, N = 15SE +/- 0.13, N = 6SE +/- 0.18, N = 5SE +/- 0.12, N = 5SE +/- 0.11, N = 5SE +/- 0.06, N = 431.0130.4826.5421.0820.3318.0713.50MIN: 28.41 / MAX: 33.92MIN: 4.31 / MAX: 33.71MIN: 23.12 / MAX: 29.17MIN: 19.53 / MAX: 22.8MIN: 18.07 / MAX: 21.88MIN: 15.25 / MAX: 19.46MIN: 12.02 / MAX: 14.47

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dAXPYRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3060 TiRTX 3070RTX 3060150300450600750SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.00, N = 3SE +/- 0.33, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 37177076325294003973271. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

LuxCoreRender

LuxCoreRender is an open-source 3D physically based renderer formerly known as LuxRender. LuxCoreRender supports CPU-based rendering as well as GPU acceleration via OpenCL, NVIDIA CUDA, and NVIDIA OptiX interfaces. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgM samples/sec, More Is BetterLuxCoreRender 2.6Scene: Orange Juice - Acceleration: GPURTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30603691215SE +/- 0.01, N = 3SE +/- 0.03, N = 3SE +/- 0.09, N = 5SE +/- 0.02, N = 3SE +/- 0.04, N = 3SE +/- 0.05, N = 3SE +/- 0.03, N = 310.3410.239.037.216.986.144.87MIN: 8.41 / MAX: 13.65MIN: 8.3 / MAX: 13.23MIN: 7.24 / MAX: 11.66MIN: 5.79 / MAX: 8.76MIN: 5.6 / MAX: 8.45MIN: 4.37 / MAX: 7.27MIN: 3.84 / MAX: 6.13

IndigoBench

This is a test of Indigo Renderer's IndigoBench benchmark. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgM samples/s, More Is BetterIndigoBench 4.4Acceleration: OpenCL GPU - Scene: SupercarRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30601224364860SE +/- 0.03, N = 3SE +/- 0.02, N = 3SE +/- 0.03, N = 3SE +/- 0.02, N = 3SE +/- 0.03, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 352.7151.9047.0039.1337.3934.0825.93

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dDOTRTX 3080 TiRTX 3090RTX 3080RTX 3070 TiRTX 3060 TiRTX 3070RTX 3060140280420560700SE +/- 0.67, N = 3SE +/- 0.00, N = 3SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 36546446055274034023311. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

FinanceBench

FinanceBench is a collection of financial program benchmarks with support for benchmarking on the GPU via OpenCL and CPU benchmarking with OpenMP. The FinanceBench test cases are focused on Black-Sholes-Merton Process with Analytic European Option engine, QMC (Sobol) Monte-Carlo method (Equity Option Example), Bonds Fixed-rate bond with flat forward curve, and Repo Securities repurchase agreement. FinanceBench was originally written by the Cavazos Lab at University of Delaware. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterFinanceBench 2016-07-25Benchmark: Monte-Carlo OpenCLRTX 3090RTX 3080 TiRTX 3080RTX 3060 TiRTX 3070RTX 3070 TiRTX 3060130260390520650SE +/- 0.69, N = 7SE +/- 0.41, N = 7SE +/- 0.29, N = 7SE +/- 0.34, N = 7SE +/- 0.28, N = 7SE +/- 0.04, N = 7SE +/- 0.29, N = 7325.27326.09388.82465.28465.30474.09622.231. (CXX) g++ options: -O3 -march=native -fopenmp

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dCOPYRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3060 TiRTX 3070RTX 3060130260390520650SE +/- 0.67, N = 3SE +/- 0.67, N = 3SE +/- 0.33, N = 3SE +/- 0.00, N = 3SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.00, N = 36025915584823723683161. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

FAHBench

FAHBench is a Folding@Home benchmark on the GPU. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgNs Per Day, More Is BetterFAHBench 2.3.2RTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 306080160240320400SE +/- 0.37, N = 3SE +/- 0.03, N = 3SE +/- 0.51, N = 3SE +/- 0.15, N = 3SE +/- 0.08, N = 3SE +/- 0.04, N = 3SE +/- 0.05, N = 3353.45345.60333.59275.12274.57245.80186.72

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: NoRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30603691215SE +/- 0.014, N = 7SE +/- 0.012, N = 7SE +/- 0.012, N = 7SE +/- 0.008, N = 6SE +/- 0.006, N = 6SE +/- 0.008, N = 6SE +/- 0.004, N = 55.5875.6416.0177.0577.7928.32810.451

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: Texture Read BandwidthRTX 3090RTX 3080RTX 3080 TiRTX 3070RTX 3060 TiRTX 3070 TiRTX 30605001000150020002500SE +/- 6.59, N = 3SE +/- 7.52, N = 3SE +/- 4.90, N = 3SE +/- 5.40, N = 4SE +/- 4.32, N = 4SE +/- 4.43, N = 4SE +/- 1.37, N = 42246.052227.532182.892150.102137.682102.791269.691. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - sAXPYRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060110220330440550SE +/- 0.33, N = 3SE +/- 0.67, N = 3SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 34984924774373613593071. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

Waifu2x-NCNN Vulkan

Waifu2x-NCNN is an NCNN neural network implementation of the Waifu2x converter project and accelerated using the Vulkan API. 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 with Vulkan. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterWaifu2x-NCNN Vulkan 20200818Scale: 2x - Denoise: 3 - TAA: YesRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 30601.10232.20463.30694.40925.5115SE +/- 0.004, N = 9SE +/- 0.005, N = 9SE +/- 0.007, N = 9SE +/- 0.005, N = 9SE +/- 0.004, N = 8SE +/- 0.003, N = 8SE +/- 0.010, N = 73.2663.2683.3373.7184.0264.1904.899

Darktable

Darktable is an open-source photography / workflow application this will use any system-installed Darktable program or on Windows will automatically download the pre-built binary from the project. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterDarktable 3.8.1Test: Masskrug - Acceleration: OpenCLRTX 3090RTX 3060 TiRTX 3070 TiRTX 3070RTX 3080RTX 3080 TiRTX 30600.59091.18181.77272.36362.9545SE +/- 0.005, N = 10SE +/- 0.005, N = 9SE +/- 0.008, N = 9SE +/- 0.006, N = 9SE +/- 0.007, N = 9SE +/- 0.010, N = 9SE +/- 0.008, N = 91.9212.6042.6052.6062.6122.6182.626

cl-mem

A basic OpenCL memory benchmark. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is Bettercl-mem 2017-01-13Benchmark: CopyRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3060 TiRTX 3070RTX 306080160240320400SE +/- 0.24, N = 9SE +/- 0.02, N = 10SE +/- 0.03, N = 9SE +/- 0.11, N = 9SE +/- 0.11, N = 8SE +/- 0.16, N = 8SE +/- 0.05, N = 7357.8356.2354.6338.6297.3296.4270.31. (CC) gcc options: -O2 -flto -lOpenCL

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - sCOPYRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3060 TiRTX 3070RTX 306080160240320400SE +/- 0.58, N = 3SE +/- 0.00, N = 3SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.67, N = 3SE +/- 0.33, N = 3SE +/- 0.00, N = 33623593593392972872761. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

JuliaGPU

JuliaGPU is an OpenCL benchmark with this version containing various PTS-specific enhancements. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSamples/sec, More Is BetterJuliaGPU 1.2pts1OpenCL Device: GPURTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3070RTX 3060 TiRTX 3060110M220M330M440M550MSE +/- 2811742.88, N = 8SE +/- 2408858.67, N = 9SE +/- 2361749.05, N = 8SE +/- 3560374.60, N = 8SE +/- 3561096.98, N = 8SE +/- 1945793.57, N = 8SE +/- 642390.88, N = 8492011904.2472565820.3459314633.0427836038.7426397082.5411524637.1378413490.21. (CC) gcc options: -O3 -march=native -ftree-vectorize -funroll-loops -lglut -lOpenCL -lGL -lm

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - sDOTRTX 3080RTX 3090RTX 3070 TiRTX 3080 TiRTX 3070RTX 3060 TiRTX 306080160240320400SE +/- 0.88, N = 3SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.58, N = 3SE +/- 0.00, N = 33753743733673313302941. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: ReductionRTX 3080RTX 3090RTX 3080 TiRTX 3070 TiRTX 3070RTX 3060 TiRTX 306080160240320400SE +/- 0.08, N = 13SE +/- 0.07, N = 13SE +/- 0.17, N = 13SE +/- 0.13, N = 13SE +/- 0.31, N = 12SE +/- 0.51, N = 12SE +/- 0.09, N = 12390.63389.58383.71381.56327.41325.50307.571. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dGEMV-TRTX 3080RTX 3090RTX 3070 TiRTX 3080 TiRTX 3070RTX 3060 TiRTX 306080160240320400SE +/- 0.67, N = 3SE +/- 0.58, N = 3SE +/- 0.58, N = 3SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.67, N = 3SE +/- 0.33, N = 33793743723693353353041. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

OpenBenchmarking.orgGFLOPs/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dGEMM-TTRTX 3090RTX 3080 TiRTX 3080130260390520650SE +/- 1.00, N = 3SE +/- 0.00, N = 25935655061. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

clpeak

Clpeak is designed to test the peak capabilities of OpenCL devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgus, Fewer Is BetterclpeakOpenCL Test: Kernel LatencyRTX 3070 TiRTX 3070RTX 3060RTX 3080RTX 3060 TiRTX 3080 TiRTX 30900.92031.84062.76093.68124.6015SE +/- 0.00, N = 15SE +/- 0.01, N = 15SE +/- 0.00, N = 15SE +/- 0.01, N = 15SE +/- 0.01, N = 15SE +/- 0.00, N = 15SE +/- 0.01, N = 153.723.753.773.783.813.864.091. (CXX) g++ options: -O3 -rdynamic -lOpenCL

ViennaCL

ViennaCL is an open-source linear algebra library written in C++ and with support for OpenCL and OpenMP. This test profile makes use of ViennaCL's built-in benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterViennaCL 1.7.1Test: OpenCL BLAS - dGEMV-NRTX 3080RTX 3070 TiRTX 3090RTX 3080 TiRTX 3060 TiRTX 3060RTX 307050100150200250SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.33, N = 3SE +/- 0.00, N = 3SE +/- 0.33, N = 32432382352322242232221. (CXX) g++ options: -fopenmp -O3 -rdynamic -lOpenCL

Rodinia

Rodinia is a suite focused upon accelerating compute-intensive applications with accelerators. CUDA, OpenMP, and OpenCL parallel models are supported by the included applications. This profile utilizes select OpenCL, NVIDIA CUDA and OpenMP test binaries at the moment. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterRodinia 3.1Test: OpenCL MyocyteRTX 3060 TiRTX 3080RTX 3070RTX 3070 TiRTX 3080 TiRTX 3060RTX 3090714212835SE +/- 0.03, N = 3SE +/- 0.04, N = 3SE +/- 0.12, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.06, N = 3SE +/- 0.10, N = 327.6427.9828.1428.4028.5628.5728.941. (CXX) g++ options: -O2 -lOpenCL

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: TriadRTX 3090RTX 3080 TiRTX 3080RTX 3070 TiRTX 3060 TiRTX 3070RTX 3060612182430SE +/- 0.03, N = 13SE +/- 0.00, N = 13SE +/- 0.00, N = 13SE +/- 0.01, N = 15SE +/- 0.01, N = 13SE +/- 0.01, N = 13SE +/- 0.01, N = 1324.4424.3924.2524.0423.6923.6623.351. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

clpeak

Clpeak is designed to test the peak capabilities of OpenCL devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGBPS, More Is BetterclpeakOpenCL Test: Transfer Bandwidth enqueueWriteBufferRTX 3070RTX 3080 TiRTX 3080RTX 3060 TiRTX 3070 TiRTX 3060RTX 3090510152025SE +/- 0.07, N = 8SE +/- 0.08, N = 5SE +/- 0.10, N = 5SE +/- 0.06, N = 8SE +/- 0.06, N = 8SE +/- 0.06, N = 5SE +/- 0.08, N = 821.9221.8921.8821.8021.7721.3721.361. (CXX) g++ options: -O3 -rdynamic -lOpenCL

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: Bus Speed DownloadRTX 3060RTX 3090RTX 3070RTX 3070 TiRTX 3080 TiRTX 3080RTX 3060 Ti612182430SE +/- 0.01, N = 13SE +/- 0.00, N = 13SE +/- 0.03, N = 13SE +/- 0.02, N = 13SE +/- 0.02, N = 13SE +/- 0.01, N = 13SE +/- 0.01, N = 1325.1025.0324.7224.7124.7024.7024.691. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

clpeak

Clpeak is designed to test the peak capabilities of OpenCL devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGBPS, More Is BetterclpeakOpenCL Test: Transfer Bandwidth enqueueReadBufferRTX 3090RTX 3060RTX 3070 TiRTX 3080RTX 3060 TiRTX 3080 TiRTX 307048121620SE +/- 0.03, N = 8SE +/- 0.03, N = 5SE +/- 0.08, N = 8SE +/- 0.05, N = 5SE +/- 0.07, N = 8SE +/- 0.05, N = 5SE +/- 0.05, N = 818.2618.2618.1818.1718.1718.1318.101. (CXX) g++ options: -O3 -rdynamic -lOpenCL

SmallPT GPU

SmallPT GPU is an OpenCL benchmark that's run with various PTS changes compared to upstream and multiple rendering scenes are available. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSamples/sec, More Is BetterSmallPT GPU 1.6pts1OpenCL Device: GPU - Scene: ComplexRTX 3070 TiRTX 3060 TiRTX 3080 TiRTX 3080RTX 3070RTX 3060RTX 3090400M800M1200M1600M2000MSE +/- 19.92, N = 3SE +/- 19.34, N = 3SE +/- 20.21, N = 3SE +/- 21.36, N = 3SE +/- 19.92, N = 3SE +/- 19.92, N = 3SE +/- 21.07, N = 316576703471657660261165764873816576379191657628359165761985116573388921. (CC) gcc options: -O3 -lm -ftree-vectorize -funroll-loops -lglut -lOpenCL -lGL

OpenBenchmarking.orgSamples/sec, More Is BetterSmallPT GPU 1.6pts1OpenCL Device: GPU - Scene: CornellRTX 3070 TiRTX 3060 TiRTX 3080 TiRTX 3080RTX 3070RTX 3060RTX 3090400M800M1200M1600M2000MSE +/- 24.54, N = 3SE +/- 24.25, N = 3SE +/- 24.54, N = 3SE +/- 24.54, N = 3SE +/- 24.54, N = 3SE +/- 24.25, N = 3SE +/- 24.54, N = 316576704811657660394165764887316576380551657628493165761998416573390281. (CC) gcc options: -O3 -lm -ftree-vectorize -funroll-loops -lglut -lOpenCL -lGL

OpenBenchmarking.orgSamples/sec, More Is BetterSmallPT GPU 1.6pts1OpenCL Device: GPU - Scene: Caustic3RTX 3070 TiRTX 3060 TiRTX 3080 TiRTX 3080RTX 3070RTX 3060RTX 3090400M800M1200M1600M2000MSE +/- 24.83, N = 3SE +/- 24.54, N = 3SE +/- 24.83, N = 3SE +/- 24.83, N = 3SE +/- 24.54, N = 3SE +/- 24.83, N = 3SE +/- 25.12, N = 316576706241657660536165764901716576381991657628635165762012616573391721. (CC) gcc options: -O3 -lm -ftree-vectorize -funroll-loops -lglut -lOpenCL -lGL

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: Bus Speed ReadbackRTX 3070 TiRTX 3060 TiRTX 3070RTX 3080RTX 3080 TiRTX 3090RTX 3060612182430SE +/- 0.00, N = 13SE +/- 0.00, N = 13SE +/- 0.00, N = 13SE +/- 0.00, N = 13SE +/- 0.00, N = 15SE +/- 0.00, N = 13SE +/- 0.00, N = 1326.3526.3526.3526.3526.3526.3526.351. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

Darktable

MinAvgMaxRTX 3060 Ti33.034.435.0RTX 307035.036.338.0RTX 308038.039.541.0RTX 3070 Ti38.040.543.0RTX 3080 Ti39.040.943.0RTX 306049.050.352.0OpenBenchmarking.orgCelsius, Fewer Is BetterDarktable 3.8.1GPU Temperature Monitor1530456075

MinAvgMaxRTX 307016.027.242.0RTX 306019.227.939.3RTX 3060 Ti18.728.239.9RTX 3070 Ti21.846.272.4RTX 308019.251.084.0RTX 3080 Ti23.061.5103.7OpenBenchmarking.orgWatts, Fewer Is BetterDarktable 3.8.1GPU Power Consumption Monitor20406080100

GPU Temperature Monitor

OpenBenchmarking.orgCelsiusGPU Temperature MonitorPhoronix Test Suite System MonitoringRTX 3060 TiRTX 3060RTX 3070RTX 3090RTX 3070 TiRTX 3080RTX 3080 Ti1428425670Min: 26 / Avg: 54.14 / Max: 70Min: 42 / Avg: 54.3 / Max: 65Min: 27 / Avg: 55.86 / Max: 72Min: 34 / Avg: 56.12 / Max: 68Min: 34 / Avg: 59.76 / Max: 75Min: 25 / Avg: 60.01 / Max: 73Min: 28 / Avg: 60.86 / Max: 75

GPU Power Consumption Monitor

OpenBenchmarking.orgWattsGPU Power Consumption MonitorPhoronix Test Suite System MonitoringRTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 3090RTX 3080 Ti60120180240300Min: 18.7 / Avg: 88.76 / Max: 170.06Min: 9.47 / Avg: 107.18 / Max: 199.84Min: 8.61 / Avg: 115.13 / Max: 219.94Min: 13.41 / Avg: 151.19 / Max: 289.69Min: 8.28 / Avg: 188.35 / Max: 323.12Min: 26.71 / Avg: 190.85 / Max: 348.7Min: 13.97 / Avg: 211.78 / Max: 349.8

Darktable

MinAvgMaxRTX 309040.042.044.0OpenBenchmarking.orgCelsius, Fewer Is BetterDarktableGPU Temperature Monitor1224364860

MinAvgMaxRTX 309028.662.4102.1OpenBenchmarking.orgWatts, Fewer Is BetterDarktableGPU Power Consumption Monitor20406080100

OpenBenchmarking.orgSeconds, Fewer Is BetterDarktable 3.8.1Test: Server Rack - Acceleration: OpenCLRTX 3060 TiRTX 3070 TiRTX 3090RTX 3060RTX 3070RTX 3080RTX 3080 Ti0.02880.05760.08640.11520.144SE +/- 0.003, N = 15SE +/- 0.004, N = 15SE +/- 0.001, N = 15SE +/- 0.004, N = 15SE +/- 0.004, N = 15SE +/- 0.004, N = 15SE +/- 0.004, N = 150.1220.1240.1240.1260.1260.1260.128