NVIDIA Windows vs. Linux GPU Compute

Benchmarks by Michael Larabel for a future article.

HTML result view exported from: https://openbenchmarking.org/result/2301024-NE-NVCOMPUTE35&grr.

ProcessorMotherboardMemoryDiskGraphicsAudioMonitorNetworkChipsetOSKernelDisplay DriverCompilerFile-SystemScreen ResolutionDesktopDisplay ServerOpenGLOpenCLVulkanRTX 3080RTX 3090RTX 3080RTX 3090 Windows 11 Windows 11 Ubuntu 22.10 Ubuntu 22.10Intel Core i9-13900K (24 Cores / 32 Threads)ASUS PRIME Z790-P WIFI (0602 BIOS)2 x 16384 MB 6000MHz Corsair CMK32GX5M2B6000C361908GB SOLIDIGM SSDPFKKW020X7NVIDIA GeForce RTX 3080 10GBHD Audio Device + NVIDIA HD AudioASUS V8UBluetooth Device (Personal Area )Microsoft Windows 11 Pro Build 2262110.0.22621.963 (x86_64)527.56 (31.0.15.2756)GCC 8.3.0NTFS3840x2160NVIDIA GeForce RTX 3090 24GBIntel Device 7a2732GB2048GB SOLIDIGM SSDPFKKW020X7 + 2000GBNVIDIA GeForce RTX 3080 10GBRealtek ALC897ASUS VP28UIntel Device 7a70Ubuntu 22.105.19.0-26-generic (x86_64)GNOME Shell 43.1X Server 1.21.1.4NVIDIA 525.60.114.6.0OpenCL 3.0 CUDA 12.0.891.3.224GCC 12.2.0ext4NVIDIA GeForce RTX 3090 24GBOpenBenchmarking.orgProcessor Details- RTX 3080: Windows 11: CPU Microcode: 000000000E010000- RTX 3090: Windows 11: CPU Microcode: 000000000E010000- RTX 3080: Ubuntu 22.10: Scaling Governor: intel_pstate performance (EPP: performance) - CPU Microcode: 0x10e - Thermald 2.5.1- RTX 3090: Ubuntu 22.10: Scaling Governor: intel_pstate performance (EPP: performance) - CPU Microcode: 0x10e - Thermald 2.5.1Security Details- RTX 3080: Windows 11: VBS: Enabled and running + HVCI: Running + MBEC: Available- RTX 3090: Windows 11: VBS: Enabled and running + HVCI: Running + MBEC: Available- RTX 3080: Ubuntu 22.10: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: 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 PBRSB-eIBRS: SW sequence + srbds: Not affected + tsx_async_abort: Not affected - RTX 3090: Ubuntu 22.10: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: 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 PBRSB-eIBRS: SW sequence + srbds: Not affected + tsx_async_abort: Not affected Kernel Details- RTX 3080: Ubuntu 22.10, RTX 3090: Ubuntu 22.10: Transparent Huge Pages: madviseGraphics Details- RTX 3080: Ubuntu 22.10: BAR1 / Visible vRAM Size: 16384 MiB - vBIOS Version: 94.02.20.00.07- RTX 3090: Ubuntu 22.10: BAR1 / Visible vRAM Size: 32768 MiB - vBIOS Version: 94.02.27.00.02OpenCL Details- RTX 3080: Ubuntu 22.10: GPU Compute Cores: 8704- RTX 3090: Ubuntu 22.10: GPU Compute Cores: 10496

fahbench: octanebench: Total Scoreblender: Barbershop - NVIDIA CUDAv-ray: NVIDIA RTX GPUv-ray: NVIDIA CUDA GPUblender: Barbershop - NVIDIA OptiXblender: Pabellon Barcelona - NVIDIA CUDAblender: Fishy Cat - NVIDIA CUDAblender: Classroom - NVIDIA CUDAblender: Fishy Cat - NVIDIA OptiXblender: BMW27 - NVIDIA OptiXblender: Pabellon Barcelona - NVIDIA OptiXblender: Classroom - NVIDIA OptiXblender: BMW27 - NVIDIA CUDAhashcat: MD5hashcat: SHA-512hashcat: SHA1hashcat: TrueCrypt RIPEMD160 + XTShashcat: 7-ZipRTX 3080RTX 3090RTX 3080RTX 3090 Windows 11 Windows 11 Ubuntu 22.10 Ubuntu 22.10241.9267553.473864104.442342178361.4560.7326.2724.5912.087.0519.3616.9512.3560704666667240920000019182000000716100976633242.6347663.42598488.222825209152.2649.6622.1320.5210.446.1216.1814.5710.48718370333332862100000227926000008391001134800314.1259572.856139100.202409177959.3659.2724.7623.9911.776.6118.2516.3111.81608938000002427200000192747333337183331018200333.7580689.7782483.822960211650.4248.6220.9620.039.955.7015.2613.9310.03716071333332844866667226903333338371671173100OpenBenchmarking.org

FAHBench

RTX 3080RTX 3090OpenBenchmarking.orgNs Per Day, More Is BetterFAHBench 2.3.2Windows 11Ubuntu 22.1070140210280350SE +/- 1.08, N = 3SE +/- 0.25, N = 3SE +/- 2.35, N = 15SE +/- 0.38, N = 3241.93314.13242.63333.76

OctaneBench

Total Score

RTX 3080RTX 3090OpenBenchmarking.orgScore, More Is BetterOctaneBench 2020.1Total ScoreWindows 11Ubuntu 22.10150300450600750553.47572.86663.43689.78

Blender

Blend File: Barbershop - Compute: NVIDIA CUDA

RTX 3080RTX 3090OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 3.4Blend File: Barbershop - Compute: NVIDIA CUDAWindows 11Ubuntu 22.1020406080100SE +/- 0.19, N = 3SE +/- 0.08, N = 3SE +/- 0.17, N = 3SE +/- 0.10, N = 3104.44100.2088.2283.82

Chaos Group V-RAY

Mode: NVIDIA RTX GPU

RTX 3080RTX 3090OpenBenchmarking.orgvrays, More Is BetterChaos Group V-RAY 5.02Mode: NVIDIA RTX GPUWindows 11Ubuntu 22.106001200180024003000SE +/- 7.86, N = 3SE +/- 5.46, N = 3SE +/- 1.53, N = 3SE +/- 2.08, N = 32342240928252960

Chaos Group V-RAY

Mode: NVIDIA CUDA GPU

RTX 3080RTX 3090OpenBenchmarking.orgvpaths, More Is BetterChaos Group V-RAY 5.02Mode: NVIDIA CUDA GPUWindows 11Ubuntu 22.105001000150020002500SE +/- 0.88, N = 3SE +/- 5.03, N = 3SE +/- 0.00, N = 3SE +/- 1.15, N = 31783177920912116

Blender

Blend File: Barbershop - Compute: NVIDIA OptiX

RTX 3080RTX 3090OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 3.4Blend File: Barbershop - Compute: NVIDIA OptiXWindows 11Ubuntu 22.101428425670SE +/- 0.09, N = 3SE +/- 0.10, N = 3SE +/- 0.10, N = 3SE +/- 0.02, N = 361.4559.3652.2650.42

Blender

Blend File: Pabellon Barcelona - Compute: NVIDIA CUDA

RTX 3080RTX 3090OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 3.4Blend File: Pabellon Barcelona - Compute: NVIDIA CUDAWindows 11Ubuntu 22.101428425670SE +/- 0.02, N = 3SE +/- 0.06, N = 3SE +/- 0.07, N = 3SE +/- 0.00, N = 360.7359.2749.6648.62

Blender

Blend File: Fishy Cat - Compute: NVIDIA CUDA

RTX 3080RTX 3090OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 3.4Blend File: Fishy Cat - Compute: NVIDIA CUDAWindows 11Ubuntu 22.10612182430SE +/- 0.24, N = 3SE +/- 0.04, N = 3SE +/- 0.09, N = 3SE +/- 0.00, N = 326.2724.7622.1320.96

Blender

Blend File: Classroom - Compute: NVIDIA CUDA

RTX 3080RTX 3090OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 3.4Blend File: Classroom - Compute: NVIDIA CUDAWindows 11Ubuntu 22.10612182430SE +/- 0.07, N = 3SE +/- 0.02, N = 3SE +/- 0.03, N = 3SE +/- 0.02, N = 324.5923.9920.5220.03

Blender

Blend File: Fishy Cat - Compute: NVIDIA OptiX

RTX 3080RTX 3090OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 3.4Blend File: Fishy Cat - Compute: NVIDIA OptiXWindows 11Ubuntu 22.103691215SE +/- 0.01, N = 3SE +/- 0.12, N = 6SE +/- 0.08, N = 9SE +/- 0.01, N = 312.0811.7710.449.95

Blender

Blend File: BMW27 - Compute: NVIDIA OptiX

RTX 3080RTX 3090OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 3.4Blend File: BMW27 - Compute: NVIDIA OptiXWindows 11Ubuntu 22.10246810SE +/- 0.04, N = 3SE +/- 0.05, N = 14SE +/- 0.06, N = 15SE +/- 0.00, N = 37.056.616.125.70

Blender

Blend File: Pabellon Barcelona - Compute: NVIDIA OptiX

RTX 3080RTX 3090OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 3.4Blend File: Pabellon Barcelona - Compute: NVIDIA OptiXWindows 11Ubuntu 22.10510152025SE +/- 0.04, N = 3SE +/- 0.01, N = 3SE +/- 0.06, N = 3SE +/- 0.00, N = 319.3618.2516.1815.26

Blender

Blend File: Classroom - Compute: NVIDIA OptiX

RTX 3080RTX 3090OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 3.4Blend File: Classroom - Compute: NVIDIA OptiXWindows 11Ubuntu 22.1048121620SE +/- 0.04, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.03, N = 316.9516.3114.5713.93

Blender

Blend File: BMW27 - Compute: NVIDIA CUDA

RTX 3080RTX 3090OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 3.4Blend File: BMW27 - Compute: NVIDIA CUDAWindows 11Ubuntu 22.103691215SE +/- 0.04, N = 3SE +/- 0.01, N = 3SE +/- 0.02, N = 3SE +/- 0.03, N = 312.3511.8110.4810.03

Hashcat

Benchmark: MD5

RTX 3080RTX 3090OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: MD5Windows 11Ubuntu 22.1015000M30000M45000M60000M75000MSE +/- 60645151.86, N = 3SE +/- 143096785.43, N = 3SE +/- 128630508.74, N = 3SE +/- 87864238.71, N = 360704666667608938000007183703333371607133333

Hashcat

Benchmark: SHA-512

RTX 3080RTX 3090OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: SHA-512Windows 11Ubuntu 22.10600M1200M1800M2400M3000MSE +/- 1115048.58, N = 3SE +/- 850490.05, N = 3SE +/- 907377.17, N = 3SE +/- 1836966.10, N = 32409200000242720000028621000002844866667

Hashcat

Benchmark: SHA1

RTX 3080RTX 3090OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: SHA1Windows 11Ubuntu 22.105000M10000M15000M20000M25000MSE +/- 25474758.75, N = 3SE +/- 24230307.01, N = 3SE +/- 9242474.42, N = 3SE +/- 16442255.86, N = 319182000000192747333332279260000022690333333

Hashcat

Benchmark: TrueCrypt RIPEMD160 + XTS

RTX 3080RTX 3090OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: TrueCrypt RIPEMD160 + XTSWindows 11Ubuntu 22.10200K400K600K800K1000KSE +/- 2891.94, N = 3SE +/- 1267.98, N = 3SE +/- 1305.12, N = 3SE +/- 5217.07, N = 3716100718333839100837167

Hashcat

Benchmark: 7-Zip

RTX 3080RTX 3090OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: 7-ZipWindows 11Ubuntu 22.10300K600K900K1200K1500KSE +/- 5976.71, N = 3SE +/- 3955.17, N = 3SE +/- 3946.31, N = 3SE +/- 1415.39, N = 3976633101820011348001173100

Phoronix Test Suite v10.8.4