nvidia RTX 5080 rtx 5090 compute benchmarks

Benchmarks for a future article.

HTML result view exported from: https://openbenchmarking.org/result/2501295-PTS-NVIDIART22&sor.

nvidia RTX 5080 rtx 5090 compute benchmarksProcessorMotherboardChipsetMemoryDiskGraphicsAudioMonitorNetworkOSKernelDesktopDisplay ServerDisplay DriverOpenGLOpenCLCompilerFile-SystemScreen ResolutionRTX 5090RTX 5080NVIDIA RTX 5080Intel Core Ultra 9 285K @ 5.10GHz (24 Cores)ASUS ROG MAXIMUS Z890 HERO (1203 BIOS)Intel Device ae7f2 x 16GB DDR5-6400MT/s Micron CP16G64C38U5B.M8D11000GB Western Digital WDS100T1X0E-00AFY0 + 4001GB Western Digital WD_BLACK SN850X 4000GBASUS NVIDIA GeForce RTX 5090 32GBIntel Device 7f50ASUS VP28URealtek Device 8126 + Intel I226-V + Intel Wi-Fi 7Ubuntu 24.106.11.0-13-generic (x86_64)GNOME Shell 47.0X Server 1.21.1.13NVIDIA 570.86.104.6.0OpenCL 3.0 CUDA 12.8.51 + OpenCL 3.0GCC 14.2.0ext43840x2160ASUS NVIDIA GeForce RTX 5080 16GB6.11.0-14-generic (x86_64)OpenCL 3.0 CUDA 12.8.51GCC 14.2.0 + CUDA 12.8OpenBenchmarking.orgKernel Details- RTX 5090: nouveau.modeset=0 - Transparent Huge Pages: madvise- RTX 5080: Transparent Huge Pages: madvise- NVIDIA RTX 5080: 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,d,fortran,objc,obj-c++,m2,rust --enable-libphobos-checking=release --enable-libstdcxx-backtrace --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-link-serialization=2 --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-14-zdkDXv/gcc-14-14.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-14-zdkDXv/gcc-14-14.2.0/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: default) - CPU Microcode: 0x114 - Thermald 2.5.8Graphics Details- RTX 5090: BAR1 / Visible vRAM Size: 32768 MiB - vBIOS Version: 98.02.2e.00.03- RTX 5080: BAR1 / Visible vRAM Size: 16384 MiB - vBIOS Version: 98.03.3b.00.01- NVIDIA RTX 5080: BAR1 / Visible vRAM Size: 16384 MiB - vBIOS Version: 98.03.3b.00.01OpenCL Details- RTX 5090: GPU Compute Cores: 21760- RTX 5080: GPU Compute Cores: 10752- NVIDIA RTX 5080: GPU Compute Cores: 10752Security Details- gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + reg_file_data_sampling: Not affected + retbleed: Not affected + spec_rstack_overflow: 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 / Automatic IBRS; IBPB: conditional; RSB filling; PBRSB-eIBRS: Not affected; BHI: Not affected + srbds: Not affected + tsx_async_abort: Not affected

nvidia RTX 5080 rtx 5090 compute benchmarksrealsr-ncnn: 4x - Norealsr-ncnn: 4x - Yeswaifu2x-ncnn: 2x - 3 - Yeshashcat: MD5hashcat: SHA1hashcat: 7-Ziphashcat: SHA-512hashcat: TrueCrypt RIPEMD160 + XTSshoc: OpenCL - S3Dshoc: OpenCL - Triadshoc: OpenCL - FFT SPshoc: OpenCL - MD5 Hashshoc: OpenCL - Reductionshoc: OpenCL - GEMM SGEMM_Nshoc: OpenCL - Bus Speed Downloadshoc: OpenCL - Bus Speed Readbackshoc: OpenCL - Texture Read Bandwidthopencl-benchmark: FP64 Computeopencl-benchmark: FP32 Computeopencl-benchmark: FP16 Computeopencl-benchmark: INT64 Computeopencl-benchmark: INT32 Computeopencl-benchmark: INT16 Computeopencl-benchmark: INT8 Computeopencl-benchmark: Memory Bandwidth Coalesced Readopencl-benchmark: Memory Bandwidth Coalesced Writenamd-cuda: ATPase Simulation - 327,506 Atomsvkresample: 2x - Doublevkresample: 2x - Singlefluidx3d: FP32-FP32fluidx3d: FP32-FP16Cfluidx3d: FP32-FP16Sclpeak: Kernel Latencyclpeak: Integer Computeclpeak: Integer 24-bit Computeclpeak: Global Memory Bandwidthclpeak: Double-Precision Computeclpeak: Single-Precision Computeclpeak: Transfer Bandwidth enqueueReadBufferclpeak: Transfer Bandwidth enqueueWriteBufferncnn: 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 GPUv2-yolov3v2-yolov3 - mobilenetv2-yolov3ncnn: Vulkan GPU - yolov4-tinyncnn: Vulkan GPU - squeezenet_ssdncnn: Vulkan GPU - regnety_400mncnn: Vulkan GPU - vision_transformerncnn: Vulkan GPU - FastestDetblender: BMW27 - NVIDIA CUDAblender: BMW27 - NVIDIA OptiXblender: Junkshop - NVIDIA CUDAblender: Classroom - NVIDIA CUDAblender: Fishy Cat - NVIDIA CUDAblender: Junkshop - NVIDIA OptiXblender: Barbershop - NVIDIA CUDAblender: Classroom - NVIDIA OptiXblender: Fishy Cat - NVIDIA OptiXblender: Barbershop - NVIDIA OptiXblender: Pabellon Barcelona - NVIDIA CUDAblender: Pabellon Barcelona - NVIDIA OptiXindigobench: OpenCL GPU - Bedroomindigobench: OpenCL GPU - Supercarv-ray: NVIDIA RTX GPUv-ray: NVIDIA CUDA GPUnamd-cuda: ATPase with 327,506 Atomsnamd-cuda: STMV with 1,066,628 Atomsllama-cpp: NVIDIA CUDA - Llama-3.1-Tulu-3-8B-Q8_0 - Text Generation 128llama-cpp: NVIDIA CUDA - Llama-3.1-Tulu-3-8B-Q8_0 - Prompt Processing 512llama-cpp: NVIDIA CUDA - Llama-3.1-Tulu-3-8B-Q8_0 - Prompt Processing 1024llama-cpp: NVIDIA CUDA - Llama-3.1-Tulu-3-8B-Q8_0 - Prompt Processing 2048llama-cpp: NVIDIA CUDA - granite-3.0-3b-a800m-instruct-Q8_0 - Text Generation 128llama-cpp: NVIDIA CUDA - granite-3.0-3b-a800m-instruct-Q8_0 - Prompt Processing 512llama-cpp: NVIDIA CUDA - granite-3.0-3b-a800m-instruct-Q8_0 - Prompt Processing 1024llama-cpp: NVIDIA CUDA - granite-3.0-3b-a800m-instruct-Q8_0 - Prompt Processing 2048llama-cpp: NVIDIA CUDA - Mistral-7B-Instruct-v0.3-Q8_0 - Text Generation 128llama-cpp: NVIDIA CUDA - Mistral-7B-Instruct-v0.3-Q8_0 - Prompt Processing 512llama-cpp: NVIDIA CUDA - Mistral-7B-Instruct-v0.3-Q8_0 - Prompt Processing 1024llama-cpp: NVIDIA CUDA - Mistral-7B-Instruct-v0.3-Q8_0 - Prompt Processing 2048RTX 5090RTX 5080NVIDIA RTX 50804.6313.4842.299106848250000688525000003272300890040000027760001117.5427.83294398.39142.407837.20735937.228.786728.68952870.681.95117.847122.9144.39661.75954.01841.7951596.241687.490.05810103.5055.648952419140184995.1562151.9461843.111562.971976.9121415.5313.8318.4142.4410.214.684.539.827.442.6713.9237.0510.938.7428.5842.4439.3422.0247.962.7611.014.722.928.998.388.925.6635.146.164.5524.3317.35742.75992.71192348515.22118.6762.939943500000032522500000165700042430000001294700591.43627.84622381.869.5635858.04218948.128.786828.37982728.640.9557.85159.4684.27630.2127.2220.368909.66911.11212.34810.278517410328102764.8730128.0430086.5849.23962.3959218.8913.9518.6471.2741.524.2210.8237.5963.396.4236.5842.0117.4218.5444.7671.2744.0968.02165.97107.87437.344.1213.814.0814.47.8458.39.477.1539.4831.5510.626.29368.3467388252814.572544.02164977804.827292.846627.59105.093878.923831.533705.87101.937865.657327.356642.525.15318.642.7339981580000032633100000166450042655000001294200592.11527.54382381.9269.3868857.86119013.128.786428.55722734.060.9558.11659.4694.330.19727.25520.169910.07911.96212.31310.276517410330102754.8730173.7330059.43849.74962.4559305.8513.718.4971.0233.9669.7529.3365.774.0842.5442.4716.3117.657.5771.0244.5568.01155.22108.8732.487.294.112.9414.0514.437.7758.079.437.1939.0931.5410.3926.27268.3177372252814.436034.0115597.037799.247304.836630.67105.253849.783832.033687.77101.847864.887331.656653.97OpenBenchmarking.org

RealSR-NCNN

Scale: 4x - TAA: No

OpenBenchmarking.orgSeconds, Fewer Is BetterRealSR-NCNN 20200818Scale: 4x - TAA: NoRTX 5090NVIDIA RTX 5080RTX 50801.17472.34943.52414.69885.87354.6305.1535.221

RealSR-NCNN

Scale: 4x - TAA: Yes

OpenBenchmarking.orgSeconds, Fewer Is BetterRealSR-NCNN 20200818Scale: 4x - TAA: YesRTX 5090NVIDIA RTX 5080RTX 508051015202513.4818.6418.68

Waifu2x-NCNN Vulkan

Scale: 2x - Denoise: 3 - TAA: Yes

OpenBenchmarking.orgSeconds, Fewer Is BetterWaifu2x-NCNN Vulkan 20200818Scale: 2x - Denoise: 3 - TAA: YesRTX 5090NVIDIA RTX 5080RTX 50800.65931.31861.97792.63723.29652.2992.7332.930

Hashcat

Benchmark: MD5

OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: MD5RTX 5090NVIDIA RTX 5080RTX 508020000M40000M60000M80000M100000MSE +/- 102551750000.00, N = 21068482500009981580000099435000000

Hashcat

Benchmark: SHA1

OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: SHA1RTX 5090NVIDIA RTX 5080RTX 508015000M30000M45000M60000M75000M688525000003263310000032522500000

Hashcat

Benchmark: 7-Zip

OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: 7-ZipRTX 5090NVIDIA RTX 5080RTX 5080700K1400K2100K2800K3500K327230016645001657000

Hashcat

Benchmark: SHA-512

OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: SHA-512RTX 5090NVIDIA RTX 5080RTX 50802000M4000M6000M8000M10000M890040000042655000004243000000

Hashcat

Benchmark: TrueCrypt RIPEMD160 + XTS

OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: TrueCrypt RIPEMD160 + XTSRTX 5090RTX 5080NVIDIA RTX 5080600K1200K1800K2400K3000K277600012947001294200

SHOC Scalable HeterOgeneous Computing

Target: OpenCL - Benchmark: S3D

OpenBenchmarking.orgGFLOPS, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: S3DRTX 5090NVIDIA RTX 5080RTX 508020040060080010001117.54592.12591.441. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

SHOC Scalable HeterOgeneous Computing

Target: OpenCL - Benchmark: Triad

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: TriadRTX 5080RTX 5090NVIDIA RTX 508071421283527.8527.8327.541. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

SHOC Scalable HeterOgeneous Computing

Target: OpenCL - Benchmark: FFT SP

OpenBenchmarking.orgGFLOPS, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: FFT SPRTX 5090NVIDIA RTX 5080RTX 508090018002700360045004398.392381.922381.801. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

SHOC Scalable HeterOgeneous Computing

Target: OpenCL - Benchmark: MD5 Hash

OpenBenchmarking.orgGHash/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: MD5 HashRTX 5090RTX 5080NVIDIA RTX 5080306090120150142.4169.5669.391. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

SHOC Scalable HeterOgeneous Computing

Target: OpenCL - Benchmark: Reduction

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: ReductionRTX 5080NVIDIA RTX 5080RTX 50902004006008001000858.04857.86837.211. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

SHOC Scalable HeterOgeneous Computing

Target: OpenCL - Benchmark: GEMM SGEMM_N

OpenBenchmarking.orgGFLOPS, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: GEMM SGEMM_NRTX 5090NVIDIA RTX 5080RTX 50808K16K24K32K40K35937.219013.118948.11. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

SHOC Scalable HeterOgeneous Computing

Target: OpenCL - Benchmark: Bus Speed Download

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: Bus Speed DownloadRTX 5080RTX 5090NVIDIA RTX 508071421283528.7928.7928.791. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

SHOC Scalable HeterOgeneous Computing

Target: OpenCL - Benchmark: Bus Speed Readback

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: Bus Speed ReadbackRTX 5090NVIDIA RTX 5080RTX 508071421283528.6928.5628.381. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

SHOC Scalable HeterOgeneous Computing

Target: OpenCL - Benchmark: Texture Read Bandwidth

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: Texture Read BandwidthRTX 5090NVIDIA RTX 5080RTX 508060012001800240030002870.682734.062728.641. (CXX) g++ options: -O2 -lSHOCCommonMPI -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt -lmpi_cxx -lmpi

ProjectPhysX OpenCL-Benchmark

Operation: FP64 Compute

OpenBenchmarking.orgTFLOPs/s, More Is BetterProjectPhysX OpenCL-Benchmark 1.6Operation: FP64 ComputeRTX 5090NVIDIA RTX 5080RTX 50800.43880.87761.31641.75522.1941.950.950.951. (CXX) g++ options: -std=c++17 -pthread -lOpenCL

ProjectPhysX OpenCL-Benchmark

Operation: FP32 Compute

OpenBenchmarking.orgTFLOPs/s, More Is BetterProjectPhysX OpenCL-Benchmark 1.6Operation: FP32 ComputeRTX 5090NVIDIA RTX 5080RTX 5080306090120150117.8558.1257.851. (CXX) g++ options: -std=c++17 -pthread -lOpenCL

ProjectPhysX OpenCL-Benchmark

Operation: FP16 Compute

OpenBenchmarking.orgTFLOPs/s, More Is BetterProjectPhysX OpenCL-Benchmark 1.6Operation: FP16 ComputeRTX 5090NVIDIA RTX 5080RTX 5080306090120150122.9159.4759.471. (CXX) g++ options: -std=c++17 -pthread -lOpenCL

ProjectPhysX OpenCL-Benchmark

Operation: INT64 Compute

OpenBenchmarking.orgTIOPs/s, More Is BetterProjectPhysX OpenCL-Benchmark 1.6Operation: INT64 ComputeRTX 5090NVIDIA RTX 5080RTX 50800.98911.97822.96733.95644.94554.3964.3004.2761. (CXX) g++ options: -std=c++17 -pthread -lOpenCL

ProjectPhysX OpenCL-Benchmark

Operation: INT32 Compute

OpenBenchmarking.orgTIOPs/s, More Is BetterProjectPhysX OpenCL-Benchmark 1.6Operation: INT32 ComputeRTX 5090RTX 5080NVIDIA RTX 5080142842567061.7630.2130.201. (CXX) g++ options: -std=c++17 -pthread -lOpenCL

ProjectPhysX OpenCL-Benchmark

Operation: INT16 Compute

OpenBenchmarking.orgTIOPs/s, More Is BetterProjectPhysX OpenCL-Benchmark 1.6Operation: INT16 ComputeRTX 5090NVIDIA RTX 5080RTX 5080122436486054.0227.2627.221. (CXX) g++ options: -std=c++17 -pthread -lOpenCL

ProjectPhysX OpenCL-Benchmark

Operation: INT8 Compute

OpenBenchmarking.orgTIOPs/s, More Is BetterProjectPhysX OpenCL-Benchmark 1.6Operation: INT8 ComputeRTX 5090RTX 5080NVIDIA RTX 5080102030405041.8020.3720.171. (CXX) g++ options: -std=c++17 -pthread -lOpenCL

ProjectPhysX OpenCL-Benchmark

Operation: Memory Bandwidth Coalesced Read

OpenBenchmarking.orgGB/s, More Is BetterProjectPhysX OpenCL-Benchmark 1.6Operation: Memory Bandwidth Coalesced ReadRTX 5090NVIDIA RTX 5080RTX 5080300600900120015001596.24910.07909.661. (CXX) g++ options: -std=c++17 -pthread -lOpenCL

ProjectPhysX OpenCL-Benchmark

Operation: Memory Bandwidth Coalesced Write

OpenBenchmarking.orgGB/s, More Is BetterProjectPhysX OpenCL-Benchmark 1.6Operation: Memory Bandwidth Coalesced WriteRTX 5090NVIDIA RTX 5080RTX 50804008001200160020001687.49911.96911.111. (CXX) g++ options: -std=c++17 -pthread -lOpenCL

NAMD CUDA

ATPase Simulation - 327,506 Atoms

OpenBenchmarking.orgdays/ns, Fewer Is BetterNAMD CUDA 2.14ATPase Simulation - 327,506 AtomsRTX 50900.01310.02620.03930.05240.06550.05810

VkResample

Upscale: 2x - Precision: Double

OpenBenchmarking.orgms, Fewer Is BetterVkResample 1.0Upscale: 2x - Precision: DoubleRTX 5090NVIDIA RTX 5080RTX 508050100150200250103.51212.31212.351. (CXX) g++ options: -O3

VkResample

Upscale: 2x - Precision: Single

OpenBenchmarking.orgms, Fewer Is BetterVkResample 1.0Upscale: 2x - Precision: SingleRTX 5090NVIDIA RTX 5080RTX 508036912155.64810.27610.2781. (CXX) g++ options: -O3

FluidX3D

Test: FP32-FP32

OpenBenchmarking.orgMLUPs/s, More Is BetterFluidX3D 3.0Test: FP32-FP32RTX 5090NVIDIA RTX 5080RTX 50802K4K6K8K10K952451745174

FluidX3D

Test: FP32-FP16C

OpenBenchmarking.orgMLUPs/s, More Is BetterFluidX3D 3.0Test: FP32-FP16CRTX 5090NVIDIA RTX 5080RTX 50804K8K12K16K20K191401033010328

FluidX3D

Test: FP32-FP16S

OpenBenchmarking.orgMLUPs/s, More Is BetterFluidX3D 3.0Test: FP32-FP16SRTX 5090RTX 5080NVIDIA RTX 50804K8K12K16K20K184991027610275

clpeak

OpenCL Test: Kernel Latency

OpenBenchmarking.orgus, Fewer Is Betterclpeak 1.1.2OpenCL Test: Kernel LatencyRTX 5080NVIDIA RTX 5080RTX 50901.15882.31763.47644.63525.7944.874.875.151. (CXX) g++ options: -O3

clpeak

OpenCL Test: Integer Compute

OpenBenchmarking.orgGIOPS, More Is Betterclpeak 1.1.2OpenCL Test: Integer ComputeRTX 5090NVIDIA RTX 5080RTX 508013K26K39K52K65K62151.9430173.7330128.041. (CXX) g++ options: -O3

clpeak

OpenCL Test: Integer 24-bit Compute

OpenBenchmarking.orgGIOPS, More Is Betterclpeak 1.1.2OpenCL Test: Integer 24-bit ComputeRTX 5090RTX 5080NVIDIA RTX 508013K26K39K52K65K61843.1130086.5030059.431. (CXX) g++ options: -O3

clpeak

OpenCL Test: Global Memory Bandwidth

OpenBenchmarking.orgGBPS, More Is Betterclpeak 1.1.2OpenCL Test: Global Memory BandwidthRTX 5090NVIDIA RTX 5080RTX 5080300600900120015001562.97849.74849.231. (CXX) g++ options: -O3

clpeak

OpenCL Test: Double-Precision Compute

OpenBenchmarking.orgGFLOPS, More Is Betterclpeak 1.1.2OpenCL Test: Double-Precision ComputeRTX 5090NVIDIA RTX 5080RTX 50804008001200160020001976.90962.45962.391. (CXX) g++ options: -O3

clpeak

OpenCL Test: Single-Precision Compute

OpenBenchmarking.orgGFLOPS, More Is Betterclpeak 1.1.2OpenCL Test: Single-Precision ComputeRTX 5090NVIDIA RTX 5080RTX 508030K60K90K120K150K121415.5359305.8559218.891. (CXX) g++ options: -O3

clpeak

OpenCL Test: Transfer Bandwidth enqueueReadBuffer

OpenBenchmarking.orgGBPS, More Is Betterclpeak 1.1.2OpenCL Test: Transfer Bandwidth enqueueReadBufferRTX 5080RTX 5090NVIDIA RTX 50804812162013.9513.8313.701. (CXX) g++ options: -O3

clpeak

OpenCL Test: Transfer Bandwidth enqueueWriteBuffer

OpenBenchmarking.orgGBPS, More Is Betterclpeak 1.1.2OpenCL Test: Transfer Bandwidth enqueueWriteBufferRTX 5080NVIDIA RTX 5080RTX 509051015202518.6418.4918.411. (CXX) g++ options: -O3

NCNN

Target: Vulkan GPU - Model: mobilenet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: mobilenetRTX 5090NVIDIA RTX 5080RTX 5080163248648042.4471.0271.27MIN: 8.34 / MAX: 76.17MIN: 8.7 / MAX: 81.03MIN: 8.2 / MAX: 79.421. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU-v2-v2 - Model: mobilenet-v2

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU-v2-v2 - Model: mobilenet-v2RTX 5090NVIDIA RTX 5080RTX 508091827364510.2133.9641.52MIN: 3.85 / MAX: 64.73MIN: 3.85 / MAX: 68.9MIN: 3.78 / MAX: 68.661. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU-v3-v3 - Model: mobilenet-v3

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU-v3-v3 - Model: mobilenet-v3RTX 5080RTX 5090NVIDIA RTX 50802468104.224.686.00MIN: 4.04 / MAX: 4.96MIN: 4.08 / MAX: 57.94MIN: 4.05 / MAX: 791. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: shufflenet-v2

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: shufflenet-v2RTX 5090NVIDIA RTX 5080RTX 508036912154.539.7510.82MIN: 3.88 / MAX: 57.59MIN: 3.91 / MAX: 74.73MIN: 3.9 / MAX: 74.521. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: mnasnet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: mnasnetRTX 5090NVIDIA RTX 5080RTX 50809182736459.8029.3337.59MIN: 3.75 / MAX: 63.5MIN: 3.68 / MAX: 66.87MIN: 3.7 / MAX: 66.41. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: efficientnet-b0

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: efficientnet-b0RTX 5090RTX 5080NVIDIA RTX 5080153045607527.4463.3965.77MIN: 6.34 / MAX: 109.98MIN: 6.28 / MAX: 114.02MIN: 6.28 / MAX: 115.251. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: blazeface

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: blazefaceRTX 5090NVIDIA RTX 5080RTX 50802468102.674.086.42MIN: 2.4 / MAX: 41.21MIN: 2.36 / MAX: 52.23MIN: 2.36 / MAX: 52.251. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: googlenet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: googlenetRTX 5090RTX 5080NVIDIA RTX 5080102030405013.9236.5842.54MIN: 7.49 / MAX: 98.37MIN: 7.53 / MAX: 102.56MIN: 7.59 / MAX: 104.851. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: vgg16

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: vgg16RTX 5090RTX 5080NVIDIA RTX 5080102030405037.0542.0142.47MIN: 22.53 / MAX: 46.22MIN: 23.96 / MAX: 46.31MIN: 25.67 / MAX: 47.211. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: resnet18

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: resnet18RTX 5090NVIDIA RTX 5080RTX 50804812162010.9316.3117.42MIN: 4.48 / MAX: 44.35MIN: 4.4 / MAX: 44.98MIN: 4.48 / MAX: 44.531. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: alexnet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: alexnetRTX 5090NVIDIA RTX 5080RTX 50805101520258.7417.6018.54MIN: 3.21 / MAX: 22.23MIN: 3.2 / MAX: 23.03MIN: 3.21 / MAX: 22.491. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: resnet50

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: resnet50RTX 5090RTX 5080NVIDIA RTX 5080132639526528.5844.7657.57MIN: 10.02 / MAX: 89.48MIN: 9.99 / MAX: 92.95MIN: 9.95 / MAX: 941. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPUv2-yolov3v2-yolov3 - Model: mobilenetv2-yolov3

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPUv2-yolov3v2-yolov3 - Model: mobilenetv2-yolov3RTX 5090NVIDIA RTX 5080RTX 5080163248648042.4471.0271.27MIN: 8.34 / MAX: 76.17MIN: 8.7 / MAX: 81.03MIN: 8.2 / MAX: 79.421. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: yolov4-tiny

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: yolov4-tinyRTX 5090RTX 5080NVIDIA RTX 5080102030405039.3444.0944.55MIN: 15.92 / MAX: 49.04MIN: 14.62 / MAX: 49.42MIN: 13.75 / MAX: 49.721. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: squeezenet_ssd

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: squeezenet_ssdRTX 5090NVIDIA RTX 5080RTX 5080153045607522.0268.0168.02MIN: 7.41 / MAX: 92.66MIN: 7.27 / MAX: 102.26MIN: 7.34 / MAX: 99.351. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: regnety_400m

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: regnety_400mRTX 5090NVIDIA RTX 5080RTX 5080408012016020047.90155.22165.97MIN: 21.96 / MAX: 421.33MIN: 21.96 / MAX: 495.67MIN: 21.95 / MAX: 492.351. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: vision_transformer

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: vision_transformerRTX 5090RTX 5080NVIDIA RTX 50802040608010062.76107.87108.87MIN: 40.3 / MAX: 105.61MIN: 40.48 / MAX: 118.93MIN: 45.11 / MAX: 120.31. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

NCNN

Target: Vulkan GPU - Model: FastestDet

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20241226Target: Vulkan GPU - Model: FastestDetRTX 5090NVIDIA RTX 5080RTX 5080102030405011.0132.4843.00MIN: 5.09 / MAX: 92.36MIN: 5.11 / MAX: 96.35MIN: 5.09 / MAX: 96.191. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

Blender

Blend File: BMW27 - Compute: NVIDIA CUDA

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.3Blend File: BMW27 - Compute: NVIDIA CUDARTX 5090NVIDIA RTX 5080RTX 50802468104.727.297.34

Blender

Blend File: BMW27 - Compute: NVIDIA OptiX

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.3Blend File: BMW27 - Compute: NVIDIA OptiXRTX 5090NVIDIA RTX 5080RTX 50800.9271.8542.7813.7084.6352.924.104.12

Blender

Blend File: Junkshop - Compute: NVIDIA CUDA

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.3Blend File: Junkshop - Compute: NVIDIA CUDARTX 5090NVIDIA RTX 5080RTX 5080481216208.9912.9413.80

Blender

Blend File: Classroom - Compute: NVIDIA CUDA

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.3Blend File: Classroom - Compute: NVIDIA CUDARTX 5090NVIDIA RTX 5080RTX 5080481216208.3814.0514.08

Blender

Blend File: Fishy Cat - Compute: NVIDIA CUDA

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.3Blend File: Fishy Cat - Compute: NVIDIA CUDARTX 5090RTX 5080NVIDIA RTX 5080481216208.9214.4014.43

Blender

Blend File: Junkshop - Compute: NVIDIA OptiX

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.3Blend File: Junkshop - Compute: NVIDIA OptiXRTX 5090NVIDIA RTX 5080RTX 50802468105.667.777.84

Blender

Blend File: Barbershop - Compute: NVIDIA CUDA

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.3Blend File: Barbershop - Compute: NVIDIA CUDARTX 5090NVIDIA RTX 5080RTX 5080132639526535.1458.0758.30

Blender

Blend File: Classroom - Compute: NVIDIA OptiX

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.3Blend File: Classroom - Compute: NVIDIA OptiXRTX 5090NVIDIA RTX 5080RTX 508036912156.169.439.47

Blender

Blend File: Fishy Cat - Compute: NVIDIA OptiX

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.3Blend File: Fishy Cat - Compute: NVIDIA OptiXRTX 5090RTX 5080NVIDIA RTX 50802468104.557.157.19

Blender

Blend File: Barbershop - Compute: NVIDIA OptiX

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.3Blend File: Barbershop - Compute: NVIDIA OptiXRTX 5090NVIDIA RTX 5080RTX 508091827364524.3339.0939.48

Blender

Blend File: Pabellon Barcelona - Compute: NVIDIA CUDA

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.3Blend File: Pabellon Barcelona - Compute: NVIDIA CUDARTX 5090NVIDIA RTX 5080RTX 508071421283517.3531.5431.55

Blender

Blend File: Pabellon Barcelona - Compute: NVIDIA OptiX

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.3Blend File: Pabellon Barcelona - Compute: NVIDIA OptiXRTX 5090NVIDIA RTX 5080RTX 508036912157.0010.3910.60

IndigoBench

Acceleration: OpenCL GPU - Scene: Bedroom

OpenBenchmarking.orgM samples/s, More Is BetterIndigoBench 4.4Acceleration: OpenCL GPU - Scene: BedroomRTX 5090RTX 5080NVIDIA RTX 5080102030405042.7626.2926.27

IndigoBench

Acceleration: OpenCL GPU - Scene: Supercar

OpenBenchmarking.orgM samples/s, More Is BetterIndigoBench 4.4Acceleration: OpenCL GPU - Scene: SupercarRTX 5090RTX 5080NVIDIA RTX 50802040608010092.7068.3568.32

Chaos Group V-RAY

Mode: NVIDIA RTX GPU

OpenBenchmarking.orgvpaths, More Is BetterChaos Group V-RAY 6.0Mode: NVIDIA RTX GPURTX 5090RTX 5080NVIDIA RTX 50803K6K9K12K15K1192373887372

Chaos Group V-RAY

Mode: NVIDIA CUDA GPU

OpenBenchmarking.orgvpaths, More Is BetterChaos Group V-RAY 6.0Mode: NVIDIA CUDA GPURTX 5090NVIDIA RTX 5080RTX 508010002000300040005000485125282528

NAMD CUDA

Input: ATPase with 327,506 Atoms

OpenBenchmarking.orgns/day, More Is BetterNAMD CUDA 3.0.1Input: ATPase with 327,506 AtomsRTX 5080NVIDIA RTX 50804812162014.5714.44

NAMD CUDA

Input: STMV with 1,066,628 Atoms

OpenBenchmarking.orgns/day, More Is BetterNAMD CUDA 3.0.1Input: STMV with 1,066,628 AtomsRTX 5080NVIDIA RTX 50800.90491.80982.71473.61964.52454.021644.01155

Llama.cpp

Backend: NVIDIA CUDA - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Text Generation 128

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: NVIDIA CUDA - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Text Generation 128NVIDIA RTX 5080RTX 50802040608010097.0397.001. (CXX) g++ options: -O3

Llama.cpp

Backend: NVIDIA CUDA - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Prompt Processing 512

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: NVIDIA CUDA - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Prompt Processing 512RTX 5080NVIDIA RTX 50802K4K6K8K10K7804.827799.241. (CXX) g++ options: -O3

Llama.cpp

Backend: NVIDIA CUDA - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Prompt Processing 1024

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: NVIDIA CUDA - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Prompt Processing 1024NVIDIA RTX 5080RTX 5080160032004800640080007304.837292.841. (CXX) g++ options: -O3

Llama.cpp

Backend: NVIDIA CUDA - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Prompt Processing 2048

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: NVIDIA CUDA - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Prompt Processing 2048NVIDIA RTX 5080RTX 5080140028004200560070006630.676627.591. (CXX) g++ options: -O3

Llama.cpp

Backend: NVIDIA CUDA - Model: granite-3.0-3b-a800m-instruct-Q8_0 - Test: Text Generation 128

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: NVIDIA CUDA - Model: granite-3.0-3b-a800m-instruct-Q8_0 - Test: Text Generation 128NVIDIA RTX 5080RTX 508020406080100105.25105.091. (CXX) g++ options: -O3

Llama.cpp

Backend: NVIDIA CUDA - Model: granite-3.0-3b-a800m-instruct-Q8_0 - Test: Prompt Processing 512

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: NVIDIA CUDA - Model: granite-3.0-3b-a800m-instruct-Q8_0 - Test: Prompt Processing 512RTX 5080NVIDIA RTX 508080016002400320040003878.923849.781. (CXX) g++ options: -O3

Llama.cpp

Backend: NVIDIA CUDA - Model: granite-3.0-3b-a800m-instruct-Q8_0 - Test: Prompt Processing 1024

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: NVIDIA CUDA - Model: granite-3.0-3b-a800m-instruct-Q8_0 - Test: Prompt Processing 1024NVIDIA RTX 5080RTX 508080016002400320040003832.033831.531. (CXX) g++ options: -O3

Llama.cpp

Backend: NVIDIA CUDA - Model: granite-3.0-3b-a800m-instruct-Q8_0 - Test: Prompt Processing 2048

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: NVIDIA CUDA - Model: granite-3.0-3b-a800m-instruct-Q8_0 - Test: Prompt Processing 2048RTX 5080NVIDIA RTX 508080016002400320040003705.873687.771. (CXX) g++ options: -O3

Llama.cpp

Backend: NVIDIA CUDA - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Text Generation 128

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: NVIDIA CUDA - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Text Generation 128RTX 5080NVIDIA RTX 508020406080100101.93101.841. (CXX) g++ options: -O3

Llama.cpp

Backend: NVIDIA CUDA - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Prompt Processing 512

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: NVIDIA CUDA - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Prompt Processing 512RTX 5080NVIDIA RTX 50802K4K6K8K10K7865.657864.881. (CXX) g++ options: -O3

Llama.cpp

Backend: NVIDIA CUDA - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Prompt Processing 1024

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: NVIDIA CUDA - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Prompt Processing 1024NVIDIA RTX 5080RTX 5080160032004800640080007331.657327.351. (CXX) g++ options: -O3

Llama.cpp

Backend: NVIDIA CUDA - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Prompt Processing 2048

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b4397Backend: NVIDIA CUDA - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Prompt Processing 2048NVIDIA RTX 5080RTX 5080140028004200560070006653.976642.521. (CXX) g++ options: -O3

Phoronix Test Suite v10.8.5