[Tutorial CUDA] Nvidia GPU: CUDA Compute Capability
文章推薦指數: 80 %
When you are compiling CUDA code for Nvidia GPUs it's important to know which is the Compute Capability of the GPU that you are going to use. HomeTutorials[TutorialCUDA]NvidiaGPU:CUDAComputeCapability [TutorialCUDA]NvidiaGPU:CUDAComputeCapability WhenyouarecompilingCUDAcodeforNvidiaGPUsit’simportanttoknowwhichistheComputeCapabilityoftheGPUthatyouaregoingtouse.Howmanytimesyougottheerror nvccfatal:Unsupportedgpuarchitecture'compute_XX' andyoudonotknowhowtocorrectlysolveit. Thesolutionisrelativelysimple,youmustaddthecorrectFLAGto“nvcc”call: -gencodearch=compute_XX,code=[sm_XX,compute_XX] where“XX”istheComputeCapabilityoftheNvidiaGPUboardthatyouaregoingtouse. Nowyouneedtoknowthecorrectvaluetoreplace“XX“,Nvidiahelpsuswiththeuseful“CUDAGPUs”webpage. Forexample,ifyourGPUisanNvidiaTitanXp,youknowthatitisa“GeForceproduct“,yousearchforitintherighttableandyoufindthatitsComputeCapabilityis 6.1,sothecorrectFLAGtouseinthecompileris -gencodearch=compute_61,code=[sm_61,compute_61] ifyouareinsteadanembeddeddeveloperandyouarealuckyownerofanNvidiaJetsonTX2,youmustsearchin“TEGRA/Jetsonproducts”andyouwillfindthattheComputeCapabilityoftheTX2is6.2,soyouneedtousethisconfiguration: -gencodearch=compute_62,code=[sm_62,compute_62] straightforward! IfyouarecompilingsoftwarethatwillrunondifferentmachinesandyoudonotknowexactlywhichGPUisinstalledonthem,youcangenerateanapplicationthatsupportsmorethanoneNvidiaGPUtechnology: Shell ARCH=-gencodearch=compute_30,code=[sm_30,compute_30]\ -gencodearch=compute_35,code=[sm_35,compute_35]\ -gencodearch=compute_50,code=[sm_50,compute_50]\ -gencodearch=compute_52,code=[sm_52,compute_52]\ -gencodearch=compute_61,code=[sm_61,compute_61] nvcc$(ARCH)[othernvccoptions] 12345678 ARCH=-gencodearch=compute_30,code=[sm_30,compute_30]\ -gencodearch=compute_35,code=[sm_35,compute_35]\ -gencodearch=compute_50,code=[sm_50,compute_50]\ -gencodearch=compute_52,code=[sm_52,compute_52]\ -gencodearch=compute_61,code=[sm_61,compute_61] nvcc$(ARCH)[othernvccoptions] FollowingalistofthecomputecapabilitiesforthemostcommonGPUs: JetsonProducts GPUComputeCapabilityJetsonAGXXavier7.2JetsonNano5.3JetsonTX26.2JetsonTX15.3TegraX15.3 TeslaWorkstationProducts GPUComputeCapabilityTeslaK803.7TeslaK403.5TeslaK203.5TeslaC20752.0TeslaC2050/C20702.0 TeslaNVIDIADataCenterProducts GPUComputeCapabilityNVIDIAA1008.0NVIDIAT47.5NVIDIAV1007.0TeslaP1006.0TeslaP406.1TeslaP46.1TeslaM605.2TeslaM405.2TeslaK803.7TeslaK403.5TeslaK203.5TeslaK103.0 QuadroDesktopProducts GPUComputeCapabilityQuadroRTX80007.5QuadroRTX60007.5QuadroRTX50007.5QuadroRTX40007.5QuadroGV1007.0QuadroGP1006.0QuadroP60006.1QuadroP50006.1QuadroP40006.1QuadroP22006.1QuadroP20006.1QuadroP10006.1QuadroP6206.1QuadroP6006.1QuadroP4006.1QuadroM600024GB5.2QuadroM60005.2QuadroK60003.5QuadroM50005.2QuadroK52003.5QuadroK50003.0QuadroM40005.2QuadroK42003.0QuadroK40003.0QuadroM20005.2QuadroK22003.0QuadroK20003.0QuadroK2000D3.0QuadroK12005.0QuadroK6205.0QuadroK6003.0QuadroK4203.0Quadro4103.0QuadroPlex70002.0 QuadroMobileProducts GPUComputeCapabilityRTX50007.5RTX40007.5RTX30007.5T20007.5T10007.5P6206.1P5206.1QuadroP52006.1QuadroP42006.1QuadroP32006.1QuadroP50006.1QuadroP40006.1QuadroP30006.1QuadroP20006.1QuadroP10006.1QuadroP6006.1QuadroP5006.1QuadroM5500M5.2QuadroM22005.2QuadroM12005.0QuadroM6205.2QuadroM5205.0QuadroK6000M3.0QuadroK5200M3.0QuadroK5100M3.0QuadroM5000M5.0QuadroK500M3.0QuadroK4200M3.0QuadroK4100M3.0QuadroM4000M5.0QuadroK3100M3.0QuadroM3000M5.0QuadroK2200M3.0QuadroK2100M3.0QuadroM2000M5.0QuadroK1100M3.0QuadroM1000M5.0QuadroK620M5.0QuadroK610M3.5QuadroM600M5.0QuadroK510M3.5QuadroM500M5.0 NVSDesktopProducts GPUComputeCapabilityNVIDIANVS8105.0NVIDIANVS5103.0NVIDIANVS3152.1NVIDIANVS3102.1 NVSMobileProducts GPUComputeCapabilityNVS5400M2.1NVS5200M2.1NVS4200M2.1 GeForceandTITANProducts GPUComputeCapabilityGeForceRTX30908.6GeForceRTX30808.6GeForceRTX30708.6NVIDIATITANRTX7.5GeforceRTX2080Ti7.5GeforceRTX20807.5GeforceRTX20707.5GeforceRTX20607.5NVIDIATITANV7.0NVIDIATITANXp6.1NVIDIATITANX6.1GeForceGTX1080Ti6.1GeForceGTX10806.1GeForceGTX1070Ti6.1GeForceGTX10706.1GeForceGTX10606.1GeForceGTX10506.1GeForceGTXTITANX5.2GeForceGTXTITANZ3.5GeForceGTXTITANBlack3.5GeForceGTXTITAN3.5GeForceGTX980Ti5.2GeForceGTX9805.2GeForceGTX9705.2GeForceGTX9605.2GeForceGTX9505.2GeForceGTX780Ti3.5GeForceGTX7803.5GeForceGTX7703.0GeForceGTX7603.0GeForceGTX750Ti5.0GeForceGTX7505.0GeForceGTX6903.0GeForceGTX6803.0GeForceGTX6703.0GeForceGTX660Ti3.0GeForceGTX6603.0GeForceGTX650TiBOOST3.0GeForceGTX650Ti3.0GeForceGTX6503.0GeForceGTX560Ti2.1GeForceGTX550Ti2.1GeForceGTX4602.1GeForceGTS4502.1GeForceGTS450*2.1GeForceGTX5902.0GeForceGTX5802.0GeForceGTX5702.0GeForceGTX4802.0GeForceGTX4702.0GeForceGTX4652.0GeForceGT7403.0GeForceGT7303.5GeForceGT730 DDR3,128bit2.1GeForceGT7203.5GeForceGT705*3.5GeForceGT640(GDDR5)3.5GeForceGT640 (GDDR3)2.1GeForceGT6302.1GeForceGT6202.1GeForceGT6102.1GeForceGT5202.1GeForceGT4402.1GeForceGT440*2.1GeForceGT4302.1GeForceGT430*2.1 GeForceNotebookProducts GPUComputeCapabilityGeforceRTX20807.5GeforceRTX20707.5GeforceRTX20607.5GeForceGTX10806.1GeForceGTX10706.1GeForceGTX10606.1GeForceGTX9805.2GeForceGTX980M5.2GeForceGTX970M5.2GeForceGTX965M5.2GeForceGTX960M5.0GeForceGTX950M5.0GeForce940M5.0GeForce930M5.0GeForce920M3.5GeForce910M5.2GeForceGTX880M3.0GeForceGTX870M3.0GeForceGTX860M3.0/5.0GeForceGTX850M5.0GeForce840M5.0GeForce830M5.0GeForce820M2.1GeForce800M2.1GeForceGTX780M3.0GeForceGTX770M3.0GeForceGTX765M3.0GeForceGTX760M3.0GeForceGTX680MX3.0GeForceGTX680M3.0GeForceGTX675MX3.0GeForceGTX675M2.1GeForceGTX670MX3.0GeForceGTX670M2.1GeForceGTX660M3.0GeForceGT755M3.0GeForceGT750M3.0GeForceGT650M3.0GeForceGT745M3.0GeForceGT645M3.0GeForceGT740M3.0GeForceGT730M3.0GeForceGT640M3.0GeForceGT640MLE3.0GeForceGT735M3.0GeForceGT635M2.1GeForceGT730M3.0GeForceGT630M2.1GeForceGT625M2.1GeForceGT720M2.1GeForceGT620M2.1GeForce710M2.1GeForce705M2.1GeForce610M2.1GeForceGTX580M2.1GeForceGTX570M2.1GeForceGTX560M2.1GeForceGT555M2.1GeForceGT550M2.1GeForceGT540M2.1GeForceGT525M2.1GeForceGT520MX2.1GeForceGT520M2.1GeForceGTX485M2.1GeForceGTX470M2.1GeForceGTX460M2.1GeForceGT445M2.1GeForceGT435M2.1GeForceGT420M2.1GeForceGT415M2.1GeForceGTX480M2.0GeForce710M2.1GeForce410M2.1 [Source]NvidiaDeveloper–Lastupdate:January19th,2021 TweetTelegramWhatsAppMorePocketShareonTumblrEmailPrint Commentsareclosed. Searchfor: Search OpenSourceProjectPayacoffee…supportMyzharBotTopPosts&Pages [TutorialOpenCV]"BallTracker"usingKalmanfilter [TutorialCUDA]NvidiaGPU:CUDAComputeCapability [Octoprint]ConfigureanUSBwebcam ThermalImagesonJetson™NanowithFLIRLepton3 [Tutorial]TeamViewer14onNvidiaJetsonTX2 MonthlyArchives MonthlyArchives SelectMonth March2021 (1) October2020 (1) July2020 (1) May2020 (1) July2019 (1) January2019 (1) July2018 (1) June2018 (1) March2018 (1) January2018 (1) October2017 (1) March2017 (3) May2016 (2) March2016 (2) February2016 (3) January2016 (1) November2015 (2) October2015 (2) August2015 (2) July2015 (1) June2015 (1) May2015 (2) April2015 (4) March2015 (6) February2015 (6) January2015 (7) December2014 (4) November2014 (1) October2014 (2) September2014 (5) August2014 (6) July2014 (3) June2014 (4) May2014 (4) April2014 (4) March2014 (4) February2014 (3) January2014 (1) December2013 (6) October2013 (3) September2013 (5) August2013 (2) FollowMyzharBotonFacebook FollowMyzharBotonFacebook FollowmeonTwitterMyTweets SendtoEmailAddress YourName YourEmailAddress Cancel Postwasnotsent-checkyouremailaddresses! Emailcheckfailed,pleasetryagain Sorry,yourblogcannotsharepostsbyemail. LoadingComments... WriteaComment... Email(Required) Name(Required) Website
延伸文章資訊
- 1CUDA FAQ | NVIDIA Developer
The compute capability of a GPU determines its general specifications and available features. For...
- 2CUDA - Wikipedia
CUDA (or Compute Unified Device Architecture) is a parallel computing platform and application pr...
- 3Compute Capability - Cornell Virtual Workshop
Understanding GPU Architecture: Compute Capability ... Are you thinking about writing CUDA progra...
- 4CUDA | FSU Research Computing Center
The GPUs currently available are NVIDIA GeForce GTX1080 Ti, which is of the Pascal micro-architec...
- 5CUDA GPUs | NVIDIA Developer
Are you looking for the compute capability for your GPU, then check the tables below. You can lea...