CUDA_Getting_Started_Windows.pdf

《CUDA_Getting_Started_Windows.pdf》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、NVIDIACUDAGETTINGSTARTEDGUIDEFORMICROSOFTWINDOWSDU-05349-001_v04

2、April2012InstallationandVerificationonWindowsDOCUMENTCHANGEHISTORYDU-05349-001_v04VersionDateAuthorsDescriptionofChange01April20,2010CW,TSRelease02August19,2010CWUpdatedforCUDAToolkit3.203March3,2011CWUpda

3、tedforCUDAToolkit4.004April6,2012CW,RCAddedinformationonVisualStudio2010buildcustomizationsNVIDIACUDAGettingStartedGuideforMicrosoftWindowsDU-05349-001_v04

4、iiTABLEOFCONTENTSIntroduction................................................................................1Syst

5、emRequirements...........................................................................2AboutThisDocument...........................................................................2InstallingCUDADevelopmentTools................................................3VerifyYo

6、uHaveaCUDA-EnabledSystem....................................................3DownloadtheCUDASoftware.................................................................4InstalltheCUDASoftware.....................................................................5VerifytheIns

7、tallation...........................................................................6CompilingtheExamples.....................................................................6CompilingCUDAPrograms.............................................................8CompilingSam

8、pleProjects.....................................................................8SampleProjects.................................................................................9BuildCustomizationsforExistingProjects..................................................10Add

9、itionalConsiderations............................................................11NVIDIACUDAGettingStartedGuideforMicrosoftWindowsDU-05349-001_v04

10、iiiINTRODUCTIONNVIDIA®CUDATMisageneralpurposeparallelcomputingarchitectureintroducedbyNVIDIA.ItincludestheCUDAInstructionS

11、etArchitecture(ISA)andtheparallelcomputeengineintheGPU.ToprogramtotheCUDAarchitecture,developerscanuseC,oneoft