1 | CUDA Version: v9.2 |
When I compiled a sample named “vectorAdd”, it occurs an error with message:
1 | C1189 #error: -- unsupported Microsoft Visual Studio version! Only the versions 2012, 2013, 2015 and 2017 are supported! vectorAdd c:\program files\nvidia gpu computing toolkit\cuda\v9.2\include\crt\host_config.h 133 |
It’s the file host_config.h
who has limited the version. So we can change the version limit in this file:
1 | 131 #if _MSC_VER < 1600 || _MSC_VER > 1913 |
to
1 | 131 #if _MSC_VER < 1600 || _MSC_VER > 1920 |
It depends on your Visual Studio version that which version should you indicate, which means that probably _MSC_VER > 19xx
is written in your host_config.h
, while your actual _MSC_VER
could also be greater than 1920. It’s all up to your situation, the principle is to modify the limit, you’ve got it!
For another error:
If we open an individual sample in the grand project, it will occur another problem.
The message shows that all the head files for CUDA have not been added. So we can add them manually: Right click on the solution->Properties->VC++ Directories->include path
, add "C:\ProgramData\NVIDIA Corporation\CUDA Samples\v9.2\common\inc;"
(which is your CUDA head files directory).
All will be well. God bless you!