Compiler Warning: Double is not supported. Swimming demonstration

I get the following message:

Dual is not supported. Dumping for swimming

I added the following command to the compiler: -arch=sm_20 , but I still get this message.
im using sdk 4 with Nvida G105M.

what does this message mean?

+4
source share
2 answers

The G105M is only the computing power of 1.1 and therefore does not have double precision floating point capabilities. You will need a device with at least 1.3 computing power, if you really need double precision, and even then this is not a good idea at all, since it is usually a big killer of performance on all but some of the new, high-performance GPUs.

+6
source

From your comment that you are using sdk 4, I suppose you are trying to compile a sample SDK. The makefile SDK already contains many -arch or -gencode command-line options for nvcc, so even if you add -arch = sm_20, there are probably other options on the command line that lead to the compiler to generate the pre-sm_13 code, probably therefore he warns.

+6
source

Source: https://habr.com/ru/post/1395312/


All Articles