cublasDnrm2/cublasSnrm2 Snrm2 works but Dnrm2 does not

Hi,

please, help me to understand what I did wrong.

Here is a small CUBLAS program. It runs perfectly with DATATYPE equal to float (so with cublasSnrm2), but produce incorrect result with DATATYPE equal to double.

Please, help me to correct it, i wish to use double precision cublas library!

Sincerely

Serge Pisarev
test.txt (1 KB)

Hi,

sorry for my stupid stuff. I am using 8800 GTX, so, it seems that doubles are implemented only from 260/280 and it explains everything.

Sincerely

Serge