I am trying to compile a cuda fortran program using PVF on win32 platform. However, I keep receiving the following message:
Warning: Olimit was exceeded on function vmain_vr_kernel; will not perform function-scope optimization.
To still perform function-scope optimization, use -OPT:Olimit=0 (no limit) or -OPT:Olimit=51794
Moreover, compiling takes unusually very long time, more than half an hour. I don’t quite understand what are the sources of these problems… I appreciate any help.
I’ll need an example to understand why NVIDIA’s back end tools are emitting this error and why the compilation time is so long. Would it possible to obtain your source? Please send a note PGI Customer Service (trs@pgroup.com) and ask them to forward it to me. We can then work one on one to determine the answer and post our findings.
Customer service sent me your code and I was able to reproduce the problem on Win32. I have opened a technical problem report (TPR#17239) and sent it to our engineers for further investigation. Note that the error does not appear when compiling on Linux so may be a short term work around until a work around or fix can be found.