On Thu, 29 Oct 2015, Jakub Jelinek wrote: > > rather provide a dummy 'usleep' under #ifdef __nvptx__. WDYT? > > Such ifdefs aren't really easily possible in OpenMP right now, the > preprocessing is done with the host compiler only, you'd need to arrange for > usleep being defined only in the PTX path and nowhere else.
Right, I should have remembered that. Providing it as a weak symbol doesn't work either. Providing it in a separate static library should work I think (if the library goes after -lc in host linking, the linker should pick up the dynamic symbol from libc). What do you think about building tests for each offload target separately, like OpenACC does? > > I've updated my patches locally in response to reviews to the last series, > > except where shared memory is involved. Should I resend? > > Are there any dependencies in your patch series against stuff still in > gomp-4_0-branch rather than already on the trunk? >From reviewing the diff I've found one dependency: code to place 'omp_data_o' in shared memory needs pass_late_lower_omp. Thanks. Alexander
