Hi! On Tue, 13 Dec 2016 13:42:23 +0100, Martin Jambor <mjam...@suse.cz> wrote: > On Tue, Dec 13, 2016 at 12:43:16PM +0100, Jakub Jelinek wrote: > > On Tue, Dec 13, 2016 at 12:39:01PM +0100, Thomas Schwinge wrote: > > > On Fri, 9 Dec 2016 14:08:21 +0100, Martin Jambor <mjam...@suse.cz> wrote: > > > > this is the promised attempt at splitting omp-low.c [...]
> > > > - move all pre-lowering gridification stuff to a new file > > > > omp-grid.c. [...] > > > > > > Is that code generic enough to not call this file "omp-hsa.c" or similar? > > > > Not at the moment, but... > > > And this as well. But omp-grid.c is fine too. > > ...I prefer omp-grid.c because I plan to use gridification also for > GCN targets, though hopefully only as an optimization rather than a > hard requirement ...and in fact I still think it is a good > optimization of simple loops for execution on all CUDA-like > environments with block/thread grids because it removes conditions > which the run-time can handle better. That certainly is reason enough to go with the generic name -- thanks for the explanation! > > > > - I moved stuff that was used from all over the place to a new file > > > > omp-general.c (unless it would mean exposing omp_region or > > > > omp_context types). > > > > > > I'd have called that simply "omp.c". > > > > The problem with that is that the corresponding header can't be called > > omp.h for obvious reasons, we already have one with very different meaning. > > That is exactly the reason why I chose omp-general. OK, that wasn't obvious to me, but yeah, it's probably best to avoid the name clash with libgomp's omp.h. I still couldn't allocate time to review the patch, but at least I now have tested it -- no regressions. As I suppose you want to commit this as sooner than later ;-) and you already have approval as I understand it, how about you do commit it, and I'll then later follow up, with any changes I'd additionally like to get done. Grüße Thomas