https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115640
--- Comment #4 from rguenther at suse dot de <rguenther at suse dot de> --- On Tue, 25 Jun 2024, ams at gcc dot gnu.org wrote: > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115640 > > --- Comment #3 from Andrew Stubbs <ams at gcc dot gnu.org> --- > (In reply to Richard Biener from comment #2) > > If you force GCN to use fixed length vectors (how?), does it work? How's > > it behaving on aarch64 with SVE? (the CI was happy, but maybe doesn't > > enable SVE) > > I believe "--param vect-partial-vector-usage=0" will disable the use of > WHILE_ULT? The default is "2" for the standalone toolchain, and last I checked > the value is inherited from the host in the offload toolchain; the default for > x86_64 was "1", meaning approximately "only use partial vectors in epilogue > loops". For unknown reasons x86-64 refuses to use partial vectors.
