Oluwatamilore Adebayo <oluwatamilore.adeb...@arm.com> writes: > From: oluade01 <oluwatamilore.adeb...@arm.com> > > This adds a recognition pattern for the non-widening > absolute difference (ABD). > > gcc/ChangeLog: > > * doc/md.texi (sabd, uabd): Document them. > * internal-fn.def (ABD): Use new optab. > * optabs.def (sabd_optab, uabd_optab): New optabs, > * tree-vect-patterns.cc (vect_recog_absolute_difference): > Recognize the following idiom abs (a - b). > (vect_recog_sad_pattern): Refactor to use > vect_recog_absolute_difference. > (vect_recog_abd_pattern): Use patterns found by > vect_recog_absolute_difference to build a new ABD > internal call.
Thanks. I was going to push this with a trivial change, namely that: > +#define SAME_TYPE(A, B) (TYPE_PRECISION (A) == TYPE_PRECISION (B)) ...this is no longer needed. But it turns out that this conflicts with a patch that Andre pushed yesterday (2f482a07365d9f4a94a56edd13b7f01b8f78b5a0). Could you rebase on top of that and send the new patch? It should just be a case of using IFN_WIDEN_MINUS instead of WIDEN_MINUS_EXPR. Thanks, and sorry for the hassle. Richard
