Re: V2 [PATCH] i386: Add pass_remove_partial_avx_dependency

2019-01-11 Thread Jeff Law
On 12/30/18 9:50 AM, H.J. Lu wrote: > On Wed, Nov 28, 2018 at 12:21 PM Jan Hubicka wrote: >> >>> On 11/28/18 12:48 PM, H.J. Lu wrote: On Mon, Nov 5, 2018 at 7:29 AM Jan Hubicka wrote: > >> On 11/5/18 7:21 AM, Jan Hubicka wrote: Did you mean "the nearest common domin

Re: V2 [PATCH] i386: Add pass_remove_partial_avx_dependency

2018-12-30 Thread H.J. Lu
On Wed, Nov 28, 2018 at 12:21 PM Jan Hubicka wrote: > > > On 11/28/18 12:48 PM, H.J. Lu wrote: > > > On Mon, Nov 5, 2018 at 7:29 AM Jan Hubicka wrote: > > >> > > >>> On 11/5/18 7:21 AM, Jan Hubicka wrote: > > > > > > Did you mean "the nearest common dominator"? > > > > If the ne

Re: V2 [PATCH] i386: Add pass_remove_partial_avx_dependency

2018-12-30 Thread H.J. Lu
On Wed, Nov 28, 2018 at 12:17 PM Jeff Law wrote: > > On 11/28/18 12:48 PM, H.J. Lu wrote: > > On Mon, Nov 5, 2018 at 7:29 AM Jan Hubicka wrote: > >> > >>> On 11/5/18 7:21 AM, Jan Hubicka wrote: > > > > Did you mean "the nearest common dominator"? > > If the nearest common domina

Re: V2 [PATCH] i386: Add pass_remove_partial_avx_dependency

2018-11-28 Thread Jan Hubicka
> On 11/28/18 12:48 PM, H.J. Lu wrote: > > On Mon, Nov 5, 2018 at 7:29 AM Jan Hubicka wrote: > >> > >>> On 11/5/18 7:21 AM, Jan Hubicka wrote: > > > > Did you mean "the nearest common dominator"? > > If the nearest common dominator appears in the loop while all uses are > ou

Re: V2 [PATCH] i386: Add pass_remove_partial_avx_dependency

2018-11-28 Thread Jeff Law
On 11/28/18 12:48 PM, H.J. Lu wrote: > On Mon, Nov 5, 2018 at 7:29 AM Jan Hubicka wrote: >> >>> On 11/5/18 7:21 AM, Jan Hubicka wrote: > > Did you mean "the nearest common dominator"? If the nearest common dominator appears in the loop while all uses are out of loops, this w

Re: V2 [PATCH] i386: Add pass_remove_partial_avx_dependency

2018-11-28 Thread H.J. Lu
On Mon, Nov 5, 2018 at 7:29 AM Jan Hubicka wrote: > > > On 11/5/18 7:21 AM, Jan Hubicka wrote: > > >> > > >> Did you mean "the nearest common dominator"? > > > > > > If the nearest common dominator appears in the loop while all uses are > > > out of loops, this will result in suboptimal xor placem

Re: V2 [PATCH] i386: Add pass_remove_partial_avx_dependency

2018-11-05 Thread Jan Hubicka
> On 11/5/18 7:21 AM, Jan Hubicka wrote: > >> > >> Did you mean "the nearest common dominator"? > > > > If the nearest common dominator appears in the loop while all uses are > > out of loops, this will result in suboptimal xor placement. > > In this case you want to split edges out of the loop. >

Re: V2 [PATCH] i386: Add pass_remove_partial_avx_dependency

2018-11-05 Thread Jeff Law
On 11/5/18 7:21 AM, Jan Hubicka wrote: >> >> Did you mean "the nearest common dominator"? > > If the nearest common dominator appears in the loop while all uses are > out of loops, this will result in suboptimal xor placement. > In this case you want to split edges out of the loop. > > In general

Re: V2 [PATCH] i386: Add pass_remove_partial_avx_dependency

2018-11-05 Thread Jan Hubicka
> > Did you mean "the nearest common dominator"? If the nearest common dominator appears in the loop while all uses are out of loops, this will result in suboptimal xor placement. In this case you want to split edges out of the loop. In general this is what the LCM framework will do for you if t