Re: [7/7] Pool alignment information for common bases

2017-07-28 Thread Richard Biener
On Thu, Jul 27, 2017 at 2:50 PM, Richard Sandiford wrote: > Richard Biener writes: >> On Tue, Jul 4, 2017 at 2:01 PM, Richard Sandiford >> wrote: >>> Richard Biener writes: On Mon, Jul 3, 2017 at 9:49 AM, Richard Sandiford wrote: > @@ -2070,8 +2143,7 @@ vect_find_same_alignment_d

Re: [7/7] Pool alignment information for common bases

2017-07-27 Thread Richard Sandiford
Richard Biener writes: > On Tue, Jul 4, 2017 at 2:01 PM, Richard Sandiford > wrote: >> Richard Biener writes: >>> On Mon, Jul 3, 2017 at 9:49 AM, Richard Sandiford >>> wrote: @@ -2070,8 +2143,7 @@ vect_find_same_alignment_drs (struct dat if (dra == drb) return;

Re: [7/7] Pool alignment information for common bases

2017-07-04 Thread Richard Biener
On Tue, Jul 4, 2017 at 2:01 PM, Richard Sandiford wrote: > Richard Biener writes: >> On Mon, Jul 3, 2017 at 9:49 AM, Richard Sandiford >> wrote: >>> @@ -2070,8 +2143,7 @@ vect_find_same_alignment_drs (struct dat >>>if (dra == drb) >>> return; >>> >>> - if (!operand_equal_p (DR_BASE_OBJ

Re: [7/7] Pool alignment information for common bases

2017-07-04 Thread Richard Sandiford
Richard Biener writes: > On Mon, Jul 3, 2017 at 9:49 AM, Richard Sandiford > wrote: >> @@ -2070,8 +2143,7 @@ vect_find_same_alignment_drs (struct dat >>if (dra == drb) >> return; >> >> - if (!operand_equal_p (DR_BASE_OBJECT (dra), DR_BASE_OBJECT (drb), >> - OEP_ADD

Re: [7/7] Pool alignment information for common bases

2017-07-04 Thread Richard Biener
On Mon, Jul 3, 2017 at 9:49 AM, Richard Sandiford wrote: > This patch is a follow-on to the fix for PR81136. The testcase for that > PR shows that we can (correctly) calculate different base alignments > for two data_references but still tell that their misalignments wrt the > vector size are equ

[7/7] Pool alignment information for common bases

2017-07-03 Thread Richard Sandiford
This patch is a follow-on to the fix for PR81136. The testcase for that PR shows that we can (correctly) calculate different base alignments for two data_references but still tell that their misalignments wrt the vector size are equal. This is because we calculate the base alignments for each dr

Pool alignment information for common bases

2017-06-28 Thread Richard Sandiford
We know that if a vectorised loop is reached, all statements in that loop execute at least once, so it should be safe to pool the alignment information for all the statements we're vectorising. The only catch is that DR_REFs for masked loads and stores only occur if the mask value is nonzero. For