This fixes PR60012 by considering TBAA for all DDRs in vectorizer
dependence analysis because the vectorizer can rely on at least
covering two scalar iterations.
It _should_ be a regression from 4.4 but I can't make a testcase
that has two TBAA disambiguatable types and is vectorized with 4.4
it seems ...
Bootstrapped and tested on x86_64-unknown-linux-gnu, ok at this stage?
Thanks,
Richard.
2014-02-03 Richard Biener <rguent...@suse.de>
PR tree-optimization/60012
* tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
TBAA disambiguation to all DDRs.
* gcc.dg/vect/pr60012.c: New testcase.
Index: gcc/tree-vect-data-refs.c
===================================================================
--- gcc/tree-vect-data-refs.c (revision 207417)
+++ gcc/tree-vect-data-refs.c (working copy)
@@ -235,6 +235,18 @@ vect_analyze_data_ref_dependence (struct
|| (DR_IS_READ (dra) && DR_IS_READ (drb)))
return false;
+ /* Even if we have an anti-dependence then, as the vectorized loop covers at
+ least two scalar iterations, there is always also a true dependence.
+ As the vectorizer does not re-order loads and stores we can ignore
+ the anti-dependence if TBAA can disambiguate both DRs similar to the
+ case with known negative distance anti-dependences (positive
+ distance anti-dependences would violate TBAA constraints). */
+ if (((DR_IS_READ (dra) && DR_IS_WRITE (drb))
+ || (DR_IS_WRITE (dra) && DR_IS_READ (drb)))
+ && !alias_sets_conflict_p (get_alias_set (DR_REF (dra)),
+ get_alias_set (DR_REF (drb))))
+ return false;
+
/* Unknown data dependence. */
if (DDR_ARE_DEPENDENT (ddr) == chrec_dont_know)
{
Index: gcc/testsuite/gcc.dg/vect/pr60012.c
===================================================================
*** gcc/testsuite/gcc.dg/vect/pr60012.c (revision 0)
--- gcc/testsuite/gcc.dg/vect/pr60012.c (working copy)
***************
*** 0 ****
--- 1,21 ----
+ /* { dg-do compile } */
+ /* { dg-additional-options "--param vect-max-version-for-alias-checks=0" } */
+
+ typedef struct
+ {
+ short real;
+ short imag;
+ } complex16_t;
+
+ void
+ libvector_AccSquareNorm_ref (unsigned long long *acc,
+ const complex16_t *x, unsigned len)
+ {
+ unsigned i;
+ for (i = 0; i < len; i++)
+ acc[i] += ((unsigned long long)((int)x[i].real * x[i].real))
+ + ((unsigned long long)((int)x[i].imag * x[i].imag));
+ }
+
+ /* { dg-final { scan-tree-dump "LOOP VECTORIZED" "vect" } } */
+ /* { dg-final { cleanup-tree-dump "vect" } } */
