* [PATCH] Fix PR50067
@ 2011-08-22 11:27 Richard Guenther
0 siblings, 0 replies; 2+ messages in thread
From: Richard Guenther @ 2011-08-22 11:27 UTC (permalink / raw)
To: gcc-patches
Applied.
Richard.
2011-08-22 Richard Guenther <rguenther@suse.de>
PR testsuite/50145
* gcc.dg/torture/pr50067-1.c: Run on little-endian systems only.
* gcc.dg/torture/pr50067-2.c: Likewise.
Index: gcc/testsuite/gcc.dg/torture/pr50067-1.c
===================================================================
--- gcc/testsuite/gcc.dg/torture/pr50067-1.c (revision 177949)
+++ gcc/testsuite/gcc.dg/torture/pr50067-1.c (working copy)
@@ -9,10 +9,15 @@ short a[32] = { 0, 1, 2, 3, 4, 5, 6, 7,
short b[32] = { 4, 0, 5, 0, 6, 0, 7, 0, 8, 0, };
int main()
{
+#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
int i;
- for (i = 0; i < 32; ++i)
- (*((unsigned short(*)[32])&a[0]))[i] = (*((char(*)[32])&a[0]))[i+8];
- if (memcmp (&a, &b, sizeof (a)) != 0)
- abort ();
+ if (sizeof (short) == 2)
+ {
+ for (i = 0; i < 32; ++i)
+ (*((unsigned short(*)[32])&a[0]))[i] = (*((char(*)[32])&a[0]))[i+8];
+ if (memcmp (&a, &b, sizeof (a)) != 0)
+ abort ();
+ }
+#endif
return 0;
}
Index: gcc/testsuite/gcc.dg/torture/pr50067-2.c
===================================================================
--- gcc/testsuite/gcc.dg/torture/pr50067-2.c (revision 177949)
+++ gcc/testsuite/gcc.dg/torture/pr50067-2.c (working copy)
@@ -9,12 +9,17 @@ short a[32] = { 0, 1, 2, 3, 4, 5, 6, 7,
short b[32] = { 4, 0, 5, 0, 6, 0, 7, 0, 8, 0, };
int main()
{
+#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
int i;
- for (i = 0; i < 32; ++i)
+ if (sizeof (short) == 2)
{
- a[i] = (*((char(*)[32])&a[0]))[i+8];
+ for (i = 0; i < 32; ++i)
+ {
+ a[i] = (*((char(*)[32])&a[0]))[i+8];
+ }
+ if (memcmp (&a, &b, sizeof (a)) != 0)
+ abort ();
}
- if (memcmp (&a, &b, sizeof (a)) != 0)
- abort ();
+#endif
return 0;
}
^ permalink raw reply [flat|nested] 2+ messages in thread
* [PATCH] Fix PR50067
@ 2011-08-24 11:15 Richard Guenther
0 siblings, 0 replies; 2+ messages in thread
From: Richard Guenther @ 2011-08-24 11:15 UTC (permalink / raw)
To: gcc-patches
This concludes the series to try to fixup data-dependence analysis.
The code is still "wrong", but it appears that the following split
out patch cures the testcases I have and does not degenerate any
in the testsuite. The patch also makes sense - we analyze the
MEM_REF base pointer for its evolution to cover the cases
of (mixed array and) pointer accesses like (*p)[i] when p
accesses into the 2st dimension of a two-dimensional array.
But it is not a good idea to claim that, if there is no evolution
of p in the analyzed nest, the MEM_REF offset (whether it is zero
or not) is the index of an independent dimension and can be used
for disambiguation by data dependence analysis. What we also
did is for not analyzable p's, put scev_not_known as the base
of the MEM_REF and pass that to the oracle(s) via dr_may_alias_p.
Something that leads to very funny disambiguations. Thus this
patch restricts us to handling POLYNOMIAL_CHRECs.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to trunk.
I'll try to forget about data-dependence analysis again now, given
it kept me busy for three days :/
Richard.
2011-08-24 Richard Guenther <rguenther@suse.de>
PR tree-optimization/50067
* tree-data-ref.c (dr_analyze_indices): Do not add an access
function for a MEM_REF base that has no evolution in the loop
nest or that is not analyzable.
* gcc.dg/torture/pr50067-3.c: New testcase.
* gcc.dg/torture/pr50067-4.c: Likewise.
* gcc.dg/torture/pr50067-5.c: Likewise.
Index: gcc/tree-data-ref.c
===================================================================
*** gcc/tree-data-ref.c.orig 2011-08-24 11:43:10.000000000 +0200
--- gcc/tree-data-ref.c 2011-08-24 12:05:24.000000000 +0200
*************** dr_analyze_indices (struct data_referenc
*** 881,905 ****
aref = TREE_OPERAND (aref, 0);
}
if (TREE_CODE (aref) == MEM_REF)
{
op = TREE_OPERAND (aref, 0);
access_fn = analyze_scalar_evolution (loop, op);
access_fn = instantiate_scev (before_loop, loop, access_fn);
! base = initial_condition (access_fn);
! split_constant_offset (base, &base, &off);
! if (!integer_zerop (TREE_OPERAND (aref, 1)))
{
! off = size_binop (PLUS_EXPR, off,
! fold_convert (ssizetype, TREE_OPERAND (aref, 1)));
! TREE_OPERAND (aref, 1)
! = build_int_cst (TREE_TYPE (TREE_OPERAND (aref, 1)), 0);
}
- access_fn = chrec_replace_initial_condition (access_fn,
- fold_convert (TREE_TYPE (base), off));
-
- TREE_OPERAND (aref, 0) = base;
- VEC_safe_push (tree, heap, access_fns, access_fn);
}
if (TREE_CODE (ref) == MEM_REF
--- 881,912 ----
aref = TREE_OPERAND (aref, 0);
}
+ /* If the address operand of a MEM_REF base has an evolution in the
+ analyzed nest, add it as an additional independent access-function. */
if (TREE_CODE (aref) == MEM_REF)
{
op = TREE_OPERAND (aref, 0);
access_fn = analyze_scalar_evolution (loop, op);
access_fn = instantiate_scev (before_loop, loop, access_fn);
! if (TREE_CODE (access_fn) == POLYNOMIAL_CHREC)
{
! base = initial_condition (access_fn);
! split_constant_offset (base, &base, &off);
! /* Fold the MEM_REF offset into the evolutions initial
! value to make more bases comparable. */
! if (!integer_zerop (TREE_OPERAND (aref, 1)))
! {
! off = size_binop (PLUS_EXPR, off,
! fold_convert (ssizetype,
! TREE_OPERAND (aref, 1)));
! TREE_OPERAND (aref, 1)
! = build_int_cst (TREE_TYPE (TREE_OPERAND (aref, 1)), 0);
! }
! access_fn = chrec_replace_initial_condition
! (access_fn, fold_convert (TREE_TYPE (base), off));
! TREE_OPERAND (aref, 0) = base;
! VEC_safe_push (tree, heap, access_fns, access_fn);
}
}
if (TREE_CODE (ref) == MEM_REF
Index: gcc/testsuite/gcc.dg/torture/pr50067-3.c
===================================================================
*** /dev/null 1970-01-01 00:00:00.000000000 +0000
--- gcc/testsuite/gcc.dg/torture/pr50067-3.c 2011-08-24 11:50:43.000000000 +0200
***************
*** 0 ****
--- 1,20 ----
+ /* { dg-do run } */
+ /* { dg-options "-fpredictive-commoning" } */
+
+ extern void abort (void);
+ int a[6] = { 0, 0, 0, 0, 7, 0 };
+ static int *p = &a[4];
+
+ int
+ main ()
+ {
+ int i;
+ for (i = 0; i < 4; ++i)
+ {
+ a[i + 1] = a[i + 2] > i;
+ *p &= ~1;
+ }
+ if (a[4] != 0)
+ abort ();
+ return 0;
+ }
Index: gcc/testsuite/gcc.dg/torture/pr50067-4.c
===================================================================
*** /dev/null 1970-01-01 00:00:00.000000000 +0000
--- gcc/testsuite/gcc.dg/torture/pr50067-4.c 2011-08-24 11:50:43.000000000 +0200
***************
*** 0 ****
--- 1,25 ----
+ /* { dg-do run } */
+
+ /* Verify we do not get a bogus access function with 0B vs. 1B which
+ disambiguates both accesses and leads to vectorization. */
+
+ extern int memcmp(const void *, const void *, __SIZE_TYPE__);
+ extern void abort (void);
+ short a[33] = { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31 };
+ short b[33] = { 0, };
+ int main()
+ {
+ #if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
+ int i;
+ if (sizeof (short) == 2)
+ {
+ for (i = 0; i < 64; ++i)
+ {
+ (*((char(*)[])&a[1]))[i] = (*((char(*)[])&a[0]))[i+1];
+ }
+ if (memcmp (&a, &b, sizeof (a)) != 0)
+ abort ();
+ }
+ #endif
+ return 0;
+ }
Index: gcc/testsuite/gcc.dg/torture/pr50067-5.c
===================================================================
*** /dev/null 1970-01-01 00:00:00.000000000 +0000
--- gcc/testsuite/gcc.dg/torture/pr50067-5.c 2011-08-24 11:50:43.000000000 +0200
***************
*** 0 ****
--- 1,28 ----
+ /* { dg-do run } */
+
+ /* Verify we do not get a bogus access function pairs with
+ exchanged dimensions, 0, {1, +, 1}_1 vs. {2B, +, 1}_1, 0 which
+ disambiguates both accesses and leads to vectorization. */
+
+ extern int memcmp(const void *, const void *, __SIZE_TYPE__);
+ extern void abort (void);
+ short a[33] = { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31 };
+ short b[33] = { 0, };
+ char * volatile ap_ = (char *)&a[0];
+ int main()
+ {
+ #if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
+ int i;
+ char *ap = ap_;
+ if (sizeof (short) == 2)
+ {
+ for (i = 0; i < 64; ++i)
+ {
+ (*((char(*)[])&ap[i+2]))[0] = (*((char(*)[])&ap[0]))[i+1];
+ }
+ if (memcmp (&a, &b, sizeof (a)) != 0)
+ abort ();
+ }
+ #endif
+ return 0;
+ }
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2011-08-24 10:06 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-08-22 11:27 [PATCH] Fix PR50067 Richard Guenther
2011-08-24 11:15 Richard Guenther
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).