tree-optimization/98674 - improve dependence analysis

This improves dependence analysis on refs that access the same
array but with different typed but same sized accesses.  That's
obviously safe for the case of types that cannot have any
access function based off them.  For the testcase this is
signed short vs. unsigned short.

2021-01-14  Richard Biener  <rguenther@suse.de>

	PR tree-optimization/98674
	* tree-data-ref.c (base_supports_access_fn_components_p): New.
	(initialize_data_dependence_relation): For two bases without
	possible access fns resort to type size equality when determining
	shape compatibility.

	* gcc.dg/vect/pr98674.c: New testcase.
This commit is contained in:
Richard Biener 2021-01-14 14:08:41 +01:00
parent a512079ef4
commit 2182274f51
2 changed files with 40 additions and 2 deletions

View file

@ -0,0 +1,16 @@
/* { dg-do compile } */
/* { dg-additional-options "-msse2" { target sse2 } } */
void swap(short *p, int cnt)
{
while (cnt-- > 0)
{
*p = ((*p << 8) & 0xFF00) | ((*p >> 8) & 0x00FF);
++p;
}
}
/* Dependence analysis should not fail. */
/* { dg-final { scan-tree-dump "dependence distance == 0" "vect" } } */
/* On x86 with SSE2 we can vectorize this with psllw/psrlw. */
/* { dg-final { scan-tree-dump "loop vectorized" "vect" { target sse2 } } } */

View file

@ -1291,6 +1291,23 @@ access_fn_component_p (tree op)
}
}
/* Returns whether BASE can have a access_fn_component_p with BASE
as base. */
static bool
base_supports_access_fn_components_p (tree base)
{
switch (TREE_CODE (TREE_TYPE (base)))
{
case COMPLEX_TYPE:
case ARRAY_TYPE:
case RECORD_TYPE:
return true;
default:
return false;
}
}
/* Determines the base object and the list of indices of memory reference
DR, analyzed in LOOP and instantiated before NEST. */
@ -3272,8 +3289,13 @@ initialize_data_dependence_relation (struct data_reference *a,
&& full_seq.start_b + full_seq.length == num_dimensions_b
&& DR_UNCONSTRAINED_BASE (a) == DR_UNCONSTRAINED_BASE (b)
&& operand_equal_p (base_a, base_b, OEP_ADDRESS_OF)
&& types_compatible_p (TREE_TYPE (base_a),
TREE_TYPE (base_b))
&& (types_compatible_p (TREE_TYPE (base_a),
TREE_TYPE (base_b))
|| (!base_supports_access_fn_components_p (base_a)
&& !base_supports_access_fn_components_p (base_b)
&& operand_equal_p
(TYPE_SIZE (TREE_TYPE (base_a)),
TYPE_SIZE (TREE_TYPE (base_b)), 0)))
&& (!loop_nest.exists ()
|| (object_address_invariant_in_loop_p
(loop_nest[0], base_a))));