OpenE2K
/
gcc
2
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

re PR rtl-optimization/34522 (inefficient code for long long multiply when only low bits are needed) 

2008-03-12  Paolo Bonzini  <bonzini@gnu.org>

	PR tree-opt/35422
	* fold-const.c (fold_unary) <NOP_EXPR>: Distribute a narrowing
	conversion to the operands of a multiplication.

testsuite:
2008-03-12  Paolo Bonzini  <bonzini@gnu.org>

	PR tree-opt/35422
	* gcc.dg/vect/slp-7.c: Change target keywords required for vectorizing
	third loop.
	* gcc.target/i386/pr35422.c: New.

From-SVN: r133144
This commit is contained in:
Paolo Bonzini 2008-03-12 15:33:45 +00:00 committed by Paolo Bonzini
parent 70fef63aef
commit c83bd37c9d
5 changed files with 50 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
gcc/ChangeLog
View File

@ -1,3 +1,9 @@
2008-03-12 Paolo Bonzini <bonzini@gnu.org>
PR tree-opt/35422
* fold-const.c (fold_unary) <NOP_EXPR>: Distribute a narrowing
conversion to the operands of a multiplication.
2008-03-12 Richard Guenther <rguenther@suse.de>
* Makefile.in (OBJS-common): Add tree-ssa-phiprop.o

20
gcc/fold-const.c
View File

@ -7926,6 +7926,26 @@ fold_unary (enum tree_code code, tree type, tree op0)
return fold_build1 (BIT_NOT_EXPR, type, fold_convert (type, tem));
}
/* Convert (T1)(X * Y) into (T1)X * (T1)Y if T1 is narrower than the
type of X and Y (integer types only). */
if (INTEGRAL_TYPE_P (type)
&& TREE_CODE (op0) == MULT_EXPR
&& INTEGRAL_TYPE_P (TREE_TYPE (op0))
&& TYPE_PRECISION (type) < TYPE_PRECISION (TREE_TYPE (op0)))
{
/* Be careful not to introduce new overflows. */
tree mult_type;
if (TYPE_OVERFLOW_WRAPS (type))
mult_type = type;
else
mult_type = unsigned_type_for (type);
tem = fold_build2 (MULT_EXPR, mult_type,
fold_convert (mult_type, TREE_OPERAND (op0, 0)),
fold_convert (mult_type, TREE_OPERAND (op0, 1)));
return fold_convert (type, tem);
}
tem = fold_convert_const (code, type, op0);
return tem ? tem : NULL_TREE;

7
gcc/testsuite/ChangeLog
View File

@ -1,3 +1,10 @@
2008-03-12 Paolo Bonzini <bonzini@gnu.org>
PR tree-opt/35422
* gcc.dg/vect/slp-7.c: Change target keywords required for vectorizing
third loop.
* gcc.target/i386/pr35422.c: New.
2008-03-11 Andrew Pinski <andrew_pinski@playstation.sony.com>
PR tree-opt/35403

7
gcc/testsuite/gcc.dg/vect/slp-7.c
View File

@ -120,8 +120,9 @@ int main (void)
return 0;
}
/* { dg-final { scan-tree-dump-times "vectorized 3 loops" 1 "vect" {target { vect_strided && vect_int_mult } } } }*/
/* { dg-final { scan-tree-dump-times "vectorized 2 loops" 1 "vect" {target { ! { vect_strided && vect_int_mult } } } } } */
/* { dg-final { scan-tree-dump-times "vectorizing stmts using SLP" 2 "vect" } } */
/* { dg-final { scan-tree-dump-times "vectorized 3 loops" 1 "vect" { target { vect_unpack && vect_int_mult } } } }*/
/* { dg-final { scan-tree-dump-times "vectorized 2 loops" 1 "vect" { target { ! { vect_unpack && vect_int_mult } } } } }*/
/* { dg-final { scan-tree-dump-times "vectorizing stmts using SLP" 3 "vect" { target { vect_unpack && vect_int_mult } } } } */
/* { dg-final { scan-tree-dump-times "vectorizing stmts using SLP" 2 "vect" { target { ! { vect_unpack && vect_int_mult } } } } } */
/* { dg-final { cleanup-tree-dump "vect" } } */

13
gcc/testsuite/gcc.target/i386/pr34522.c Normal file
View File

@ -0,0 +1,13 @@
/* { dg-options "-O2" } */
/* { dg-do compile } */
/* { dg-require-effective-target ilp32 } */
int test(long long a, long long b)
{
return a * b;
}
/* Check that we did not spill anything. This is all that is needed
to qualify the generated code as "decent"... */
/* { dg-final { scan-assembler-not "%e[sd]i" } } */