public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug rtl-optimization/20306] New: [4.0/4.1 Regression] FP complex modes moved via GPRs
@ 2005-03-03 19:24 dje at gcc dot gnu dot org
  2005-03-03 19:26 ` [Bug rtl-optimization/20306] " dje at gcc dot gnu dot org
                   ` (5 more replies)
  0 siblings, 6 replies; 7+ messages in thread
From: dje at gcc dot gnu dot org @ 2005-03-03 19:24 UTC (permalink / raw)
  To: gcc-bugs

The patch for PR rtl-opt/15289, changed the behavior of emit_move_insn_1 to 
invoke emit_block_move on memory to memory moves for "optimal behavior", 
instead of invoking a movMM pattern on the inner mode.  emit_block_move 
changes the mode to BLKmode masking the original mode.  This sequence of 
decisions results in GCC choosing GPRs to move a value that explicitly is 
declared floating point in the user code, hurting performance.

-- 
           Summary: [4.0/4.1 Regression] FP complex modes moved via GPRs
           Product: gcc
           Version: 4.0.0
            Status: UNCONFIRMED
          Keywords: missed-optimization
          Severity: normal
          Priority: P2
         Component: rtl-optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: dje at gcc dot gnu dot org
                CC: gcc-bugs at gcc dot gnu dot org,pthaugen at us dot ibm
                    dot com
 GCC build triplet: powerpc*-*-*
  GCC host triplet: powerpc*-*-*
GCC target triplet: powerpc*-*-*


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20306


^ permalink raw reply	[flat|nested] 7+ messages in thread

* [Bug rtl-optimization/20306] [4.0/4.1 Regression] FP complex modes moved via GPRs
  2005-03-03 19:24 [Bug rtl-optimization/20306] New: [4.0/4.1 Regression] FP complex modes moved via GPRs dje at gcc dot gnu dot org
@ 2005-03-03 19:26 ` dje at gcc dot gnu dot org
  2005-03-03 19:38 ` pinskia at gcc dot gnu dot org
                   ` (4 subsequent siblings)
  5 siblings, 0 replies; 7+ messages in thread
From: dje at gcc dot gnu dot org @ 2005-03-03 19:26 UTC (permalink / raw)
  To: gcc-bugs


------- Additional Comments From dje at gcc dot gnu dot org  2005-03-03 19:26 -------
Example Fortran code derived from BLAS ZGEMM routine.

      SUBROUTINE Z ( M, N, C, LDC )
*     .. Scalar Arguments ..
      INTEGER            M, N, LDC
*     .. Array Arguments ..
      COMPLEX*16         C( LDC, * )
*     .. Local Scalars ..
      INTEGER            I, J
*     .. Parameters ..
      COMPLEX*16         ZERO
      PARAMETER        ( ZERO = ( 0.0D+0, 0.0D+0 ) )
*
            DO 20, J = 1, N
               DO 10, I = 1, M
                  C( I, J ) = ZERO
   10          CONTINUE
   20       CONTINUE
*
      RETURN
      END

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20306


^ permalink raw reply	[flat|nested] 7+ messages in thread

* [Bug rtl-optimization/20306] [4.0/4.1 Regression] FP complex modes moved via GPRs
  2005-03-03 19:24 [Bug rtl-optimization/20306] New: [4.0/4.1 Regression] FP complex modes moved via GPRs dje at gcc dot gnu dot org
  2005-03-03 19:26 ` [Bug rtl-optimization/20306] " dje at gcc dot gnu dot org
@ 2005-03-03 19:38 ` pinskia at gcc dot gnu dot org
  2005-03-03 19:48 ` pinskia at gcc dot gnu dot org
                   ` (3 subsequent siblings)
  5 siblings, 0 replies; 7+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2005-03-03 19:38 UTC (permalink / raw)
  To: gcc-bugs



-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pinskia at gcc dot gnu dot
                   |                            |org
   Target Milestone|---                         |4.0.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20306


^ permalink raw reply	[flat|nested] 7+ messages in thread

* [Bug rtl-optimization/20306] [4.0/4.1 Regression] FP complex modes moved via GPRs
  2005-03-03 19:24 [Bug rtl-optimization/20306] New: [4.0/4.1 Regression] FP complex modes moved via GPRs dje at gcc dot gnu dot org
  2005-03-03 19:26 ` [Bug rtl-optimization/20306] " dje at gcc dot gnu dot org
  2005-03-03 19:38 ` pinskia at gcc dot gnu dot org
@ 2005-03-03 19:48 ` pinskia at gcc dot gnu dot org
  2005-03-11 22:59 ` cvs-commit at gcc dot gnu dot org
                   ` (2 subsequent siblings)
  5 siblings, 0 replies; 7+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2005-03-03 19:48 UTC (permalink / raw)
  To: gcc-bugs


------- Additional Comments From pinskia at gcc dot gnu dot org  2005-03-03 19:48 -------
Confirmed, C testcase which shows that this is defintely a regression:
void f(double _Complex *f,int len)
{
  int i = 0;
  for(i = 0 ;i<20;i++)
  {
    *f = 0.0;
    f++;
  }
}


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
      Known to fail|                            |4.0.0 4.1.0
      Known to work|                            |3.3.3
   Last reconfirmed|0000-00-00 00:00:00         |2005-03-03 19:48:05
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20306


^ permalink raw reply	[flat|nested] 7+ messages in thread

* [Bug rtl-optimization/20306] [4.0/4.1 Regression] FP complex modes moved via GPRs
  2005-03-03 19:24 [Bug rtl-optimization/20306] New: [4.0/4.1 Regression] FP complex modes moved via GPRs dje at gcc dot gnu dot org
                   ` (2 preceding siblings ...)
  2005-03-03 19:48 ` pinskia at gcc dot gnu dot org
@ 2005-03-11 22:59 ` cvs-commit at gcc dot gnu dot org
  2005-03-11 23:04 ` cvs-commit at gcc dot gnu dot org
  2005-03-21 16:07 ` dje at gcc dot gnu dot org
  5 siblings, 0 replies; 7+ messages in thread
From: cvs-commit at gcc dot gnu dot org @ 2005-03-11 22:59 UTC (permalink / raw)
  To: gcc-bugs


------- Additional Comments From cvs-commit at gcc dot gnu dot org  2005-03-11 22:59 -------
Subject: Bug 20306

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	dje@gcc.gnu.org	2005-03-11 22:58:53

Modified files:
	gcc            : ChangeLog expr.c 

Log message:
	PR rtl-optimization/20306
	* expr.c (emit_move_complex): Set try_int false if mode is
	MODE_COMPLEX_FLOAT and mov_optab exists for inner mode.  Only try
	emit_block_move if try_int is true.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.7789&r2=2.7790
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/expr.c.diff?cvsroot=gcc&r1=1.781&r2=1.782



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20306


^ permalink raw reply	[flat|nested] 7+ messages in thread

* [Bug rtl-optimization/20306] [4.0/4.1 Regression] FP complex modes moved via GPRs
  2005-03-03 19:24 [Bug rtl-optimization/20306] New: [4.0/4.1 Regression] FP complex modes moved via GPRs dje at gcc dot gnu dot org
                   ` (3 preceding siblings ...)
  2005-03-11 22:59 ` cvs-commit at gcc dot gnu dot org
@ 2005-03-11 23:04 ` cvs-commit at gcc dot gnu dot org
  2005-03-21 16:07 ` dje at gcc dot gnu dot org
  5 siblings, 0 replies; 7+ messages in thread
From: cvs-commit at gcc dot gnu dot org @ 2005-03-11 23:04 UTC (permalink / raw)
  To: gcc-bugs


------- Additional Comments From cvs-commit at gcc dot gnu dot org  2005-03-11 23:04 -------
Subject: Bug 20306

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	gcc-4_0-branch
Changes by:	dje@gcc.gnu.org	2005-03-11 23:03:56

Modified files:
	gcc            : ChangeLog expr.c 

Log message:
	PR rtl-optimization/20306
	* expr.c (emit_move_complex): Set try_int false if mode is
	MODE_COMPLEX_FLOAT and mov_optab exists for inner mode.  Only try
	emit_block_move if try_int is true.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-4_0-branch&r1=2.7592.2.43&r2=2.7592.2.44
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/expr.c.diff?cvsroot=gcc&only_with_tag=gcc-4_0-branch&r1=1.778&r2=1.778.6.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20306


^ permalink raw reply	[flat|nested] 7+ messages in thread

* [Bug rtl-optimization/20306] [4.0/4.1 Regression] FP complex modes moved via GPRs
  2005-03-03 19:24 [Bug rtl-optimization/20306] New: [4.0/4.1 Regression] FP complex modes moved via GPRs dje at gcc dot gnu dot org
                   ` (4 preceding siblings ...)
  2005-03-11 23:04 ` cvs-commit at gcc dot gnu dot org
@ 2005-03-21 16:07 ` dje at gcc dot gnu dot org
  5 siblings, 0 replies; 7+ messages in thread
From: dje at gcc dot gnu dot org @ 2005-03-21 16:07 UTC (permalink / raw)
  To: gcc-bugs


------- Additional Comments From dje at gcc dot gnu dot org  2005-03-21 16:07 -------
Patch confirmed to restore correct behavior.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20306


^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2005-03-21 16:07 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2005-03-03 19:24 [Bug rtl-optimization/20306] New: [4.0/4.1 Regression] FP complex modes moved via GPRs dje at gcc dot gnu dot org
2005-03-03 19:26 ` [Bug rtl-optimization/20306] " dje at gcc dot gnu dot org
2005-03-03 19:38 ` pinskia at gcc dot gnu dot org
2005-03-03 19:48 ` pinskia at gcc dot gnu dot org
2005-03-11 22:59 ` cvs-commit at gcc dot gnu dot org
2005-03-11 23:04 ` cvs-commit at gcc dot gnu dot org
2005-03-21 16:07 ` dje at gcc dot gnu dot org

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).