This is the mail archive of the
gcc-prs@gcc.gnu.org
mailing list for the GCC project.
Re: optimization/6883: Fails to optimize temporary objects.
- From: Richard Guenther <rguenth at tat dot physik dot uni-tuebingen dot de>
- To: nobody at gcc dot gnu dot org
- Cc: gcc-prs at gcc dot gnu dot org,
- Date: 11 May 2003 12:06:01 -0000
- Subject: Re: optimization/6883: Fails to optimize temporary objects.
- Reply-to: Richard Guenther <rguenth at tat dot physik dot uni-tuebingen dot de>
The following reply was made to PR optimization/6883; it has been noted by GNATS.
From: Richard Guenther <rguenth@tat.physik.uni-tuebingen.de>
To: Dara Hazeghi <dhazeghi@yahoo.com>
Cc: gcc-gnats@gcc.gnu.org, <rguenth@tat.physik.uni-tuebingen.de>
Subject: Re: optimization/6883: Fails to optimize temporary objects.
Date: Sun, 11 May 2003 14:00:37 +0200 (CEST)
On Sat, 10 May 2003, Dara Hazeghi wrote:
> http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-
> trail&database=gcc&pr=6883
>
> Hello,
>
> this bug was reported against gcc 3.1. Would it be possible to test
> your testcase against a more current version of gcc (ie 3.2.3 or 3.3
> prerelease) and report back on the results? Thanks,
Some more information, based on my last reply - the Intel compiler
Intel(R) C++ Compiler for 32-bit applications, Version 7.1 Build
20030424Z produces with -O2
iterators with temporaries
..B1.18: # Preds ..B1.18 ..B1.17
lea 1(%esi), %edi #65.24
cmpl $254, %edi #66.16
fldl -8(%edx,%esi,8) #65.21
faddl 8(%edx,%esi,8) #65.28
fmull (%eax,%esi,8) #65.32
fstpl (%ecx,%esi,8) #65.9
movl %edi, %esi #65.24
jle ..B1.18 # Prob 97% #66.16
iterators without temporaries
..B1.21: # Preds ..B1.21 ..B1.20 # Infreq
fldl -8(%edx,%esi,8) #74.17
faddl 8(%edx,%esi,8) #74.25
fmull (%eax,%esi,8) #74.34
fstpl (%ecx,%esi,8) #74.9
incl %esi #75.16
cmpl $254, %esi #75.16
jle ..B1.21 # Prob 97% #75.16
no iterators, just int
..B1.24: # Preds ..B1.24 ..B1.23 # Infreq
lea 1(%edi), %esi #83.24
fldl -8(%eax,%edi,8) #83.17
faddl 8(%eax,%edi,8) #83.24
fmull (%edx,%edi,8) #83.32
fstpl (%ecx,%edi,8) #83.9
addl $2, %edi #83.24
cmpl $254, %edi #84.16
fldl -8(%eax,%esi,8) #83.17
faddl 8(%eax,%esi,8) #83.24
fmull (%edx,%esi,8) #83.32
fstpl (%ecx,%esi,8) #83.9
jle ..B1.24 # Prob 97% #84.16
it even detects the loop is run even times and unrolls it 2-times ;)
This is what I like the code to get optimized to, at least all three
versions look nearly the same.
Richard.