This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug pending/30773] Spec cpu2k6/h264ref and sphinx3 miscompare regression
- From: "grigory_zagorodnev at linux dot intel dot com" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 14 Feb 2007 13:40:22 -0000
- Subject: [Bug pending/30773] Spec cpu2k6/h264ref and sphinx3 miscompare regression
- References: <bug-30773-7719@http.gcc.gnu.org/bugzilla/>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Comment #2 from grigory_zagorodnev at linux dot intel dot com 2007-02-14 13:40 -------
Created an attachment (id=13048)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=13048&action=view)
Source code for minimal reproducer
There is a minimal reproducer for the problem. Test-case is meaningless from
the execution point of view. To see the bug, you need to generate and compare
assembly output with "good" and "bad" compiler (gcc -O2 -S utt.i).
bash-3.00$ diff -up utt.s utt.s.bad
--- utt.s 2007-02-14 16:27:58.000000000 +0300
+++ utt.s.bad 2007-02-14 16:27:48.000000000 +0300
@@ -145,9 +145,9 @@ utt_end:
movl %edx, %eax
sarl $31, %edx
idivl 8(%ecx)
- movl $.LC2, 4(%esp)
- movl %eax, 12(%esp)
+ movl %eax, -24628(%ebp)
movl stdout, %eax
+ movl $.LC2, 4(%esp)
movl %eax, (%esp)
call fprintf
cmpl %ebx, %edi
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30773