Bug 46578 - 416.gamess in SPEC CPU 2006 is miscompiled by LTO with gold
Summary: 416.gamess in SPEC CPU 2006 is miscompiled by LTO with gold
Status: RESOLVED FIXED
Alias: None
Product: gcc
Classification: Unclassified
Component: lto (show other bugs)
Version: 4.6.0
: P3 normal
Target Milestone: 4.7.0
Assignee: Not yet assigned to anyone
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-11-20 14:41 UTC by H.J. Lu
Modified: 2012-05-07 15:55 UTC (History)
1 user (show)

See Also:
Host:
Target:
Build:
Known to work:
Known to fail:
Last reconfirmed: 2012-05-07 00:00:00

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description H.J. Lu 2010-11-20 14:41:29 UTC 
On Linux/ia32/x86-64, 416.gamess in SPEC CPU 2006 is miscompiled by

-O3 -mfpmath=sse -msse2 -funroll-loops -ffast-math  -fwhole-program
-flto=jobserver -fuse-linker-plugin  

when GCC is configured with --with-plugin-ld=ld.gold.

  Running 416.gamess ref peak lto default

416.gamess: copy 0 non-zero return code (exit code=0, signal=11)


****************************************
Contents of cytosine.2.err
****************************************
STOP IN ABRT

****************************************

****************************************
Contents of triazolium.err
****************************************
STOP IN ABRT

The same benchmark works fine when GCC is configured with --with-plugin-ld=ld,
using binutils as of 2010-10-20 CVS.
Comment 1 Jan Hubicka 2011-01-08 20:44:28 UTC 
Are there any noticeable differences in the resolution files passed to the plugin?
Comment 2 Richard Biener 2012-05-07 12:01:17 UTC 
Is this still an issue?  Please close this if it does not persist with 4.7.
Comment 3 H.J. Lu 2012-05-07 15:55:14 UTC 
Gold 20120430 on trunk works OK with GCC 4.7/4.8.