Bug 63895 - [5 Regression] r217588 miscompled 178.galgel from cpu2000
Summary: [5 Regression] r217588 miscompled 178.galgel from cpu2000
Status: RESOLVED FIXED
Alias: None
Product: gcc
Classification: Unclassified
Component: rtl-optimization (show other bugs)
Version: 5.0
: P1 normal
Target Milestone: 5.0
Assignee: Not yet assigned to anyone
URL:
Keywords: wrong-code
Depends on:
Blocks:
 
Reported: 2014-11-15 17:59 UTC by H.J. Lu
Modified: 2014-11-18 21:26 UTC (History)
3 users (show)

See Also:
Host:
Target: x86_64-*-*
Build:
Known to work:
Known to fail:
Last reconfirmed: 2014-11-15 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 2014-11-15 17:59:27 UTC 
On Linux/x86-64, r217588 miscompiled 178.galgel from cpu2000:

# gfortran     -O3 -funroll-loops -ffast-math  -DSPEC_CPU2000_LP64 ....
...
0203:                 C1 = (  -0.3411    ,  0.2388    )
                      C1 = (   0.2181E+13, -0.3477E+12)
                                        ^
0206:                 Error in right eigenvector is   0.000    
                      Error in right eigenvector is   156.7    
                                                          ^
0249:                 D(Lambda)/D(Gr) = (   0.3254E-04,  0.6779E-04)
                      D(Lambda)/D(Gr) = (  -0.4526E-13, -0.6501E-13)
                                                     ^
0251:            Calculated estimations of errors:
        Error in BifGel: D(alpha)/D(Anu) = 0
       ^
0252:                 Error in calculation of Omega       =  0.000    
        BifGel: weakly nonlinear analysis failed
       ^
0253:                 Error in calculation of Amu2        =  0.000    
        Error return from BifGel called by Cyklap: ier=  139
             ^
'galgel.out' short
Comment 1 H.J. Lu 2014-11-15 18:21:20 UTC 
It also miscompiled

 1x171.swim 1x176.gcc 1x178.galgel 1x253.perlbmk 1x301.apsi

from SPEC CPU 2000 on ia32 with

-m32     -O3 -funroll-loops -msse2 -mfpmath=sse -ffast-math
Comment 2 Vladimir Makarov 2014-11-18 21:19:51 UTC 
Sorry, I can not reproduce any reported failures.  I guess it was before the patch

https://gcc.gnu.org/ml/gcc-patches/2014-11/msg01995.html

The patch most probably fixed it.

Could you confirm that the problem was gone (or still exists).
Comment 3 H.J. Lu 2014-11-18 21:26:37 UTC 
Fixed by r217624.