optimization/2495: -O optimization creates roundoff errors on Linux machine

mlister@nrao.edu mlister@nrao.edu
Thu Apr 5 12:06:00 GMT 2001


>Number:         2495
>Category:       optimization
>Synopsis:       -O optimization creates roundoff errors on Linux machine
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    unassigned
>State:          open
>Class:          wrong-code
>Submitter-Id:   net
>Arrival-Date:   Thu Apr 05 12:06:00 PDT 2001
>Closed-Date:
>Last-Modified:
>Originator:     Matthew Lister
>Release:        gcs-2.91.66 19990314/Linux (egcs-1.1.2 release)
>Organization:
>Environment:
Dell Pentium III (Coppermine) 650 MHz
Red Hat Linux Distribution Release 6.2 (Zoot) 
Operating system version 2.2.17-14
>Description:
When the gcc compiler flag '-O' is invoked on the above 
system, floating point precision appears to be altered for 
variables passed to subroutines. The result
is incorrect behaviour of certain mathematical 
functions such as floor(), as demonstrated by the
attached program. The error goes away if '-O' is not
used. The problem appears to be limited to Linux machines,
I have tried it on Solaris with the '-O' option and the
test program shows no errors. 
>How-To-Repeat:
Compile the attached C program on a Linux system with
'gcc -lm -O '
>Fix:
Compile the program without the '-O' optimization option. 
>Release-Note:
>Audit-Trail:
>Unformatted:
 
----gnatsweb-attachment----
Content-Type: application/x-tar; name="test.tar.gz"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="test.tar.gz"
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More information about the Gcc-bugs mailing list