| Summary: | Result of conversion of double to long long differs in valgrind and non-valgrind runs | ||
|---|---|---|---|
| Product: | [Developer tools] valgrind | Reporter: | rudolf.hornig |
| Component: | general | Assignee: | Julian Seward <jseward> |
| Status: | RESOLVED DUPLICATE | ||
| Severity: | normal | CC: | njn, tom |
| Priority: | NOR | ||
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Platform: | Ubuntu | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed In: | ||
This is probably a dup of bug 197915. Rudolf, what do you think? Confirmed. After some research, I have recompiled the code with -msse2 -mfpmath=sse options (forcing to use 64bit floats on an intel CPU). and the valgrind and non-valgind output is now the same. Sorry for the confusion. This is really a dauplicate. *** This bug has been marked as a duplicate of bug 197915 *** |
Version: valgrind-3.4.1-Debian (using KDE 4.2.4) Compiler: gcc 4.3 OS: Linux Installed from: Ubuntu Packages uname -a Linux hal90000 2.6.28-13-generic #45-Ubuntu SMP Tue Jun 30 19:49:51 UTC 2009 i686 GNU/Linux The output of the followinfg program differs it it is run under valgrind: #include <stdio.h> int main() { char *x = "\x19\x6d\xb2\x9a\xba\x7e\x15\x40"; double d = *(double *)x; printf("\nlong long:%lld \n", (long long)(d*(1e12))); } The output without valgrind: long long:5373758713855 using valgrind: long long:5373758713856