|
From: James A. <jam...@me...> - 2004-08-02 14:32:35
|
i was testing out the 2.1.2 version of valgrind on my redhat 7.3 box, and it was virtually unuseable. i ran 2.0.0 fine with our application (slow, but reasonable). with 2.1.2, the screen takes forever for each repaint, such that it's almost impossible to use. any ideas on what changes might have caused this? any ideas how i could diagnose the slowdown? thanks, -james |
|
From: Nicholas N. <nj...@ca...> - 2004-08-03 17:30:09
|
On Mon, 2 Aug 2004, James Ahlborn wrote: > i was testing out the 2.1.2 version of valgrind on my redhat 7.3 box, and it was > virtually unuseable. i ran 2.0.0 fine with our application (slow, but > reasonable). with 2.1.2, the screen takes forever for each repaint, such that > it's almost impossible to use. any ideas on what changes might have caused > this? any ideas how i could diagnose the slowdown? That sounds unusual... I'd guess it's caused by the changes in the way system calls are handled, esp. block system calls. The output of --trace-syscalls=yes might help identify the problem if it's a particular system call that's the problem. You could also try the little-used --lowlat-signals=yes and --lowlat-syscalls=yes options? If they don't work, it would be great if you could file a bug report (see valgrind.kde.org/bugs.html). Thanks. HTH N |