|
From: Dennis L. <pla...@in...> - 2005-06-26 19:24:03
|
Hi, I tried to compile valgrind 2.4 svn (today) with compiler optimization (maybe it runs faster then) and tried to use combinations of: -O3 -march=pentium4 -mtune=pentium4 mfpmath=sse,387 -ffast-math and some others but even -O3 seems to produce a valgrind that isnt able to run and crashes with some assertions. Is this known, intended or buggy behaviour ? It would be nice if I could at least compile valgrind with a quite high optimization and support for current processor, since things are really time-intensive when running with valgrind. greets Dennis Carpe quod tibi datum est |
|
From: Nicholas N. <nj...@cs...> - 2005-06-26 19:38:27
|
On Sun, 26 Jun 2005, Dennis Lubert wrote: > I tried to compile valgrind 2.4 svn (today) with compiler optimization (maybe > it runs faster then) and tried to use combinations of: > -O3 -march=pentium4 -mtune=pentium4 mfpmath=sse,387 -ffast-math and some > others but even -O3 seems to produce a valgrind that isnt able to run and > crashes with some assertions. Is this known, intended or buggy behaviour ? It > would be nice if I could at least compile valgrind with a quite high > optimization and support for current processor, since things are really > time-intensive when running with valgrind. I don't know why this would be the case. But compiling with more optimisations may not help much, since much of Valgrind's runtime is spent executing code that it has compiled itself. Nick |