Branches
Comments
[»]
Exceptions
by edg - Jun 21st 2007 02:28:06
At least 1.5.4 doesn't seem to handle exceptions well. Functions that exit
with exceptions result in having some unrealistic timings.
[reply]
[top]
[»]
Installation
by edg - Jun 20th 2007 09:28:51
There were 2 errors I encountered during installation.
1) fncmalloc.c:34: error: `RTLD_NEXT' undeclared
2) /home/edg/FunctionCheck/fnccheck/test/sub.c:14: undefined reference to
`sin'
To fix them, after ./configure is finished I edited config.status changing
CFLAGS line to "s%@CFLAGS@%-g -D_GNU_SOURCE -lm -O2%g".
-D_GNU_SOURCE is fix for the first error and -lm for the second. Hope this
helps any other noob than me.
[reply]
[top]
[»]
there are 3 forks of this project
by jgabriels - Apr 19th 2006 17:56:30
There are 3 forks of this project
Version 1.4/2.0 is the original and is no
longer being actively developed.
Version 3.0 is a fork that is also dead.
Version 1.5.4 is the only active one and also
the only one that supports amd64 processors.
[reply]
[top]
[»]
Not for SMP
by Howard Chu - Oct 8th 2005 20:17:33
By the way, the multithreading support barely works on a single-processor
system; it is hopeless on SMP. Unfortunately it has no way of knowing when
a thread context switch occurs, so the allocation of time to functions is
unreliable. It still works fine for single-threaded programs or
single-processor machines, but I don't see any easy solutions for making
the profiler accurate on a multi-processor machine. At least it's still
usable as a malloc tracer, and it doesn't eat up as much address space as
valgrind.
[reply]
[top]
|