|
From: bsreddy <bsr...@ya...> - 2007-04-04 19:12:09
|
I am using valgrind -massif for memory profiling. When I look at the generated ps/pdf file, huge chunk of memory is categorised as "other". Any pointers why the huge memory consuming part of my "normal" code is being clubbed into "other" section?? -- View this message in context: http://www.nabble.com/valgrind--massif-generated-ps-file-has-huge-memory-in-%22other%22-section-tf3527716.html#a9843667 Sent from the Valgrind - Users mailing list archive at Nabble.com. |
|
From: Nicholas N. <nj...@cs...> - 2007-04-04 22:11:45
|
On Wed, 4 Apr 2007, bsreddy wrote:
> I am using valgrind -massif for memory profiling.
>
> When I look at the generated ps/pdf file, huge chunk of memory is
> categorised as "other".
>
> Any pointers why the huge memory consuming part of my "normal" code is being
> clubbed into "other" section??
Basically it shows the biggest 15--20, and the rest get aggregated. It's
not ideal. I've been working on Massif recently to improve things. You
might like to try the version on the MASSIF2 branch by checking it out:
svn co svn://svn.valgrind.org/valgrind/branches/MASSIF2 massif2
cd massif2
then build as per the instructions in README. The code is still
experimental, and not at all documented, but it may be of use. The main
difference is that the notion of "spacetime" is no longer used, rather it
takes detailed snapshots of the memory usage every so often and gives you
full information about those snapshot points.
Nick
|