This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: Speaking on malloc internals at FOSS.in
On Wed, Nov 28, 2012 at 06:59:34AM +0530, Siddhesh Poyarekar wrote:
> On Tue, Nov 27, 2012 at 11:37:34PM +0100, OndÅej BÃlka wrote:
> >
> > I drawn graphs how malloc behaves. It could be downloaded at
> >
> > http://kam.mff.cuni.cz/~ondra/benchmark_string/malloc_profile_26_11_2012.tar.bz2
> >
> > In NOTES file is rougthly what I would say about results.
> >
> > If you need to change figures use gen_eps script or data in appropriate
> > files.
> >
> > I now work on measuring lifetimes of allocated blocks.
> >
>
> Thanks! I had embarked on a similar project earlier this year, but
> had limited it to specific applications since I was trying to solve a
> different problem - trying to figure out resource usage.
>
> > One problem is that we cannot hook calloc yet.
>
> This is a known issue and it's because ld.so uses calloc and we go
> into this complete tailspin when we try to hook calloc. It's been on
> my list of to-look-at items since the first time I ran into it earlier
> this year.
Now I added lifetimes. I it still measured on simulated workloads.
Also calloc/memalign accounts to most allocations.
http://kam.mff.cuni.cz/~ondra/benchmark_string/malloc_profile_28_11_2012.tar.bz2