On 5/1/07, Daniel Stenberg <daniel_at_haxx.se> wrote:

> > The question is about caching memory pools for internal libcurl housekeeping
> > allocations, where many places are using the "raw" calloc, malloc and free,
> > unless this is done for purpose.
>
> Do you have any numbers/estimates on what we could gain by doing this
> suggested move?

The issue, that we have now is a certain performance degradation
within tens of minutes upon a high CPU and low memory conditions .
Shotting oprofile at such conditions is tricky, but we will try some tricks.
We are suspecting intensive memory defragmentation efforts by glibc,
which impacts the performance.

curl-loader does not make any run-time allocs/free and not much allocs
is doing libevent.

Agree, that we need to try the pools via the nice global_init_mem
facilty and collect
more performance data on our side.

-- 
Sincerely,
Robert Iakobashvili,
coroberti %x40 gmail %x2e com
...........................................................
http://curl-loader.sourceforge.net
A web testing and traffic generation tool.