Hello,
I found one more problem with ptmalloc and registration cache.
In arena.c:grow_heap() when heap is shrinking ptmalloc tries to be smart
and is using mmap() to change pages protection instead of mprotect() because
as a side effect mmap() drops underlying pages. In the case the area is
Thanks for catching this. I agree with you, the first option seems
like it is closer to what we want to have happen in Open MPI. I'm
less concerned with dropping a couple of pages of memory as I am with
searching our registration cache more often. It shall be committed
to SVN today.
Hello there,
Proposed below is a patch to remove the redundant instance of the
version algorithm, together with some cleanup (for unification of
the two algorithms, which have diverged a little) and a micro bit
more efficiency.
One could add a rebuilding rule to config/Makefile.am, if you think
On Jan 11, 2006, at 3:05 AM, Rainer Keller wrote:
Hello dear all,
I had a point on the tbd-list, that I would like to ask here:
- Shouldn't we have a while-loop condition around every occurence
of opal_condition_wait (spurious wake-ups)
As we may do a pthread_cond_wait,
e.g. in
Thanks for the patch - I'm going to try to grok it and then commit it
this evening.
Brian
On Jan 18, 2006, at 1:22 PM, Ralf Wildenhues wrote:
Hello there,
Proposed below is a patch to remove the redundant instance of the
version algorithm, together with some cleanup (for unification of