I don't know about "drama", but people did clearly explain to you why this approach was unacceptable. You simply cannot cross-link at the component level. If you need something from the opal/mca/memory framework, you have to get it from the framework level.

Doesn't seem that hard a concept to grasp and follow - failing to do so breaks things for a bunch of people, which is why we don't allow it. So I hope your "configure" approach also takes this into account, or we'll have to revert it again :-(


On Apr 13, 2012, at 11:13 AM, Mike Dubman wrote:

Too many drama - we will fix it to detect hooks availability at configure stage, this will make your life back to normal.
 
The problem is not a Mellanox hw, but Intel PCI bus implementation, which charge extra latency if buffers are not aligned.
The patch is a workaround for this problem and help to non-benchmark code as well.
 

 
On Fri, Apr 13, 2012 at 7:06 PM, Barrett, Brian W <bwbarre@sandia.gov> wrote:
r2655 is awful as a patch.  It doesn't work on any non-Linux platform,
which is unpleasant.  But worse, what does it possibly accomplish?  In
codes other than benchmarks, there's no advantage to aligning the pointer
to 32 or 64 byte boundaries, as the malloced buffer very rarely is exactly
what is sent.  So you've done a whole lot of work, screwed with the memory
allocator (which always bites OMPI in the butt), and accomplished nothing
useful.  Mellanox should fix the hardware, not make everyone's life
miserable with crappy workarounds.

MEMORY_LINUX_PTMALLOC2 is the wrong define for what they want.  They
should check for __malloc_hook and only use that code if __malloc_hook is
found.

Brian

On 4/13/12 9:32 AM, "TERRY DONTJE" <terry.dontje@oracle.com> wrote:

>
>
>
>    I am thinking MEMORY_LINUX_PTMALLOC2 is probably the right define to
>    key off of but this is really going to look gross ifdef'ing out the
>    lines that are accessing the Linux memory module.  One other idea I
>    have is to create a dummy __malloc_hook in the Solaris memory module
>    but might there be other OSes that could run into the same
>    problem?   Or what happens if PTMALLOC2 is not used (does that
>    happen)?
>
>    --td
>
>    On 4/13/2012 10:45 AM, TERRY DONTJE wrote:
>
>
>      r26255 is forcing the use of __malloc_hook which is implemented in
>      opal/mca/memory/linux however that is not compiled in the library
>      when built on Solaris thus causing a referenced symbol not found
>      when libmpi tries to load the openib btl.
>
>      I am looking how to fix this now but if someone has a good idea
>      how to detect when __malloc_hook is used (or not) I'd be
>      interested in hearing it.
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>    --
>
>
>
>
>
>
>
>              Terry D. Dontje | Principal
>                Software Engineer
>                Developer
>                    Tools
>                    Engineering | +1.781.442.2631
>
>                  Oracle
>
>                      - Performance
>                      Technologies
>
>                    95 Network Drive, Burlington, MA 01803
>                    Email terry.dontje@oracle.com
>
>
>
>
>
>
>
>
>
>
>
>
>_______________________________________________
>devel mailing list
>devel@open-mpi.org
>http://www.open-mpi.org/mailman/listinfo.cgi/devel


--
 Brian W. Barrett
 Dept. 1423: Scalable System Software
 Sandia National Laboratories






_______________________________________________
devel mailing list
devel@open-mpi.org
http://www.open-mpi.org/mailman/listinfo.cgi/devel

_______________________________________________
devel mailing list
devel@open-mpi.org
http://www.open-mpi.org/mailman/listinfo.cgi/devel