On Wed, Oct 19, 2011 at 2:27 PM, Roy Stogner <[email protected]>wrote:

>
> On Wed, 19 Oct 2011, David Fuentes wrote:
>
> > I probably need my version of libmesh, but
> > Is there an example that uses threading for Matrix assembly for FemSystem
> ?
>
> Not at the moment.  I made the DiffContext change to make the
> FEMSystem API thread-safe as soon as I noticed that issue, but I
> haven't actually added the threading yet; since PECOS is using a
> pre-FEMSystem code it hasn't been a high priority here.
> Multithreading FEMSystem would actually be pretty trivial, but it
> wouldn't be safe unless we were doing some better regression testing
> of it and it wouldn't be worthwhile unless we were doing some
> performance testing of it.
>
> Do you guys want to take care of that end of things?  I'll add the
> threading to FEMSystem::assembly/assemble_qoi*/postprocess if you'll
> do a before/after test on the resulting assembled matrix.  Different
> order of operations would mean we can't get an exact 0 diff, but in
> theory the only differences should be on the order of floating point
> epsilon as long as there's not too much cancellation between the
> overlapping terms of different element matrices.


I have up and running tests using FEMSystem for a variety of problems that
can also be used as guinea pigs. Let me know how I can help here. If it's
trivial, might be a nice stepping stone into TBB for me... ;-)
------------------------------------------------------------------------------
The demand for IT networking professionals continues to grow, and the
demand for specialized networking skills is growing even more rapidly.
Take a complimentary Learning@Ciosco Self-Assessment and learn 
about Cisco certifications, training, and career opportunities. 
http://p.sf.net/sfu/cisco-dev2dev
_______________________________________________
Libmesh-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-users

Reply via email to