I see.  Let me try that... I'm looking at example 4, which does not involve
reading in a mesh. Is there a way I can parallelize even the read-in?

Thanks,
Karen

On Tue, Mar 9, 2010 at 5:50 PM, Roy Stogner <[email protected]>wrote:

>
>
> On Tue, 9 Mar 2010, Karen Lee wrote:
>
>  Thanks! Looks like I was just missing some entries of my MeshData since I
>> was combining 2 node files...
>>
>> I'm in gdb mode and it has taken 50 min and it seems that the mesh is
>> still not completely read in yet. Got stalled
>> here:
>> (gdb) n
>> found the tetgen files to read
>> read in nodes and elements
>> n
>> n
>> n
>> n
>>
>> after I pressed n to see if it was actually already done. Is this normal?
>> (1.7M points and 11M elements...
>>
>
> Probably.  METHOD=dbg turns on a ton of debugging tests, including the
> GLIBCXX tests, and last time I checked one of those tests turned some
> basic std::set operations from O(log(N)) to O(N log(N)).  This is
> obviously a problem when N is in the millions...
>
> Try METHOD=devel, if you're trying to debug problems that only crop up
> on large runs.  You won't get the vector bounds-checking, which is a
> shame, but you'll at least still be running all the libmesh_assert()
> tests, and it'll run much faster.
> ---
> Roy
>
------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Libmesh-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-users

Reply via email to