RE: [Libmesh-devel] Optimizations

[KIRK, B. (JSC-E. (NASA) <ben...@na...>]

I started, but then I realized it is going to take more effort than I
thought and I want to make sure I do it right...  

WRT to the affine checking, this brings up an interesting issue I've been
thinking about in general.

template <class Traits=DefaultTraits>
class Elem<Traits>
{
  virtual bool has_affine_map () { return Traits::has_affine_map; }
};

...


This could be useful for a whole bunch of reasons...  Of course,
DefaultTraits::has_affine_map is false,  but the build_cube method could
force all elements to use an affine map by replacing

Elem::build(...) 

with 

Elem<Affine>::build(...)

sure, there are other implications, but I thought I'd bring the idea up.

-Ben



-----Original Message-----
From: lib...@li...
[mailto:lib...@li...] On Behalf Of Roy Stogner
Sent: Thursday, June 30, 2005 1:17 PM
To: lib...@li...
Subject: [Libmesh-devel] Optimizations



I've added affine map code paths to remove redundant derivative calculations
from compute_maps and init_shape_functions - it seems to be slightly faster,
but "slightly" means "around 2% on cubic TRI6 elements".  Since checking for
affine maps costs more for elements with more nodes and saves less for
elements with fewer quadrature points, it's entirely possible that I've just
slowed down calculations on the popular quadratic HEX27s.  Let me know if
anyone sees a slowdown - there are ways to speed up the has_affine_map
calculations that I'd like to try if necessary before reverting anything.

Ben, have you started work on the FE caching yet?  I'd offer to help, but
I'd be too tempted to saddle you with a vector of vector of multiarrays
instead of the hash table you wanted... 
---
Roy


-------------------------------------------------------
SF.Net email is sponsored by: Discover Easy Linux Migration Strategies from
IBM. Find simple to follow Roadmaps, straightforward articles, informative
Webcasts and more! Get everything you need to get up to speed, fast.
http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
_______________________________________________
Libmesh-devel mailing list
Lib...@li...
https://lists.sourceforge.net/lists/listinfo/libmesh-devel