On Sun, 28 Feb 2010, Roy Stogner wrote:

> On the other hand, none of the main libMesh developers use anything
> but double frequently,

Where by "frequently", I guess I meant "any time in the last few
years"...

> so we may have a regression there.  I'll see if I can replicate the
> problem.

We had lots of regressions there.  I wasn't able to test things as
well as I'd like (no non-double builds of PETSc or Trilinos lying
around) but I fixed things with Laspack at least.  Give the svn head a
try now.

One remaining problem: you might want to stick to .xda I/O.  The
binary I/O routines were generally written with Real==double in mind,
and I don't have time to fix those myself.
---
Roy

------------------------------------------------------------------------------
Download Intel® 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-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-devel

Reply via email to