On Oct 20, 2010, at 4:51 PM, Derek Gaston wrote: > On Oct 20, 2010, at 4:47 PM, Derek Gaston wrote: > >> On Oct 20, 2010, at 4:18 PM, Roy Stogner wrote: >> >>> Precisely. Using the default tolerance here (I think 1e-6 for >>> doubles?) would be a big no-no, but using 1e-15 or so ought to be >>> safe. >> >> With my particular setup 1e-15 was too restrictive... but it passes with >> 1e-14. I actually want to set it to 1e-13 though just to be safe. Does >> that seem too large to you? It is a "relative" fuzzy equals though... so at >> least it will scale with the mesh size (so if you have a small mesh the >> tolerance will be tighter right?). > > Check that... it passes at 4e-15.... so 1e-14 should probably be safe....
Check that again... after greatly increasing the number of elements 1e-14 failed... so my current recommendation is still 1e-13. Derek ------------------------------------------------------------------------------ Nokia and AT&T present the 2010 Calling All Innovators-North America contest Create new apps & games for the Nokia N8 for consumers in U.S. and Canada $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store http://p.sf.net/sfu/nokia-dev2dev _______________________________________________ Libmesh-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-devel
