Versio 611 problem

Log Out | Topics | Search Moderators | Register | Edit Profile

Author Message   Archie Campbell (amc1) Member Username: amc1 Post Number: 9 Registered: 03-2008 Posted on Thursday, December 31, 2009 - 09:38 am:    Since upgrading from 610 to 611 some of my programmes have stopped working, giving the error message 'singular matrix'. The attatched is typical, it seems to only affect time dependent problems. Is there a quick fix? (I think I may have sent this somewhere else earlier) Flux Penetration fieldup.pde (2.7 k)   Patricio A. Greco (pagreco) Member Username: pagreco Post Number: 13 Registered: 12-2004 Posted on Thursday, December 31, 2009 - 10:10 am:    I chaged the conductivity at vacuum bound and itīs works I set a very very small value. Let me show you how it looks in my snow lepoard OS. FLexPDE611.tiff (151.2 k) untitled.pde (2.6 k)   Patricio A. Greco (pagreco) Member Username: pagreco Post Number: 14 Registered: 12-2004 Posted on Thursday, December 31, 2009 - 10:11 am:    I changed the file format...   Patricio A. Greco (pagreco) Member Username: pagreco Post Number: 15 Registered: 12-2004 Posted on Thursday, December 31, 2009 - 10:12 am:    Excuse me for the missed characters ... this is FLEX PDE on Snow leopard....   Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 1316 Registered: 06-2003 Posted on Thursday, December 31, 2009 - 04:46 pm:    Actually, I get the "singular matrix" diagnostic on 6.10 as well (Windows). The problem stems from some recent changes we made, trying to make the internal handling of vectors more consistent. This has some repercussions on the old treatment of scalars as z-components of vectors when used in Curl operators in 2d. We will fix it in 6.12, but in the meantime you can work around it by using CURL(CURL(0,0,A)), which explicitly identifies A as the out-of-plane component of a vector. The Snow Leopard graphic problem is another issue that we are trying to identify. Something changed in Apple's graphics.