31 May
2006
31 May
'06
8:30 a.m.
Douglas Gregor wrote:
Yes, it would be. Perhaps you could work up a patch to improve the performance in this area?
Umm, what exactly is the stuff that was already in rev 1.9? http://boost.cvs.sourceforge.net/boost/boost/boost/graph/fruchterman_reingold.hpp?r1=1.9.2.1&r2=1.9 The log just says "Fix bug in enumerating grid-force pairs" ... Anyway, I have written something that still uses the for loops and for my test program is about 1% faster. Eliminating the for loops and writing all 4 cases by hand is, strangely enough, about 4% slower in my setup ... I have the code at home, so I'll send it this afternoon.