On Thu, 2009-01-01 at 14:08 -0500, DJ Delorie wrote: > > One interesting experiment I did was to pull out _all_ the drawing > > code, and measure PCB's "frame" rate. > > Can the batch HID do this for us? Or did you short-circuit the > gtk/lesstif hid?
I can't remember exactly how I did it. One vague recollection was that I spent some time chasing what I thought were nasty bottle-necks, but in fact the drawing code wasn't working. It may have been that I was calling the GL code, but hadn't set the context up right, so a lot of operations just returned NOP. I spent some time optimising things - which you saw as changes in the colour cache's hash algorithm, fixing the LRU colour heuristic etc. It might also have been cairo I was using.. before I coded GL support, I tried cairo, but it was too slow to be usable. I made an attempt to use the "glitz" backend for cairo, (which uses GL acceleration with a backend underneath the cairo drawing API), but I couldn't make that work - I was hitting fallback paths rather than fast-paths. -- Peter Clifton Electrical Engineering Division, Engineering Department, University of Cambridge, 9, JJ Thomson Avenue, Cambridge CB3 0FA Tel: +44 (0)7729 980173 - (No signal in the lab!) _______________________________________________ geda-dev mailing list geda-dev@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-dev
