May 12, 2010
view_performance branch has been merged into trunk!
A big thanks to Phoku and Helios for making this a reality. We have now removed all O(N) instance lookups from FIFE. We have heard reports of some people getting 50% + better performance when operating at a regular zoom level. Phoku did most of the work and Helios was able to find several pesky bugs that plagued the branch. Also thanks to all the people who took time to help us test everything. Your help/input was greatly appreciated!
Until next time…
hey there,
while you’re doing this performance review and such, have you taken a look at throwing parts of FIFE through the clang static analyser?
we’ve found it quite handy for finding code paths that were ignored or not thought to exist and tighten quite a few things up… and also a few notices about performance helped, specially guys that are newer to C, commiting. tho most of this wont apply as you guys use a bunch of C++.
just a thought