Updates:
Status: Assigned
Comment #2 on issue 1750 by [email protected]: Improve performance of
mesh decompression benchmark
http://code.google.com/p/v8/issues/detail?id=1750
I found two potential problems why V8 is slow here:
1. performance of String.charCodeAt: we have optimizations for string
functions. I'm going to figure out why we don't hit them here.
2. for-in statement in the function decompressSimpleMesh: Crankshaft can't
deal with for-in loops. It would be however a bigger change to allow for-in
for the optimizing compiler. So if it is easy, I recommend avoiding for-in.
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
