[PATCH] D32124: [BPI] Move tail computation out of the loop. NFC
Serguei Katkov via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Thu Apr 20 21:17:57 PDT 2017
skatkov added a comment.
After reading of the code a bit more I would say that in general no one worries about sum of probabilities is equal to 1. They follow it when operating on weights. But the trouble is that after weights are transformed to probabilities there are rounding errors which breaks the restriction.
https://reviews.llvm.org/D32124
More information about the llvm-commits
mailing list