Evaluating and comparing performance #374
ivanperez-keera
started this conversation in
General
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
The just released version of Yampa includes basic benchmarks. Lack of a proper mechanism to evaluate the performance of changes proposed to dunai has been one of the major obstacles to incorporating such changes.
I'd like to invite everyone to take a look at those benchmarks, and help determine 1) any flaws or limitations there may be in them, 2) whether similar benchmarks could be used for dunai and bearriver, 3) how the results could be compared across implementations (e.g., bearriver vs Yampa vs dunai), and 4) what other benchmarks could be added to the same battery, and what applications could be adapted so that they can be used as benchmarks.
Beta Was this translation helpful? Give feedback.
All reactions