This repository was archived by the owner on Apr 28, 2023. It is now read-only.
Autotuning improvements #501
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
nicolasvasilache
force-pushed
the
pr/autotuning-improvements
branch
from
45a4998 to
065c935
Compare
Using an unscaled fitness value for selection is problematic: -Outstanding individuals take over very quickly, this leads to premature convergence. -When fitness values are close together, very litle selection pressure is applied and selection is almost uniformly random. Having slightly better fitness does not improve an individual's survival chances. -Transposing the fitness function (e.g. adding a constant value) changes the selection probabilities even though the location of the optimum (and the "shape" of the fitness) remain unchanged. Scaling the fitness function helps ameliorate those issues. Sigma scaling is used: fitness' = max(fitness - (mean_fitness - 2 * std_fitness), 0)
Stochastic Universal Sampling is an improvement upon the roulette algorithm that was previously used
This is needed following the changes to the build system.
It is too slow to converge and does not yield perf gains
nicolasvasilache
force-pushed
the
pr/autotuning-improvements
branch
from
065c935 to
380c054
Compare
thetheodor
reviewed
Jun 12, 2018
| lastBestConf(confs[0]), | ||
| numGenerations(numGenerations), | ||
| maxPopulationSize(populationSize), | ||
| matingPoolSize(populationSize * 3), |
There was a problem hiding this comment.
In one of the subsequent commits I introduced a separate flag for the mating pool size (instead of hardcoding populationSize * 3).
|
Thank you for your pull request. We require contributors to sign our Contributor License Agreement, and yours has expired. Before we can review or merge your code, we need you to email [email protected] with your details so we can update your status. |
ftynse
approved these changes
Sep 6, 2018
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
4 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR isolates a small subset from @ttheodor's #453.
I took the commits that look independent, fix clear bugs or implement simple new behavior and do not touch autotuner-inl.h (which underwent some changes related to early exit recently and for which I need to look deeper).