[ENH]: improves heuristic in default implementation of GridIndex._chunk_io
#5191
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
mabruzzo
force-pushed
the
GridIndex-chunk_io
branch
from
d3e7ca2 to
f31d348
Compare
This commit improves the heuristic used by `GridIndex`’s default implementation of the `_chunk_io` method (with the `"auto"` chunk-sizing strategy) for determining how many grids to read in a given iteration. For context, the heuristic was historically hardcoded to a value of 1000 grids. This works well for AMR simulations with small grids (e.g. 16^3, 32^3), but the heuristic is problematic when you have unified skmulations This commit adopts a heuristic that tries to limit the number of grids in order to make sure we don't run out of memory.
mabruzzo
force-pushed
the
GridIndex-chunk_io
branch
from
f31d348 to
5110b82
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.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.
PR Summary
This commit improves the heuristic used by
GridIndex’s default implementation of the_chunk_iomethod (with the"auto"chunk-sizing strategy) for determining how many grids to read in a given iteration.For context, the heuristic was historically hardcoded to a value of 1000 grids. This works well for AMR simulations with small grids (e.g. 16^3, 32^3), but the heuristic is problematic when you have unigrid simulations composed of large grids. For example, Cholla commonly has grids of size 256^3.1
This commit adopts a heuristic that tries to limit the number of grids in order to make sure we don't run out of memory.
Footnotes
With the benefit of hindsight, some difficulties I encountered a few years back with some massive unigrid Enzo-E simulations now make a lot more sense. These improvements made here would have helped out with those difficulties ↩