Remove in-place lowering. #2055
Merged
Conversation
This file contains 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
This is one of our oldest optimisations, and mainly served to clean up the code produced by sequentialisation, as well as a few other simple things that we would today call "memory short circuiting". In-place lowering is a really complicated optimisation, so now that we have a propert short circuiting pass, we should seriously consider getting rid of it. Short circuiting is even more complicated, but it is also properly principled, and more general. We still largely have the sequentialisation problems that in-place lowering was supposed to address (mostly nested maps), but at least on the GPU backends there seems to be no real impact from it anymore. I would still like to one day fix sequentialisation so that it produces better code for nested SOACs. The big open question is how this impacts the performance of CPU code.
athas
force-pushed
the
no-in-place-lowering
branch
from
20f2e75
to
5aa2b7a
Compare
|
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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 is one of our oldest optimisations, and mainly served to clean up the code produced by sequentialisation, as well as a few other simple things that we would today call "memory short circuiting".
In-place lowering is a really complicated optimisation, so now that we have a propert short circuiting pass, we should seriously consider getting rid of it. Short circuiting is even more complicated, but it is also properly principled, and more general.
We still largely have the sequentialisation problems that in-place lowering was supposed to address (mostly nested maps), but at least on the GPU backends there seems to be no real impact from it anymore. I would still like to one day fix sequentialisation so that it produces better code for nested SOACs.
The big open question is how this impacts the performance of CPU code.