add SLJIT_REV16PACK and SLJIT_REV32PACK #206
Draft
+223
−52
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
Including implementations without emulation for ARM, MIPS and LoongArch. Allow 16/32bit bswap in a packed integer, so it could provide faster results by doing 2 or 4 operations at once. While at it add an instruction to ARM 32bit to make things slightly more efficient and do some refactoring on related test cases.
|
Is there a use case for this? When I was thinking about this before, I decided not to do it until the support needed for something. For example, the normal rev is needed for big-endian WebAssembly support, since WebAssembly must use little endian memory accesses. |
no, was more interested on feedback on the design and code, as well as the potential performance improvements if bulk bswaps were needed as well as if it was worth doing it here or going all the way to vector instructions (most likely the later which seems to be well supported at least in s390x). |
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Outdated suggestions cannot be applied.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.
Might become useful when byte swapping in bulk and therefore as it is meant to be performance sensitive, not Including (at least for now) emulation.
Not particularly fond of he name, but at least keeps the bit width of the instruction away from the bit width of the register size, even if a mouthful as
SLJIT_REV16PACK32