#14826: Remove misoptimizations from init code #14861
Open
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
nathan-TT
requested review from
aliuTT,
ubcheema,
abhullar-tt,
pgkeller,
tt-aho,
tt-dma and
tt-asaigal
as code owners
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
nathan-TT
force-pushed
the
nsidwell/trisc-14826
branch
from
6f79a1d
to
d5b4f64
Compare
|
I was dissatisfied with the volatile asms, so changed to actual assembly instructions, hiding the form of the loop from the compiler. All the git hub comments are unavoidable changes -- this is startup code, of course it's going to be 'exciting' |
1) Stop wzerorange being recognized as memset 2) Reduce insns in data image copy 3) Do not use a loop for residue Rename init code as do_crt1, to make it clearer what it is doing.
nathan-TT
force-pushed
the
nsidwell/trisc-14826
branch
from
d5b4f64
to
10c7e96
Compare
pgkeller
approved these changes
Nov 12, 2024
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Outdated suggestions cannot be applied.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.
Ticket
#14826
Problem description
Empty kernels are larger than expected.
What's changed
Stop wzerorange being recognized as memset. Memset is no longer pulled in.
Reduce insns in data image copy. Original loop was 21 isnsn (3.5 per word), new loop is 10 insns (3.3 per word).
Do not use a loop for residue. We only have to handle 0, 1 and 2 cases. A loop is more overhead.
Sprinkle a few more unroll-inhibiting pragmas around.
Rename init code as do_crt1, to make it clearer what it is doing.
These changes remove 436 bytes from a kernel code.
Checklist