Cache swizzled tensor for tuning #1686
Open
+156
−29
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
Serge45
requested review from
jichangjichang,
KKyang,
vin-huang,
imcarsonliao,
hcman2,
Jinp800125,
TonyYHsieh and
solaslin
as code owners
jichangjichang
previously approved these changes
Feb 21, 2025
| //TODO: Support more swizzling type, such as 32x32x8, currently we have 16x16x8 only. | ||
| if(needSwizzle) | ||
|
|
||
| //if no validation, skip the swizzle |
There was a problem hiding this comment.
I understand we don't need to permute the tensor when no validation. But is it correct to skip the padding?
There was a problem hiding this comment.
After doing some experiments on swizzled and non-swizzled on STA problem, I observed that there's performance disparity. In the latest commit, it uses LRUCache to manage the cached tensor for balancing memory usage and runtime performance, and now it always performs swizzle for STA/STB problem.
Serge45
force-pushed
the
feature/tensilelite-cache-swizzle
branch
from
0120521 to
4948354
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Outdated suggestions cannot be applied.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.
Re-layout will be skipped if validation disabled.