Copybara import of the project:#8143
Closed
copybara-service[bot] wants to merge 1 commit intomasterfrom
Closed
Conversation
copybara-service
bot
force-pushed
the
test_738898954
branch
9 times, most recently
from
f22be42 to
2272333
Compare
-- 7a85278 by EJ Park <eunjpark@quicinc.com>: scripts to generate x32 transposec for HVX -- 1c50176 by EJ Park <eunjpark@quicinc.com>: [WIP] tranpose for x32, vector_width=32 -- faa8310 by EJ Park <eunjpark@quicinc.com>: Temp changes to print out -- 057264d by EJ Park <eunjpark@quicinc.com>: Some generated kernels are removed. -- 8238faf by EJ Park <eunjpark@quicinc.com>: scripts and hvx.c.in changed temporarily -- a255441 by EJ Park <eunjpark@quicinc.com>: Trying scalar implementation for tail part -- f078770 by EJ Park <eunjpark@quicinc.com>: Almost working version Some tail case is not working properly. e.g., 34x34 matrix: We do HVX transpose 32x32, then 32x2, 2x32 and 2x3 with scalar code. Output has slight mismatch for the last row. -- 0b73071 by EJ Park <eunjpark@quicinc.com>: Fixed the tail code Now applying input_offset and output_offset works both for HVX code and scalar tail code. -- 51c3e8b by EJ Park <eunjpark@quicinc.com>: Remove print statements in test -- f0e0eb7 by EJ Park <eunjpark@quicinc.com>: Revert unnecessary changes for this commit -- 95c6828 by EJ Park <eunjpark@quicinc.com>: Change to output, input pointer The code to reset output and input pointer will only be executed if we go to the next iteration. Otherwise, skip. -- 4f7a434 by EJ Park <eunjpark@quicinc.com>: Change the way to use include Rather than using <>, change to use "" to include xnnpack header files. -- 4d46eb6 by EJ Park <eunjpark@quicinc.com>: Add `prod_microkernels` in Cmake build Add kernels used in configs into cmake build for hexagon FUTURE_COPYBARA_INTEGRATE_REVIEW=#7395 from ejparkqc:x32-transposec 4d46eb6 PiperOrigin-RevId: 738898954
copybara-service
bot
force-pushed
the
test_738898954
branch
from
2272333 to
5ceb081
Compare
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
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.
Copybara import of the project:
--
7a85278 by EJ Park eunjpark@quicinc.com:
scripts to generate x32 transposec for HVX
--
1c50176 by EJ Park eunjpark@quicinc.com:
[WIP] tranpose for x32, vector_width=32
--
faa8310 by EJ Park eunjpark@quicinc.com:
Temp changes to print out
--
057264d by EJ Park eunjpark@quicinc.com:
Some generated kernels are removed.
--
8238faf by EJ Park eunjpark@quicinc.com:
scripts and hvx.c.in changed temporarily
--
a255441 by EJ Park eunjpark@quicinc.com:
Trying scalar implementation for tail part
--
f078770 by EJ Park eunjpark@quicinc.com:
Almost working version
Some tail case is not working properly.
e.g.,
34x34 matrix:
We do HVX transpose 32x32, then 32x2, 2x32 and 2x3 with scalar code.
Output has slight mismatch for the last row.
--
0b73071 by EJ Park eunjpark@quicinc.com:
Fixed the tail code
Now applying input_offset and output_offset works both for HVX code and scalar tail code.
--
51c3e8b by EJ Park eunjpark@quicinc.com:
Remove print statements in test
--
f0e0eb7 by EJ Park eunjpark@quicinc.com:
Revert unnecessary changes for this commit
--
95c6828 by EJ Park eunjpark@quicinc.com:
Change to output, input pointer
The code to reset output and input pointer will only be executed if we go to the next iteration. Otherwise, skip.
--
4f7a434 by EJ Park eunjpark@quicinc.com:
Change the way to use include
Rather than using <>, change to use "" to include xnnpack header files.
--
4d46eb6 by EJ Park eunjpark@quicinc.com:
Add
prod_microkernelsin Cmake buildAdd kernels used in configs into cmake build for hexagon
FUTURE_COPYBARA_INTEGRATE_REVIEW=#7395 from ejparkqc:x32-transposec 4d46eb6