Speed up line merging in INTER_AREA #24412
Merged
Conversation
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
|
Running: Here are the results ("real" INTER_AREA is only used for non-integer scale division): |
|
My results for AMD Ryzen 7 2700X: |
|
Less obvious result for Jetson TK1 (Armv7 with NEON): |
asmorkalov
requested changes
Oct 17, 2023
There was a problem hiding this comment.
OpenCV team migrates from static-size vector code to scalable vector code to support RISC-V RVV and other platforms. Could you accommodate to it? Example: #24166
|
Thx, I added taskset -c 33 python3 .//modules/ts/misc/run.py ...... |
mshabunin
reviewed
Oct 17, 2023
asmorkalov
reviewed
Oct 18, 2023
modules/imgproc/src/resize.cpp
Outdated
Comment on lines
3025
to
3028
| const v_int32 tmp0 = v_round(vx_load(src + 0 * v_float32::nlanes)); | ||
| const v_int32 tmp1 = v_round(vx_load(src + 1 * v_float32::nlanes)); | ||
| const v_int32 tmp2 = v_round(vx_load(src + 2 * v_float32::nlanes)); | ||
| const v_int32 tmp3 = v_round(vx_load(src + 3 * v_float32::nlanes)); |
There was a problem hiding this comment.
v_float32::nlanes -> VTraits<float32>::vlanes()
There was a problem hiding this comment.
Thx, done. BTW, it could be nice to change that API to make it constexpr.
|
@mshabunin , it seems the CI is hanging for RISC-V. |
|
@vrabaud, thank you for the patch! I wonder why in several places of the patch double-precision floating-point arithmetics is used? Isn't FP32 enough for this algorithm? (unless resize is applied to FP64 images). I mean, it's fine to use double precision to construct the interpolation tables, but when we do actual interpolation and accumulation, FP32 should probably be enough, right? |
|
@vpisarev , indeed float32 is used for all integer types and float32. Double is only used for doubles: opencv/modules/imgproc/src/resize.cpp Line 3797 in 2f1d529 This is the current behavior so I kept it. BTW, any idea as to why int and schar are disabled? |
|
RSIC-V builds are broken in CI now. Manual build error messages: |
|
@vrabaud you need to use v_add, v_mul and other functions instead of overloaded +, * and other operators. |
|
Thx @asmorkalov , I believe I fixed it using |
asmorkalov
approved these changes
Oct 19, 2023
|
As summary: |
|
0.1 and 0.25 mean integer proportions (a tenth and 4th of the original image) whose dimensions are divisible by 10 or 4. In that case, a different algorithm is used (areafast) that this pull request does not speed up. It is normal to get no change there. 0.81 is a non-even scale, normal INTER_AREA is used there and that's where you can see the speed-up. |
6 tasks
Merged
IskXCr
pushed a commit
to Haosonn/opencv
that referenced
this pull request
Dec 20, 2023
Speed up line merging in INTER_AREA opencv#24412 This provides a 10 to 20% speed-up. Related perf test fix: opencv#24417 This is a split of opencv#23525 that will be updated to only deal with column merging. ### Pull Request Readiness Checklist See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request - [x] I agree to contribute to the project under Apache 2 License. - [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV - [x] The PR is proposed to the proper branch - [x] There is a reference to the original bug report and related work - [x] There is accuracy test, performance test and test data in opencv_extra repository, if applicable Patch to opencv_extra has the same branch name. - [x] The feature is well documented and sample code can be built with the project CMake
thewoz
pushed a commit
to thewoz/opencv
that referenced
this pull request
Jan 4, 2024
Speed up line merging in INTER_AREA opencv#24412 This provides a 10 to 20% speed-up. Related perf test fix: opencv#24417 This is a split of opencv#23525 that will be updated to only deal with column merging. ### Pull Request Readiness Checklist See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request - [x] I agree to contribute to the project under Apache 2 License. - [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV - [x] The PR is proposed to the proper branch - [x] There is a reference to the original bug report and related work - [x] There is accuracy test, performance test and test data in opencv_extra repository, if applicable Patch to opencv_extra has the same branch name. - [x] The feature is well documented and sample code can be built with the project CMake
thewoz
pushed a commit
to thewoz/opencv
that referenced
this pull request
May 29, 2024
Speed up line merging in INTER_AREA opencv#24412 This provides a 10 to 20% speed-up. Related perf test fix: opencv#24417 This is a split of opencv#23525 that will be updated to only deal with column merging. ### Pull Request Readiness Checklist See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request - [x] I agree to contribute to the project under Apache 2 License. - [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV - [x] The PR is proposed to the proper branch - [x] There is a reference to the original bug report and related work - [x] There is accuracy test, performance test and test data in opencv_extra repository, if applicable Patch to opencv_extra has the same branch name. - [x] The feature is well documented and sample code can be built with the project CMake
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 provides a 10 to 20% speed-up.
Related perf test fix: #24417
This is a split of #23525 that will be updated to only deal with column merging.
Pull Request Readiness Checklist
See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request
Patch to opencv_extra has the same branch name.