Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Conversation

zihaomu
Copy link
Member

@zihaomu zihaomu commented Apr 19, 2023

Address #23465
Merge with test: opencv/opencv_extra#1058

Pull Request Readiness Checklist

See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request

  • I agree to contribute to the project under Apache 2 License.
  • 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
  • The PR is proposed to the proper branch
  • There is a reference to the original bug report and related work
  • There is accuracy test, performance test and test data in opencv_extra repository, if applicable
    Patch to opencv_extra has the same branch name.
  • The feature is well documented and sample code can be built with the project CMake

@zihaomu zihaomu linked an issue Apr 19, 2023 that may be closed by this pull request
@zihaomu zihaomu added this to the 4.8.0 milestone Apr 19, 2023
@asmorkalov asmorkalov requested a review from vpisarev April 19, 2023 13:14
@asmorkalov
Copy link
Contributor

The issue is still there. See buildbot:

[ RUN      ] Conv3D.conv3d/14, where GetParam() = (GFLOPS=0.000, K=[1 x 1 x 1], IN={1, 8, 1, 10, 10}, OCN=8, G=8, P=(1, 1) x (1, 1) x (1, 1), BIAS, OCV/CPU)
C:\build\precommit_windows64\4.x\opencv\modules\ts\src\ts_perf.cpp(1975): error: Failed
Expected: PerfTestBody() doesn't throw an exception.
  Actual: it throws...
params    = (GFLOPS=0.000, K=[1 x 1 x 1], IN={1, 8, 1, 10, 10}, OCN=8, G=8, P=(1, 1) x (1, 1) x (1, 1), BIAS, OCV/CPU)
termination reason:  unhandled exception
bytesIn   =          0
bytesOut  =          0
samples   =          0 of 1
outliers  =          0
frequency =          0
[  FAILED  ] Conv3D.conv3d/14, where GetParam() = (GFLOPS=0.000, K=[1 x 1 x 1], IN={1, 8, 1, 10, 10}, OCN=8, G=8, P=(1, 1) x (1, 1) x (1, 1), BIAS, OCV/CPU) (8 ms)

Link: https://pullrequest.opencv.org/buildbot/builders/precommit_windows64/builds/101372/steps/perf_dnn/logs/stdio

@zihaomu zihaomu force-pushed the issue_23465 branch 2 times, most recently from a8ae90b to f60d9a0 Compare April 20, 2023 08:18
@zihaomu
Copy link
Member Author

zihaomu commented Apr 20, 2023

Hi @asmorkalov, I have added a compute branch for the conv3d depthwise (the crashed case) which was used in the previous code.

@asmorkalov asmorkalov self-requested a review April 20, 2023 09:25
@opencv-alalek
Copy link
Contributor

opencv-alalek commented Apr 20, 2023

@zihaomu Please take a look on valgrind reports - valgrind points on some problem places in dnn: https://pullrequest.opencv.org/buildbot/builders/4_x_valgrind-lin64-debug

@zihaomu
Copy link
Member Author

zihaomu commented Apr 20, 2023

Update: I found there is a bug at 1x1_conv3d-depthwise, and I just fixed it. This time everything should be fine.

Hi @opencv-alalek and @asmorkalov, thanks for your reminder, I will generate a test case in accuracy test to check this issue carefully.

@zihaomu
Copy link
Member Author

zihaomu commented Apr 21, 2023

Hi, @opencv-alalek, looks there is issue in built-bot CI.

@asmorkalov asmorkalov requested review from asmorkalov and opencv-alalek and removed request for asmorkalov April 21, 2023 12:30
Copy link
Contributor

@asmorkalov asmorkalov left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

👍 Thanks!

}

int MAX_STRIPES = (56 + CONV_NR - 1)/CONV_NR;
int MAX_STRIPES = conv->conv_type == CONV_TYPE_DEPTHWISE_REMAIN ? 1 : (56 + CONV_NR - 1)/CONV_NR;
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

(56 + CONV_NR - 1)/CONV_NR

BTW, there is divUp() call for that

@opencv-pushbot opencv-pushbot merged commit 9ab0ff6 into opencv:4.x Apr 22, 2023
@asmorkalov asmorkalov mentioned this pull request May 31, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

DNN: sporadic crashes in perf_dnn

5 participants