Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Assignment6] Meaning of error message about "Pipeline stall" in 'conv' integration test #78

Closed
ki6090 opened this issue Nov 21, 2024 · 8 comments
Assignees
Labels
homework - mesh-with-delays Assignment 6: Mesh-with-delays question Further information is requested

Comments

@ki6090
Copy link

ki6090 commented Nov 21, 2024

image

Hello, I already passed all unit test and matmul, matrix_add, raw_hazard integration test.

But only conv test fails...

Could I get some hints about the meaning of pipeline stall error message? ,,,

Other error message shows the different results and I solved with that messages, but this message cannot figure about my code errors...

Thanks for reading!

@JongyCysec
Copy link

JongyCysec commented Nov 21, 2024

Sadly, I encountered same problem too...

@woojinnn
Copy link

woojinnn commented Nov 21, 2024

  • After the Execute module (which is outer module of MeshWithDelays module) finishes its execution, it sends rob_id to reservation station module to notify that Execute module finished its execution. If Reservation station doesn't receive the fired rob_id more than 10000 cycles, the Reservation Station classifies it as stalled and throws an error.
  • Probably your implementation isn't sending ready signal properly, and it is blocking the execution of next command. As a hint, matmul, matrix_add, and Raw_hazard testcases send DIM*DIM (where DIM indicates the Mesh size) matrices to MeshDelays module, whilst the conv testcase sends partial inputs. (For instance, 5*5 matrices, etc...) How about checking your counter management logic etc...?

@ki6090
Copy link
Author

ki6090 commented Nov 21, 2024

Thanks for kind answer!!

@ki6090 ki6090 closed this as completed Nov 21, 2024
@minseongg
Copy link
Member

We will prepare the reference waveform for debugging purpose ASAP. Sorry for the inconvenience.

@JongyCysec
Copy link

@woojinnn
I've examined through sub-modules where stall can happen.
Considering the complexity of the design and hardness of debugging, it would have been quite hard for me to debug own myself without your sincere help.

Thanks to your explanation, I can solve the problem !!

@woojinnn
Copy link

woojinnn commented Nov 21, 2024

@JongyCysec Congrats!
image

@minseongg
Copy link
Member

minseongg commented Nov 21, 2024

We have added reference waveform files for integration tests (kaist-cp/hazardflow@53b99dd).

For more details, please refer to #71 (comment)!

@minseongg minseongg self-assigned this Nov 21, 2024
@minseongg minseongg added question Further information is requested homework - mesh-with-delays Assignment 6: Mesh-with-delays labels Nov 21, 2024
@ki6090
Copy link
Author

ki6090 commented Nov 21, 2024

Thank you!! I appreciate your help.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
homework - mesh-with-delays Assignment 6: Mesh-with-delays question Further information is requested
Projects
None yet
Development

No branches or pull requests

4 participants