Skip to content

Conversation

@Lestropie
Copy link
Member

Modification resulting from my own needing to be able to catch eddy errors and modify parameters accordingly within an automated pipeline; but likely beneficial for users executing the script directly as well.

Basically, if the CUDA version of eddy fails, the resulting error message gets essentially hidden, because the exception is caught, the OpenMP version is run, and the error message produced by that version may give no insight into why the CUDA version failed.

With this change, if both the CUDA and OpenMP versions of eddy fail, the resulting error messages are concatenated into a single exception. When that is raised, the user (/ wrapping script) is provided with the terminal outputs generated by both versions of the command, enabling proper diagnosis.

Lestropie added 2 commits June 9, 2020 16:48
In cases where both cuda and openmp versions of eddy are available, the reason for the cuda version failing may be obscured because it is only the terminal output produced by the openmp version that is reported to the user. With this change, the messages produced by both cuda and openmp eddy commands will be concatenated into a new combined exception.
@Lestropie Lestropie self-assigned this Jun 10, 2020
@Lestropie Lestropie requested a review from a team June 10, 2020 02:18
Copy link
Member

@jdtournier jdtournier left a comment

Choose a reason for hiding this comment

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

Seems OK to me.

@jdtournier jdtournier merged commit 59e7e13 into dev Jun 10, 2020
@jdtournier jdtournier deleted the dwifslpreproc_eddy_error branch June 10, 2020 11:43
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.

3 participants