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

[oneMKL samples][computed tomography] Revisions, refactoring and addition of a functional verification #2564

Open
wants to merge 3 commits into
base: development
Choose a base branch
from

Conversation

raphael-egan
Copy link
Contributor

Description

This PR revises the "computed tomography" oneAPI sample.

The sample itself was revised so that the application creates reconstructed signal/image samples that are real (no longer complex) and so that those generated real values are directly comparable to the input image's (gray-scale-converted) values themselves. These changes are motivated by the desire to introduce a legitimate functional validation step (introduced herein as well), but that need triggered more thorough revision(s) of the sample's source code.

Extensive explanatory comments were added to the source code, explaining the mathematical nature of the operations at play and/or the technical details related to the suggested implementation.

Type of change

  • Implement fixes for ONSAM Jiras

How Has This Been Tested?

This was extensively tested on a PVC GPU (512 EUs) and a SPR CPU using a machine operating Ubuntu 22.04. The default input file as well as others (not added to this PR) were considered along with an extensive range of input parameters. Due to the nature of the operations at play, one simply cannot guarantee that the mean global error will be below the arbitrarily-chosen threshold for any input file and any choice of (supported) input parameters. However,

  • the default operations were verified to pass the added tests by a large margin;
  • the application was augmented with explanatory messages advising how to tweak input parameter(s) to improve accuracy of the results, if found inaccurate for other input file and/or input parameters.

The updated sample's default behavior was also successfully tested on a BMG GPU machine operating Windows 11.

Regarding instructions to test or to run the sample with different inputs, please refer to the application's "Usage" information (printed in case of invalid usage).

@andreyfe1 andreyfe1 self-requested a review January 17, 2025 09:49
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant