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

[DOC] Codepin intro update for 2025.0 #2401

Merged
merged 3 commits into from
Oct 11, 2024

Conversation

lapel22
Copy link
Contributor

@lapel22 lapel22 commented Oct 11, 2024

Updating Debug with Codepin intro and adding new image for 2025.0 per https://jira.devtools.intel.com/browse/DOC-12786 please review & approve & merge if it looks good.

The above picture shows the data flow graph of the vectorAdd example, which is constructed by a title and execution layer. The execution layer presents a kernel execution and its inputs and outputs. The kernel node shows that kernel ``vectorAdd`` is executed on the stream of the device named GPU0, and also shows the kernel's execution time and source location. All input arguments (``d_a`` node and top ``d_result`` node) are tagged with V0, indicating initial values. The output argument (bottom ``d_result`` node) is tagged with V1 because ``d_result`` is both input and output arguments, and its value changes in the kernel.

The nodes ``d_a:V0 ``, ``d_result:V0``, and ``d_result:V1`` are colored red, indicating a value mismatch between the CUDA and SYCL runs. In this case, the result value mismatch is caused by the mismatch of the input argument values, and the mismatch between input argument values may be caused by the different behavior of memory initialization between CUDA and SYCL, as the report states.
Copy link
Contributor

Choose a reason for hiding this comment

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

The ``d_a:V0 `` should be changed to ``d_a:V0``, the space before the last two backticks needs to be removed.

In the data flow graph, each kernel execution and its input and output arguments are grouped into a layer, presenting a run status of the kernel execution. The value of input and output arguments are tagged with version information in the form of “V<num>”. For example, the initial version is tagged as V0, and once the value of the argument is updated, the version number will be increased. For a specific kernel execution, if there’s a mismatch between CUDA and SYCL results, the mismatched argument node will be colored red.

.. figure:: DataFlowGraph.png
Copy link
Contributor

Choose a reason for hiding this comment

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

The path should be ../../_images/DataFlowGraph.png

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Updates made, please review & approve & merge if it looks ok. Thanks!

Copy link
Contributor

@intwanghao intwanghao left a comment

Choose a reason for hiding this comment

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

LGTM

@zhimingwang36 zhimingwang36 merged commit 9a91951 into oneapi-src:SYCLomatic Oct 11, 2024
5 of 6 checks passed
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.

3 participants