# PyFR 2.0.2 Update question about graph feature

**URL:** https://pyfr.discourse.group/t/pyfr-2-0-2-update-question-about-graph-feature/1144
**Category:** Development
**Created:** [7 August 2024 10:26 UTC](https://pyfr.discourse.group/t/pyfr-2-0-2-update-question-about-graph-feature/1144 "2024-08-07T10:26:40Z")
**Posts on this page:** 2
**Page:** 1

<div class="post-metadata">

### Author: ![luli](https://avatars.discourse-cdn.com/v4/letter/l/5f9b8f/32.png) [@luli](https://pyfr.discourse.group/u/luli)
#### Post date: [7 August 2024 10:26 UTC](https://pyfr.discourse.group/t/pyfr-2-0-2-update-question-about-graph-feature/1144/1 "2024-08-07T10:26:40Z")

</div>

“Upon upgrading to PyFR version 2.0.2, I observed the integration of “cudagraph” into the CUDA backend, which is a significant enhancement. However, I noticed that this feature does not appear to have been implemented for the HIP backend. Could you please clarify the rationale behind this discrepancy?”

---

<div class="post-metadata">

### Author: ![fdw](https://avatars.discourse-cdn.com/v4/letter/f/a587f6/32.png) [@fdw](https://pyfr.discourse.group/u/fdw)
#### Post date: [7 August 2024 11:44 UTC](https://pyfr.discourse.group/t/pyfr-2-0-2-update-question-about-graph-feature/1144/2 "2024-08-07T11:44:27Z")

</div>

See:

> <https://github.com/PyFR/PyFR/pull/312>
>
> This PR adds full support for HIP graphs, bringing the HIP backend on par with t…he CUDA backend. However, whereas for simple test cases overhead decreased by a factor of ~2 when moving to graphs, for HIP overhead increases by ~2. The performance impact for non-trivial cases still needs to be assessed, however.

Regards. Freddie.