Skip to content

Conversation

jaeyoo
Copy link
Member

@jaeyoo jaeyoo commented May 4, 2022

This is the first draft version which worked for the google internal infrastructure.

We need to modify this code as the following steps:

  1. Update & polish the code working for the external environment. (e.g. fix build error, naming, remove unused codes and commented code lines etc.)
  2. Update the gpu_benchmark.py file to see the better result on the external gpu environment.
  3. Add qsim cuQuantum version gpu op for expectation op.

@jaeyoo
Copy link
Member Author

jaeyoo commented May 4, 2022

@Sinestro38 Please take this PR for your implementation

@MichaelBroughton
Copy link
Collaborator

This looks awesome @jaeyoo ! Can we test this on github actions with GPU or add a kokoro job to do our GPU testing ?

@jaeyoo
Copy link
Member Author

jaeyoo commented May 19, 2022

Sure Michael! Thank you so much :) I am working on designing Kokoro-based GPU CI testing infra. My ETA is early June. Let's sync it together at our next meeting.

@mhucka mhucka added area/gpu Involves interfaces to GPUs area/dependencies Involves libraries or other software that TFQ depends on labels Dec 2, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
area/dependencies Involves libraries or other software that TFQ depends on area/gpu Involves interfaces to GPUs
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants