You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi,
We are facing challenges in ensuring the compatibility and performance of our CUDA code across different GPUs. Given the significant variance in memory capacities and capabilities, it's challenging to predict the exact available memory usage for our code.
In light of this, we would like to inquire whether it would be possible for the contest organizers to provide access to the workstation, which has a setup similar to the one used for the alpha, beta, and final tests (four A100 GPUs).
Thank you!
The text was updated successfully, but these errors were encountered:
Thanks for the request! Sorry that it will be difficult for us to provide access to the workstation of testing environments. But participants could make good use of the alpha and beta submission opportunities to test their algorithms. We are also considering providing teams more opportunities to submit their solutions and run in our server. We will also release details of our testing environment soon!
Hi,
We are facing challenges in ensuring the compatibility and performance of our CUDA code across different GPUs. Given the significant variance in memory capacities and capabilities, it's challenging to predict the exact available memory usage for our code.
In light of this, we would like to inquire whether it would be possible for the contest organizers to provide access to the workstation, which has a setup similar to the one used for the alpha, beta, and final tests (four A100 GPUs).
Thank you!
The text was updated successfully, but these errors were encountered: