Skip to content

Conversation

jfeist
Copy link
Contributor

@jfeist jfeist commented Oct 5, 2021

This uses CUDA.with_workspace if it exists (CUDA 3.3 and later), and defines a minimal wrapper that uses CUDA.@workspace if not. I checked that tests pass on CUDA 2.6, 3.2, 3.3, and 3.4.

Fixes #107

@Jutho
Copy link
Member

Jutho commented Oct 5, 2021

Thanks for tackling this. Looks great.

@Jutho Jutho merged commit 45e9bf5 into QuantumKitHub:master Oct 5, 2021
@jfeist jfeist deleted the cuda3.3 branch October 5, 2021 12:37
@jfeist jfeist restored the cuda3.3 branch October 5, 2021 12:38
@jfeist jfeist deleted the cuda3.3 branch October 5, 2021 13:51
@alhirzel
Copy link

alhirzel commented Dec 28, 2021

@Jutho - any chance this PR could be incorporated into a release of TensorOperations.jl? Thank you for your work, and making these tools so smooth and easy to use!

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.

TensorOperations and CUDA 3.3, 3.4
3 participants