Conversation
|
Your PR no longer requires formatting changes. Thank you for your contribution! |
lkdvos
left a comment
lkdvos
left a comment
There was a problem hiding this comment.
In principle looks good to me!
Do we want to wait with the AMD support until the dust settles on CUDA (mostly to avoid having to duplicate things we might still change), or should we just go ahead with this?
|
This doesn't include factorization stuff which is the only inflight CUDA thing, I think? Most of the diff is the tests, tbh |
|
The tests is actually what I was thinking of, but maybe it's really not that bad |
lkdvos
left a comment
lkdvos
left a comment
There was a problem hiding this comment.
Overall looks good to me, I would be happy to merge and gradually improve
| @test ht2 == TensorKit.to_cpu(dt2) | ||
| end | ||
|
|
||
| dt3 = AMDGPU.@allowscalar repartition(t, k) |
There was a problem hiding this comment.
Are we tracking these @allowscalar calls somewhere? Technically this test is now not really testing whether or not it works :p
There was a problem hiding this comment.
At least in the tests we can just search in the file, I can make a tracker comment at the top?
|
Let me figure out where the segfaults are happening then I'm also ok to merge |
2 participants
Mostly copied from the CUDA support