Leo Fang
Hi @brandon-b-miller @gmarkall this is not urgent but it'd be nice to discuss this in the next sync-up. We recently also hit the error with undefined reference to `cudaCGGetIntrinsicHandle` in...
Need to incorporate @emcastillo's tensor bridge. The main challenge will be setting up a CI pipeline for building/testing.
Merging my offline comment to here. My 2c: - If possible I’d like to make part (or all) of this PR rendered as Sphinx docs, for both internal and external...
Ah... but the tile kernel calling convention is considered an implementation detail atm... @haijieg @brycelelbach @aterrel time to revisit the tile ABI discussion 🙂
I would prefer that we do no checks. The source provider (ex: cupy or our own StridedLayout, should we choose to make it public) is responsible for ensuring proper strides....
Isn't this a duplicate of https://github.com/NVIDIA/cuda-python/issues/762 + https://github.com/NVIDIA/cuda-python/issues/716?
Works need to happen in both repos, so I am not sure what's the duplication?
Adding @rparolin to evaluate how the APIs would look like to make comptue-graphics interop easy and performant
I think OpenGL should be the primary target. John has linked to various discussions and works in the issue description that some of us contributed to in the past few...
Not re-exporting was done on purpose, because if a downstream consumer can accept DLPack already, why don't they just take the input arrays directly instead of trampolining through SMV?