tqchen commented on issue #15874: URL: https://github.com/apache/tvm/issues/15874#issuecomment-1749784280
Looks like indeed we didn't have proper `kDLCUDAManaged ` . I think it can be added in similar ways as https://github.com/apache/tvm/blob/main/src/runtime/cuda/cuda_device_api.cc#L256 And add allocation support. I don't have experience with `CUDAManaged` before, but if it is a matter of updating cudamalloc and copy we might be able to update support -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
