nverke commented on PR #13613: URL: https://github.com/apache/tvm/pull/13613#issuecomment-1353742695
We are still discussing the pros and cons of the potential options but the possible solutions are. 1. Lower the copy into n x 1D dma copies. 2. Lower to 2D dma copies. 3. Use tensorize to make synchronous dma copies then in the lower_async_dma pass add the necessary asynchronous pieces. -- 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]
