tqchen commented on PR #18877: URL: https://github.com/apache/tvm/pull/18877#issuecomment-4008898599
Thanks @mitiskuma did a round of review. I am curious the benefit of StageBufferPool in this case. I know that it saves the blit from CPU to GPU and allows these blits to be fused into the compute. What happens if we just fuse the compute kernels and leave out the copy ones? -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
