Closed zero9178 closed 4 months ago
Up until now all wait_for_dma_transfers ops did not only wait for the transfers denoted by the tokens to be completed, but all transfers within the DMA. This blocks implementing transformations such as pipelining and is generally very unoptimized.
wait_for_dma_transfers
Up until now all
wait_for_dma_transfersops did not only wait for the transfers denoted by the tokens to be completed, but all transfers within the DMA. This blocks implementing transformations such as pipelining and is generally very unoptimized.