dmmartindale@watcgl.UUCP (Dave Martindale) (03/29/84)
Yes, DEUNAs work well under 4.2 once you get a driver from somewhere. We are running one here which is based on one written by Lou Salkind at NYU. My only complaint about the DEUNA is that it takes 16 amps of +5V current, which is a very substantial chunk of the power available in a single Unibus backplane in a BA11K box. A new subject: Has anyone measured actual DMA transfer rates on the existing controllers? It seems that the DEUNA transfers data from its own memory to the VAX at about 4 microseconds per word, while transferring data in the other direction at about 10 microseconds per word (if I remember the two figures correctly). It issues a new NPR for each cycle, rather than holding the bus for more than one cycle. How do other controllers compare in transfer rate and intelligence of using the bus?