davef@lakesys.UUCP (Dave Fenske) (09/06/89)
I'm not certain that I understand all of the rationale for the way in which SCO handles disk IO. For example, they avoid DMA in favor of program control. Does anyone have any information on this? Thanks. DF
neese@adaptex.UUCP (09/07/89)
Most disk controllers for the AT bus do not have provisions for DMA
usage. That is they don't even have the REQ/ACK or channel support
for the AT DMA. Even if the controller would support it, the AT DMA
is awfully slow when compared to PIO.
Roy Neese
Adaptec Central Field Applications Engineer
UUCP @ {texbell,killer}!cpe!adaptex!neese
merch!adaptex!neeseromkey@asylum.SF.CA.US (John Romkey) (09/07/89)
In article <1049@lakesys.UUCP> davef@lakesys.UUCP (Dave Fenske) writes: >I'm not certain that I understand all of the rationale for the way in which >SCO handles disk IO. For example, they avoid DMA in favor of program control. SCO avoids DMA because in the PC/AT architecture (and therefore the ensuing 386 clones built out of it) DMA is SLOW and a processor string instruction with a repeat prefix will improve performance tremendously over what the DMA controller will get you. - john romkey USENET/UUCP: romkey@asylum.sf.ca.us Internet: romkey@ftp.com "God is good, God is great, God's a big inverterbrate." - Boiled In Lead