home *** CD-ROM | disk | FTP | other *** search
/ NetNews Usenet Archive 1992 #30 / NN_1992_30.iso / spool / comp / arch / 11782 < prev    next >
Encoding:
Internet Message Format  |  1992-12-21  |  1.4 KB
  1. Xref: sparky comp.arch:11782 comp.parallel:2772
  2. Newsgroups: comp.arch,comp.parallel
  3. Path: sparky!uunet!usc!cs.utexas.edu!sdd.hp.com!ncr-sd!ncrcae!hubcap!fpst
  4. From: dfk@wildcat.dartmouth.edu (David Kotz)
  5. Subject: question about DMA devices
  6. Message-ID: <1992Dec18.175402.27841@hubcap.clemson.edu>
  7. Apparently-To: comp-parallel@rutgers.edu
  8. Followup-To: comp.arch
  9. Summary: can they do scatter/gather?
  10. Originator: dfk@wildcat.dartmouth.edu
  11. Keywords: DMA, scatter/gather
  12. Sender: news@dartmouth.edu (The News Manager)
  13. Organization: Dartmouth College, Hanover, NH
  14. Date: Fri, 18 Dec 1992 16:59:15 GMT
  15. Approved: parallel@hubcap.clemson.edu
  16. Lines: 21
  17.  
  18. I have a question about DMA devices, specifically for networks and disks.
  19.  
  20. Do you know if it is possible (and if so, common) for a DMA device to
  21. support scatter/gather? In particular, I'm trying to model a
  22. multiprocessor interconnect network interface, and it would be very
  23. helpful if I could avoid packetization by specifying the addresses and
  24. lengths of my message header and message body, for the network DMA to
  25. pick up. I'm guessing that this is not usually an option.
  26.  
  27. Also, in a disk interface, can it read in a whole track but scatter
  28. the blocks to different locations in memory? I'm guessing that this
  29. kind of thing is more common.
  30.  
  31. Thanks,
  32. dave
  33. -- 
  34. -----------------
  35. Mathematics and Computer Science
  36. Dartmouth College, 6188 Bradley Hall, Hanover NH 03755-3551
  37. email: David.Kotz@Dartmouth.edu    or   dfk@cs.dartmouth.edu
  38.  
  39.