home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!cis.ohio-state.edu!zaphod.mps.ohio-state.edu!cs.utexas.edu!ut-emx!ibmchs!auschs!awdprime.austin.ibm.com!scott
- From: scott@austin.ibm.com
- Newsgroups: comp.unix.aix
- Subject: Re: Virtual Memory curiosity
- Keywords: VM overflow response
- Message-ID: <1992Nov21.000017.16741@awdprime.austin.ibm.com>
- Date: 21 Nov 92 00:00:17 GMT
- References: <BxM5C5.Jyw@csn.org> <1992Nov13.034308.8602@oucsace.cs.ohiou.edu> <30381@nntp_server.ems.cdc.com> <1e530vINNrcn@network.ucsd.edu>
- Sender: news@awdprime.austin.ibm.com (USENET News)
- Distribution: usa
- Organization: IBM Austin
- Lines: 29
- Originator: scott@porter.austin.ibm.com
-
-
- In article <1e530vINNrcn@network.ucsd.edu>, dkf@chem.ucsd.edu (David Knight French) writes:
- > The question I have hear, "Is it one better?" Seems to me you would
- > have a lot more overhead using this scheme. The OS must now page fault
- > on reaads and writes if what it wants is not in memory. Maybe I am
- > reading this wrong, but some stats I have gotten from sar point to this
- > being the case. When the system was doing reads/writes, the number
- > of context switches jump dramatically. Again I don't know AIX internals,
- > but the overhead here does seem dramatic enough to suggest that the
- > kernel is page faulting on large reads and writes.
- >
- > I am very interested in what is really happening here. If anyone
- > knows for sure or could point me at the proper information I could
- > read, I would greatly appriciate it.
-
- There is an article titled "Evolution of Storage Facilities in AIX Version 3"
- in the IBM RISC System/6000 Technology book (order number SA23-2619) which
- should answer some of your questions. This same article also appeared in
- the IBM Journal of Research and Developement Volume 34, Number 1, Jan. 1990
- which is probably more accessible to you.
-
-
- Scott L. Porter IBM AWS Austin / AIX Kernel Development
- Internet: scott@austin.ibm.com
- --
- Scott L. Porter IBM AWS Austin / AIX Kernel Development
- Phone: (512) 838-3319 T/L: 678-3319
- "Beam me up, Scotty" VNET: PORTER at AUSTIN
- Internet: scott@austin.ibm.com
-