home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!news.larc.nasa.gov!news
- From: s.j.scotti@larc.nasa.gov (Stephen J. Scotti)
- Newsgroups: comp.sys.sgi.misc
- Subject: Re: Where does the CPU go, when its not doing nuthin' (gr_osview)
- Date: 31 Dec 1992 14:19:20 GMT
- Organization: Aircraft Structures Branch, NASA Langley Research Center
- Lines: 18
- Distribution: world
- Message-ID: <1huvh8INN1tf@rave.larc.nasa.gov>
- References: <1992Dec31.000257.19671@photon.com>
- Reply-To: s.j.scotti@larc.nasa.gov (Stephen J. Scotti)
- NNTP-Posting-Host: scotti.larc.nasa.gov
-
- |Your process was too big for physical memory, so it was doing a lot of
- |swapping to disk. Once you got enough memory you didn't need to swap
- |and the disk access went away (and the process worked a lot faster)
-
- I now see this point made by numerous respondents (sp?)
-
- Thanks to all who replied. My confusion was based on the fact that the
- code in question ran in only 4MB on a VAX, and I forgot the "true" meaning of
- "reduced instruction set" when I used the code on a 4d35. I tried the
- gr_osview command while running the code and, if I interpret it correctly,
- I need at least 18MB to run on the 4d35 (not counting modifed memory pages).
- Thanks again to all (esp. Scott Dorsey who brought the VAX vs. RISC aspect
- to light)
-
-
- Stephen J. Scotti
- ASB, NASA Langley Research Center
- s.j.scotti@larc.nasa.gov
-