home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.acorn
- Path: sparky!uunet!mcsun!ieunet!tcdcs!maths.tcd.ie!chughes
- From: chughes@maths.tcd.ie (Conrad Hughes)
- Subject: Re: OS differences and improvements (Was Re: new PC's, what's happening acorn?)
- Message-ID: <1992Aug27.090413.3631@maths.tcd.ie>
- Organization: Dept. of Maths, Trinity College, Dublin, Ireland.
- References: <9208250652.AA04072@ucbvax.Berkeley.EDU> <1992Aug25.134924.9601@merlin.dev.cdx.mot.com>
- Date: Thu, 27 Aug 1992 09:04:13 GMT
- Lines: 56
-
- lezz@merlin.dev.cdx.mot.com (Lezz Giles) writes:
-
- >Anybody who thinks that VM doesn't and cannot work should (a) try using
- >a decent computer, and (b) take a decent computer science course.
- MIPS Magnums & SPARCStations (~30 users on a busyish day) might satisfy
- you; I'm a mathematician, not a computer scientist though.
-
- >For
- >many years I worked on an Amdahl mainframe supporting around 200 users
- You mean the machines with 64 megabytes of SRAM cache? These
- "illusions" of speed you suffered from may not have been
- hallucinations :-)
-
- >I'm currently
- >on a DECstation - a single user workstation with X windows. I don't
- >notice the VM at all. I've also seen HP, Sun, etc. etc workstations with
- >VM and with the same illusion of no VM.
- How many processes were in VM? The Magnums here spend a lot of time
- with ~10Mb of virtual RAM, but this tends to be dead login shells
- and things lost in the mists of time. I have several times tried
- running an nth-order arithmetic coder on the machines (suffice to
- say at high orders it can easily chew its way through tens of
- megabytes of RAM) - when there were no other active users on the
- machines (maybe two idle for a couple of days), the load was
- 0.0something and all the memory was real. When it pushed past 30Mb
- of virtual memory (with the coder eating 40-50Mb) things were really
- thrashing quite badly. VM is very useful in that it enables me to
- run incredibly inefficient, memory-hungry programs with 80%
- redundant data structures on machines that wouldn't otherwise cope,
- _but_ I am under no illusion that it won't kill the machine while
- I'm doing it. BTW, the VM on these machines seems to dislike the
- concept of a process larger than physical RAM :-(
-
- >Virtual memory is a fact of life, and virtual memory works extremely well.
- At low levels of usage when all that's virtual is suspended or
- dead processes, fine. If it's virtual cos you're running one big
- yucky process it shows it up for what it is: memory with a 4ms
- access time and transfer rates of a mere few Mb per second..
-
- >Before anybody flames me back, make sure that you have satisfied both
- >conditions (a) and (b) above.
- This wasn't a flame and while I'm sure my "qualifications" will be
- totally unstatisfactory for your high standards (maybe it is a flame
- ;-}), I have tried making real _use_ of the VM on these systems with
- some quite horrible results. (I ran a MOO on the 16Mb SPARCStation 1
- for a while; (MOO is a variety of MUD) database dumps, during which
- the MOO ate up ~24Mb (two 12Mb processes) also seriously impacted
- on the machine, which was also multiuser - the MOO only existed
- outside office hours :-)
-
- Conrad
- --
- ,--------------------. Suicide City, Paranoid State. Geddout! trip to, heave'n'
- |chughes@maths.tcd.ie| ho, up, down, to'n'fro you have no word please leave us
- +-=>Conrad Hughes<=-+ here close our eyes to the octopus ride isn't it good to
- `-------<SICK>-------' be lost in the wood isn't it bad so quiet there . . .
-