Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!wuarchive!zaphod.mps.ohio-state.edu!rpi!uupsi!sunic!ugle.unit.no!nuug!ulrik!news
From: me@spirea.UiB.NO (My Account)
Newsgroups: comp.sys.next
Subject: Re: RAM and performance
Message-ID: <1991Apr22.135539.2058@ulrik.uio.no>
Date: 22 Apr 91 13:55:39 GMT
Sender: news@ulrik.uio.no (USENET News System)
Organization: University of Bergen, Norway.
Lines: 28
Nntp-Posting-Host: spirea.uib.no

>I have a mono 040 cube with 8MB:  I'm thinking about getting more RAM to speed
>the system up (it seems to swap a fair bit).  
>
>Can folks report their performance experiences with RAM upgrades in the 
>NeXT - ie, is the performance difference between 8MB and 16MB RAM very 
>noticeable/profound?  16MB->32MB?  
>
>What's the recommendation on how much RAM you "should" have in a basic 
>single-user cube for high-performance?  (Most of what I'm doing is pretty 
>vanilla, though I like to keep a lot of stuff active.)
>
>--
>Jim Black  (jjb@u.washington.edu)

Cosider 16MB a realistic minimum on a NeXT.  

I have a 040 cube at home (optical + accelerator drive) , with 2.0 and 16MB RAM.  At school I use a 030 cube under 1.0a with 16 MB RAM, and a 1.2 Gig disk.  I notice the difference - the cube at home is definately faster.  And with 16MB I avoid all those ugly swapping noises (hey gang, if I have a swap disk how come the optical is always causing a racket...?).  With 8MB things take much longer to load, run, and switch between.  

I scored 8MB for about $200 so imagine its worth it performance wise.  The 040's gotta have room to breath!  Watch out for parity and non parity mixing - I originally thought my upgrade board was bad (Exception 5 0xa something or other) but it was the mixed memory.  I've swapped the original NeXT (non parity) memory with some friends who still have PC's and don't need parity checking.  You could also buy regular non parity SIMMs (cheaper too!) and avoid the hassle of a mixed memory board.
 
 I can't say anything about 32MB (not yet anyway ;-) but I would expect a linear improvement over 16 (i.e 32 is to 16 as 16 is to 8).  If I'm wrong I'm confident someone who knows better will point out my error.  
 
 
--
Thor-Lee Legvold                 |         ppstl@nobergen.earn (or .bitnet)
University of Bergen             |         ppstl@cc.uib.no
NORWAY                                |        and now on NeXT...
Hvor mye for datter'n din...? |        me@fiol.uib.no