Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!ucsd!sdd.hp.com!uakari.primate.wisc.edu!samsung!dali!milton!oregon!cs.uoregon.edu!comix!akm
From: akm@comix.cs.uoregon.edu (Anant Kartik Mithal)
Newsgroups: comp.sys.ibm.pc
Subject: Re: 386 Memory question (simms)
Summary: 16bits = 2 simms?? not 4??
Keywords: simms 386
Message-ID: <1990May31.010851.20237@cs.uoregon.edu>
Date: 31 May 90 01:08:51 GMT
Expires: 6/7/90
References: <5923@buengc.BU.EDU> <4922@plains.UUCP>
Sender: usenet@cs.uoregon.edu (Netnews Owner)
Reply-To: akm@comix.cs.uoregon.edu (Anant Kartik Mithal)
Organization: University of Oregon Computer Science Dept.
Lines: 33

In article <4922@plains.UUCP> kperson@plains.UUCP (Kerry Person) writes:
>In article <59uengc.BU.EDU> rem@buengc.bu.edu (Robert E. Mee) writes:
>>I have a question about memory configurations
>>for a 386 system. I'm about to buy an SX that
>>comes standard with 1mb in 4 256k simms. I asked
>>the salesperson if it was possible to have 1
>>1mb simm in it's place. He went on to say that 
>>in order to use 1mb simms I have to use a minimum
>>config of 4 meg!!!
>>Is this correct?
>Keep in mind how the data is stored.  Each of those SIMMs stores four bits of
>data.  If you place a 1 Meg SIMM in place of the four 256's, the
>machine has no way to know that all sixteen bits should now be stored in one
>chip.  In fact it is physically impossible, without some major hardware
>changes.  The chips have to work parallel to each other, so if you increase
>the capacity of the first four bits, you gotta do the same with the rest.
>At least that's the way it seems it should work.  Or does the memory get 
>stored 32 bits at a time on a 386?  Conflicting opinions welcomed.
>Kerry Person

I think Kerry's analysis is correct for a system with 32 bits, like a
386DX, *not* a 386SX, which has only 16 bits. Each simm should store
*8* bits of data, not 4, so you should require 2 simms at a time.

In turn, I invite conflicting opinions... :-)

kartik


-----------------------------------------------------------------------------
Anant Kartik Mithal					akm@cs.uoregon.edu
Department of Computer Science				akm@oregon.BITNET
University of Oregon