386 XENIX system sanity check - SUMMARY OF RESPONSES

[Chip Rosenthal]

In article <349 at vector.UUCP> chip at vector.UUCP (that's me!) writes:
>I am very close to a purchase decision on a 386 XENIX system.  I'd
>like to run a couple of things by the collective mind...

I basically raised two issues:  memory architecture and disks for a
386 XENIX system.  I received two responses.

Bill Davidsen (a man of several addresses; his reply-to is davidsen at crdos1)
talked about memory cacheing.  Bill suggested that "cache give[s] 15-30%
improvement for 32 bit memory, about 50% if you go cheap and use 16 bit
memory."  The 15-30% number feels intuitively right to me.  The manufacturer
I'm looking at (ALR) takes about a 50% premium (2K$) to go to cacheing.
Bill commented that "I have the PC Designs GV386, and it is a lot less
than the numbers you quoted."  One thing which might be deceiving is that
my numbers have added memory and disk rolled into them.  But all this
does is make the cache look less and less attractive for the money, at
least on an ALR.

BTW, would anybody really buy a 386 system and run 16-bit memory?

Bill also commented that he is looking into the Perstor PS180 controller.
He has no idea whether it will work, but he said that a number of folks
have given it good reports.

William Mattil (wrm at telcomm) said that he has no experience with the Priam
ID62 drives I mentioned, however he has worked with their V-170 drives.
He called them "dependable and reasonably fast (22 ms) for ST506."

If you have any additional comments on memory or the OMTI 8626/Priam ID62
combo, please send 'em along.  I'll issue a future summary if applicable.
-- 
Chip Rosenthal /// chip at vector.UUCP /// Dallas Semiconductor /// 214-450-0400
{uunet!warble,sun!texsun!rpp386,killer}!vector!chip
I won't sing for politicians.  Ain't singing for Spuds.  This note's for you.