Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!hpda!hplabs!ucbvax!VLSI.JPL.NASA.GOV!bolef%albedo.span
From: bolef%albedo.span@VLSI.JPL.NASA.GOV
Newsgroups: comp.os.vms
Subject: Comment on uVAX disk controllers: Response to Art Stine
Message-ID: <871027094338.04o@VLSI.JPL.NASA.GOV>
Date: Tue, 27-Oct-87 12:43:37 EST
Article-I.D.: VLSI.871027094338.04o
Posted: Tue Oct 27 12:43:37 1987
Date-Received: Thu, 5-Nov-87 04:38:25 EST
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The ARPA Internet
Lines: 37


This is a response to Art Stine's query on disk controllers

	I've been using Sigma SDC-RQD11-SC's for a year with
success. These are caching controllers (1 Mbyte cache) which handle
two SMD drives of any variety. Sigma is at (714) 630-6553. They may
not be as "hot" as the Andromeda, but the relationship between
actual through-put on your CPU and the bells-'n'-whistles of the
controller is a sometimes thing. 
 
	The big advantage of Sigma is the price. I've gotten them
for less than $1000 each from a local firm, R-Squared (Rick
Pickett, 303 799-9292). Of course the Dilog's handle four drives,
though they don't have on-board cache. don't forget the Dilog
DQ-236, if your transfer rate is under 2.5 Mbyte/second, as it is
for the 9720's. Also, Plessey makes a virtually identical
controller to the DQ-236 (Rick Ramirez, 505 471-6360), if you want
to get a little bidding war started. 
 
	As you probably know, the subject of on-board controller
cache is a controversial one. VMS has lots of file caching already,
and getting data from this source is much faster than getting it
from a controller's cache, since you avoid all the QIO overhead.
The only possible advantage of controller cache comes if you open
and close the same files frequently, since the VMS caches flush
when a file is closed. If Jim Reynolds of Andromeda lays his
"super-performance" trip on you, challenge him by asking for a unit
to test against your an ordinary controller with your standard
applications!
 
Regards,
Larry Bolef                   Arpa-Internet: BOLEF%ALBEDO@ames-io.arpa
University of Colorado  SPAN/Physnet/Hepnet: ALBEDO::BOLEF -or- 33839::BOLEF
Campus Box 449                       TEXnet: UTADNX::UTSPAN::ALBEDO::BOLEF
Boulder, CO 80309                  Telemail: [LBOLEF/NASA] NASAMAIL/USA
(303) 492-8124    UUCP(forwarded to ALBEDO): bolef@cires.colorado.edu
-------