Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!rutgers!princeton!allegra!ulysses!mhuxt!m10ux!braun
From: braun@m10ux.UUCP
Newsgroups: comp.sys.ibm.pc
Subject: Re: IBM, this is 4 u
Message-ID: <204@m10ux.UUCP>
Date: Mon, 27-Apr-87 16:39:51 EDT
Article-I.D.: m10ux.204
Posted: Mon Apr 27 16:39:51 1987
Date-Received: Sat, 2-May-87 16:05:29 EDT
References: <1741@isis.UUCP> <58200035@gorgo.UUCP> <1666@bnrmtv.UUCP>
Organization: AT&T Bell Labs, Murray Hill
Lines: 28
Summary: Drive data rates

In article <1666@bnrmtv.UUCP>, connery@bnrmtv.UUCP (Glenn Connery) writes:
> Steve Blasingame writes:
> > I think that the more significant point was that the folks at the product
> > announcement told me that the new drives are ESDI-based, significantly
> > faster and are linked to a SCSI bus interface. That alone CAN mean transfer
> > rates in excess of 1 MB/Sec.[unexaggerated] (even on single-ended SCSI). This
> > is (of course) quite dependent upon the PS implementation. It is certainly
> > significantly better than the hardware limitation of ST412/506 to about
> > 625K/Sec. and may make these more worthy of multitasking applications.
> 

What is forgotten is that the ST506 data rate is exactly as fast as the
data comes in from the heads on the drive.  Nothing can be faster for a
significant length of time.  A SCSI interface cannot cause a disk transaction
to be completed any more quickly.
However, if the data is buffered on the controller and transmitted in bursts
to the computer, less time will be stolen from the computer bus for DMA
transactions.  Assuming that the computer has something else to
do while waiting for a disk transaction to complete (unlike current
implementations of MS-DOS), things will be speeded up.
However, for speeding up curent PC applications, you must
reduce the seek times and rotational latency for the drive if you want your
applications to run faster.

-- 

Doug Braun		AT+T Bell Labs, Murray Hill, NJ
m10ux!braun		201 582-7039