Path: utzoo!attcan!uunet!cs.utexas.edu!csd4.milw.wisc.edu!srcsip!tcnet!rosevax!cimcor!mike
From: mike@cimcor.mn.org (Michael Grenier)
Newsgroups: comp.sys.ibm.pc
Subject: Re: scsi rll trade off questions?
Message-ID: <809@cimcor.mn.org>
Date: 9 Jul 89 12:16:00 GMT
References: <3299@copper.MDP.TEK.COM>
Organization: Grenier & friends, Forest Lake, MN
Lines: 44

From article <3299@copper.MDP.TEK.COM>, by byronl@copper.MDP.TEK.COM (Byron Lunz):
> In article <14978@ut-emx.UUCP> allred@ut-emx.UUCP (Kevin L. Allred) writes:
>>I had pointed out to me that Segate has recently started marketing a
>>low cost SCSI addaptor (ST01 and ST02) suitable for use with its
>>ST296N 80MB hard disk.  This combination reportedly offeres about 750
>>KB/sec transfer rate, which is comparable to the 1:1 interleve RLL
> 
> I received my new Gateway 2000 386/20 a few days ago.  It arrived with
> a Seagate ST296N and SCSI controller (not sure of the model #).  Transfer
> rate was one of my reasons for purchasing this system, and I was assured
> prior to the purchase by the salesperson that I could expect 800KB/sec.
> I was quite disappointed when both Spintest and Coretest 2.7 gave me
> data transfer rates of 440-460KB/sec!  Then, just today Mark Davis

What does Spintest actually test? Keep in mine that the drive has to 
physically seek and read the blocks into the scsi buffer on the drive
before transfer can begin again on the bus. If Spintest is measuring
the total time that it took for the request to get to the BIOS to the
time it took to get results back (as it must since Spintest doesn't
know the actual harware involved) then you are measuring the drive
delays as well as the SCSI bus transfer speed. If the drive could seek
and read the data in zero time then I'm sure you would see the much faster
SCSI transfer rates.

In a UNIX environment, scsi is a bigger win since the CPU can run another
process while waiting for the Seagate ST-01 controller to interrupt
up it to tell it that the transfer is ready to begin. 

Going to a fancier controller which supports DMA may speed up the
transfer rate but Spintest is still going to report a time slower
that the what an ESDI or RLL controller will do hooked to the same 
drive. For instance, using this Adaptec 2732 RLL 1:1 controller 
on this Maxtor drive (also a 28ms drive) I get over 800K bytes per
second transfer as reported by DOS based disk programs. 

A big plus for scsi is that it can make multiple disk requests at the
same time to different devices assuming you have a real operating
system, obviously not MSDOS, which can really speed up system
performance but with only one disk and a single tasking system the
closer you get to the disk head the better. 

    -Mike Grenier
    mike@cimcor.mn.org
    uunet!rosevax!cimcor!mike