Date   : Mon, 06 Aug 1990 18:29:22 GMT
From   : usc!zaphod.mps.ohio-state.edu!rpi!bu.edu!orc!inews!cadev6!dbraun@ucsd.edu (Doug Braun ~)
Subject: How to speed up Ampro LB+ SCSI?

In article <8190004@hp-lsd.COS.HP.COM> was@hp-lsd.COS.HP.COM (Bill Stubblebine)
writes:
>I have a question for any Ampro Little Board Z80+ BIOS hackers still left
>out there.
>
>My configuration:
>
>   Ampro LB Z80+ (w/built-in SCSI interface)
>   Adaptec ACB4000 (not 4000A) SCSI hard disk controller
>   Seagate ST-125 20 MB 40 ms hard disk drive
>   3M MCD-403 40 MB QIC SCSI tape drive 
>   NZ-COM/Z-System
>
.
.
.
>
>If you're still with me, I wonder if anyone has managed to get more than
>30.7K bytes per second net throughput to/from the hard disk out of a
>configuration similar to mine.  I've read the Ampro BIOS source and the
>Adaptec technical manual several times without finding a clue to speeding
>things up further.  What's the trick?
>

Since you are already directly accessing the SCSI bus to run the tape drive,
you should do the same to access the disk.  You could then read at least
32K at a time from the disk.  In my UZI system, I swapped 32K bytes at a time.
My hardware was a 4MHz Z80, a custom-built (simple) SCSI host adapter
that used a Z80-DMA chip, a Shugart SCSI to ST-506 controller, and a
hard disk with 8 heads.  I was able to use a 2:1 interleave.  With
this setup, it takes about 4 revolutions to read 32K, which is ~68 ms.
Allowing 2 ms for overhead, this gives you thruput of over 450K bytes/sec.
The DMA chip allows me to read data fast enough for this.  If you have to use
programmed I/O, you will not do as well, had have to use a bigger interleave.

With all these SCSI disk controllers, if you do many small reads instead
of one large one, the overhead time will dominate the transfer time.
I noticed on my CP/M BIOS, which uses 1K transfers (2 disk sectors at a time),
that the performance is essentially independent of the disk interleave.

With your tape setup, if you read 8k from disk, and write it to tape,
you might keep the drive streaming msot of the time.  If not, you could
at least read 32K, and write 4 tape blocks per start/stop.
Beware. If you always let almost 250 ms go by between writing tape blocks,
you may have very large interrecord gaps, which will reduce your tape capacity.

I have dealt with most of these issues while interfacing a Memtec drive
to my system.

Doug Braun                             Intel Corp CAD
                                       408 765-4279

 / decwrl \
 | hplabs |
-| oliveb |- !intelca!mipos3!cadev4!dbraun
 | amd    |
 \ qantel /

 or:

 dbraun@scdt.intel.com