From caret@pyramid.com  Ukn Jul  4 00:28:10 1993
Received: from gossip.pyramid.com by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA22811; Sun, 4 Jul 93 00:28:09 EDT
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA07389; Sat, 3 Jul 93 21:27:36 -0700
Received: by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA17860; Sat, 3 Jul 93 21:27:25 -0700
Received: by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA17852; Sat, 3 Jul 93 21:27:23 -0700
From: caret@pyramid.com (Neil Russell)
Message-Id: <9307040427.AA17852@sword.eng.pyramid.com>
Subject: Should we have SCSI?
To: riscy@pyramid.com
Date: Sat, 3 Jul 93 21:27:23 PDT
X-Mailer: ELM [version 2.3 PL11]
Sender: riscy-request@pyramid.com
Reply-To: riscy@pyramid.com
Status: RO
X-Status: 

In my opinion, Adaptec have one of the best SCSI controllers that you
can get for a PC like machine.  Why is it good?  It provides master
DMA, scatter/gather modes, etc.;  In short, all the things that ISA makes
it difficult to get on an x86 machine.  But now we have the option
to do better.


SCSI on the motherboard:

Advantages:
	*  Higher bandwidth available to the SCSI bus gives higher
	   transfer rates and lower latencies
	*  Access to main memory is a 32-bit data path, so consumes
	   half of the main memory bandwidth (assuming the 3730 or
	   3041 DMA)
	*  Much cheaper than a performance ISA SCSI card.
	*  Not limited to first 16 MB of memory
	*  Don't have to provide ISA bus master mode
Disadvantages:
	*  Cost of motherboard goes up (maybe)
	*  More software to write to drive it, especially if we
	   do scatter/gather

If we intend to provide SCSI on ISA rather than the mother board,
we need to provide bus mastering capability to the ISA bus.  I think
that the chip count to do this would be around 10.  The chips would
probably be cheaper than a SCSI controller chip and connector, but they
would need extra board space (most of these chips are buffers
of some kind).  Also, there is a fairly large amount of time required
to design and test such a feature, and then the cost of the SCSI card.

Alternatively, a SCSI controller chip is designed to connect to a
system such as ours and would take relatively little effort to get
going.

Some cost guesses:

    ISA SCSI:
	*  Bus mastering logic:		$10
	*  Board space for logic:	$5
	*  SCSI controller card:	$300
    Motherboard SCSI:
	*  SCSI chip:			$17
	*  Glue logic:			$2
	*  SCSI connector:		$2
	*  Board space for this:	$2


For those wanting SCSI, the motherboard wins hands down.

In my last life, I used a MIPS magnum.  For those that don't know, this
is a MIPS workstation like machine which in our configuration had a
MIPS R3000/R3010 running at 33 MHz, 128 MB of memory, and several giga-
bytes of disk connected to a SCSI controller on the motherboard.
Its subjective performance is still the best I've ever seen, including
the 12 CPU Pyramid R3000 machine with its 35 GB of disk that I'm typing
on now.  Originally I put it down to a good operating system and a
fast CPU.  Now I'm sure that they have some good interface to that
SCSI bus.


Comments?
-- 
Neil Russell		(The wizard from OZ)
Pyramid Technology			Email:  caret@pyramid.com
3860 N. First Street			Voice:  (408) 428-7302
San Jose, CA 95134-1702			  FAX:  (408) 428-8845

 
From rei2!tsprad@uunet.UU.NET  Ukn Jul  4 17:26:06 1993
Received: from gossip.pyramid.com by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA07411; Sun, 4 Jul 93 17:26:05 EDT
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA27116; Sun, 4 Jul 93 14:25:37 -0700
Received: by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA29958; Sun, 4 Jul 93 14:25:24 -0700
Received: from goss.pyramid.com
	by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA29947; Sun, 4 Jul 93 14:25:22 -0700
Received: from relay1.UU.NET 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA27108; Sun, 4 Jul 93 14:25:29 -0700
Received: from spool.uu.net (via LOCALHOST) by relay1.UU.NET with SMTP 
	(5.61/UUNET-internet-primary) id AA27999; Sun, 4 Jul 93 17:24:03 -0400
Received: from rei2.UUCP by spool.uu.net with UUCP/RMAIL
	(queueing-rmail) id 172239.3868; Sun, 4 Jul 1993 17:22:39 EDT
Received: by rei.com (5.4/5.40/1.0)
	id AA16861; Sun, 4 Jul 1993 15:37:21 -0500
From: rei2!tsprad@uunet.UU.NET (6692)
Message-Id: <9307042037.AA16861@rei.com>
Subject: Should we have SCSI?
To: riscy@pyramid.com
Date: Sun, 4 Jul 1993 15:37:20 -0500 (CDT)
In-Reply-To: <9307040722.AA15258@rei.com> from "UUCP Login" at Jul 4, 93 02:22:02 am
X-Mailer: ELM [version 2.4 PL13]
Content-Type: text
Content-Length: 720       
Sender: riscy-request@pyramid.com
Reply-To: riscy@pyramid.com
Status: RO
X-Status: 

Neill said:
> 
> SCSI on the motherboard:
> 
> If we intend to provide SCSI on ISA rather than the mother board,
> we need to provide bus mastering capability to the ISA bus.  
> 
> Alternatively, a SCSI controller chip is designed to connect to a
> system such as ours and would take relatively little effort to get
> going.
> 
> For those wanting SCSI, the motherboard wins hands down.

Seems clear to me!

-- 
Ted Spradley   Recognition International, Inc.  Opinions are mine, not theirs.
2701 E Grauwyler Rd. |Your productivity is not enhanced when you're staring
Irving TX 75061      |at that thing.  Your productivity is enhanced when
214-579-6692         |the computer is working and you're doing something else.

 
From jcallen@Think.COM  Ukn Jul  6 13:38:34 1993
Received: from gossip.pyramid.com by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA08018; Tue, 6 Jul 93 13:38:14 EDT
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA01753; Tue, 6 Jul 93 10:37:33 -0700
Received: by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA21603; Tue, 6 Jul 93 10:37:17 -0700
Received: from goss.pyramid.com
	by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA21574; Tue, 6 Jul 93 10:37:14 -0700
Received: from Mail.Think.COM 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA01738; Tue, 6 Jul 93 10:37:22 -0700
Received: from Luna.Think.COM by mail.think.com; Tue, 6 Jul 93 13:35:53 -0400
From: Jerry Callen <jcallen@Think.COM>
Received: by luna.think.com (4.1/Think-1.2)
	id AA04097; Tue, 6 Jul 93 13:35:52 EDT
Date: Tue, 6 Jul 93 13:35:52 EDT
Message-Id: <9307061735.AA04097@luna.think.com>
To: riscy@pyramid.com
Subject: Should we have SCSI?
Sender: riscy-request@pyramid.com
Reply-To: riscy@pyramid.com
Status: RO
X-Status: 

   From: caret@pyramid.com (Neil Russell)
   Date: Sat, 3 Jul 93 21:27:23 PDT

   In my opinion, Adaptec have one of the best SCSI controllers that you
   can get for a PC like machine.  Why is it good?  It provides master
   DMA, scatter/gather modes, etc.

Sounds great, but you forgot to mention the part number! :-)

Seriously:

- What part are you thinking of?

- Is it easy to get in small quantities?

- How easy is it to glue to a 3051 memory bus? I looked at the NCR 53C710 a
  while back, and it would have been a pain to glue to the bus.

- Got a number for Adaptec so I can get the doc?

-- Jerry

 
From drew@caesar.cs.Colorado.EDU  Ukn Jul  6 13:55:42 1993
Received: from gossip.pyramid.com by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA08836; Tue, 6 Jul 93 13:55:39 EDT
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA03832; Tue, 6 Jul 93 10:55:01 -0700
Received: by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA28172; Tue, 6 Jul 93 10:54:43 -0700
Received: from goss.pyramid.com
	by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA28137; Tue, 6 Jul 93 10:54:40 -0700
Received: from caesar.cs.colorado.edu 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA03637; Tue, 6 Jul 93 10:54:47 -0700
Received: from localhost by caesar.cs.Colorado.EDU with SMTP id AA00998
  (5.65c/IDA-1.4.4 for <riscy@pyramid.com>); Tue, 6 Jul 1993 11:54:32 -0600
Message-Id: <199307061754.AA00998@caesar.cs.Colorado.EDU>
To: riscy@pyramid.com
Subject: Re: Should we have SCSI? 
In-Reply-To: Your message of "Tue, 06 Jul 1993 13:35:52 EDT."
             <9307061735.AA04097@luna.think.com> 
Date: Tue, 06 Jul 1993 11:54:31 -0600
From: Drew Eckhardt <drew@caesar.cs.Colorado.EDU>
Sender: riscy-request@pyramid.com
Reply-To: riscy@pyramid.com
Status: RO
X-Status: 


--------

       From: caret@pyramid.com (Neil Russell)
       Date: Sat, 3 Jul 93 21:27:23 PDT
    
       In my opinion, Adaptec have one of the best SCSI controllers that you
       can get for a PC like machine.  Why is it good?  It provides master
       DMA, scatter/gather modes, etc.
    
    Sounds great, but you forgot to mention the part number! :-)

If I interpreted his post correctly,  he was refering to the 1542 ISA 
controllers, built arround Adaptec's own chip, and commenting how 
all of the features that made it great were not needed on our 
system (especially with the Adaptec's $200 price tag).

I've never seen an Adaptec chip on a non-adaptec product, and the chip
probably isn't suited to what we need.  Personally, I like the NCR53c90A/
NCR53cf90 chips.  Not nearly as hard to interface as the 53c700 series,
not as hard to program (the 53c700 series have onboard microcontrolllers,
and are best programmed with NCR's SCRIPTS development tools), not as
featureful but probably well suited to our purposes.  In case you 
missed it, here's my posting regarding the chip :

(BTW, a QFP which would use less realestate is also available,
there are also a 53c94/53cf94 chips which offer a 8 or 16 bit 
wide bus (you have to demux it though), I'm not sure about 
pricing on these)

--
Ok, I called NCR's SCSI division, got some literature on the 53c90 
series, saw that the 53c90 looks like what we wanted, and talked to
the local NCR distributor, Electrodyne (303)-695-903 for pricing
information.

Overview : 

The NCR 53c90(A or B) can sustain 6M/sec asynchronous on a one 
foot cable, worst case 4.5M/sec on a one foot cable,
3M/sec on a maximum length (single ended) cable (provided
that the internal drivers are used.  Note that asynch. SCSI
is only spec'd to 2.3M/sec, and for higher transfer rates 
sync. SCSI must be used). It also does sync.  SCSI to 5M/sec.  
A fast version of the chip is also available, I'm less clear
on it's specs.  


The controller is totally integrated, ie there are onchip
SCSI drivers, etc - we only need to add decode circuitry 
and a 25Mhz clock (Ok, technically it requires a 10Mhz clock
for async, 12 Mhz for sync, but the sync. SCSI clock
is derived from the external clock and the minimum divisor
is five - so we need a 25Mhz crystal oscilator for this thing 
to get 5M/sec sync. SCSI).

It comes in a 68 lead PLCC, .910" square.

The host/DMA interface is 8 bits wide, so it's best on an 
8 bit bus of the 3730.

The controller is semi-intelligent, ie it supports
commands to "arbitrate, select, send message, then 6,10,
or twelve command bytes and generate interrupt when
done", "transfer information", "return status byte 
followed by message byte".  In the normal case, 
we'll service four interrupts per command (one for "Select
with ATN sequence", one when reselection completes,
one for information transfer complete, one for command
complete). We'll set up for one DMA transfer per command.
(Essentially all I/O can be done directly by
the CPU or the DMA, except for sync. SCSI DATAIN/DATAOUT
which can only be transfered via DMA) - either one through the 
16 byte FIFO buffer.  Personally, I'd DMA the DATAIN/DATAOUT
SCSI phases so we only have the overhead of one DMA controller
setup, the command bytes, message / status in, etc will fit in the 
16 byte FIFO buffer.

Pricing : 

59c90A
1-49 : $17.60
50-99 : $16.35
100-999: $15.20

( In light of this, we definately want to go with the cf flavor)

59cf90 :
1-49: $19.45
50-99 : $18.60
100-999 : $16.80

Important question : 


With SCSI, you want to minimize the command overhead per block, so 
you want to read/write the maximum number of blocks per command.  How 
ever, buffer-cache is non-contiguous, so you need scatter-gather-support 
to do it.

Without scatter/gather, you end up doing one read/write per 
block and performance can be quite abysmal (as in, before we 
implemented scatter-gather in the Linux SCSI code, we saw 
60K/sec through the file system, after 500-700K/sec on 
Seagates, more on reasonable hardware).

So : Does the 3051/3081 / 3730 combination support DMA to non-contiguous
pieces of memory for one DMA transfer? (With motorola 68k machines, 
DMA is done through the paging unit so this is automatic, 
Intel doesn't do it which is why all decent PC SCSI boards must bu$ ma$ter)?

--

Right now the SCSI disk driver tries for  a 16 sector readahead (ie, 8K),
this is small enough so that we can DMA directly to an intermediate 
buffer and copy it back on the interrupt we recieve after the information
transfer phase is complete, so I retract my whining about scatter/gather
in hardware.


 
From tim@ubitrex.mb.ca  Ukn Jul  6 15:59:48 1993
Received: from gossip.pyramid.com by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA15208; Tue, 6 Jul 93 15:59:46 EDT
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA23354; Tue, 6 Jul 93 12:59:20 -0700
Received: by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA21862; Tue, 6 Jul 93 12:59:04 -0700
Received: from goss.pyramid.com
	by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA21846; Tue, 6 Jul 93 12:58:59 -0700
Received: from ubiserver.ubitrex.mb.ca 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA23281; Tue, 6 Jul 93 12:59:04 -0700
Received: from ska.ubitrex.mb.ca ([192.75.16.23]) by ubitrex.mb.ca (4.1/SMI-4.1)
	id AA10528; Tue, 6 Jul 93 14:57:52 CDT
Date: Tue, 6 Jul 93 14:57:52 CDT
From: tim@ubitrex.mb.ca (Tim Braun)
Message-Id: <9307061957.AA10528@ubitrex.mb.ca>
To: riscy@pyramid.com
Subject: Re: Should we have SCSI?
Sender: riscy-request@pyramid.com
Reply-To: riscy@pyramid.com
Status: RO
X-Status: 


> SCSI on the motherboard:

> For those wanting SCSI, the motherboard wins hands down.

I definitely agree.  And the proliferation of CD-ROM is
going to make SCSI more desirable through the next 3-5 years.

The part could be socketed, though PLCC sockets drive me up
the wall around here with unreliable connections.  

There are a number of SCSI controller chips worth looking
at, including the NCR 5380 and derivatives (including the
NatSemi DP8490) and the Western Digital 33c93 and derivatives.

I would tend towards the 5380 type, as they are multiple sourced
(cheap), well-known (Seagate ST-01), and can provide reasonable
performance.  PLCC packages are common.

________________________________________________________________
Tim Braun                          |
Ubitrex Corporation                | Voice: 204-942-2992 ext 228
1900-155 Carlton St                | FAX:   204-942-3001
Winnipeg, Manitoba, Canada R3C 3H8 | Email: tim@ubitrex.mb.ca

 
From drew@nag.cs.Colorado.EDU  Ukn Jul  6 16:34:11 1993
Received: from gossip.pyramid.com by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA17148; Tue, 6 Jul 93 16:34:09 EDT
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA29625; Tue, 6 Jul 93 13:33:38 -0700
Received: by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA25195; Tue, 6 Jul 93 13:33:22 -0700
Received: from goss.pyramid.com
	by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA25188; Tue, 6 Jul 93 13:33:20 -0700
Received: from nag.cs.colorado.edu 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA29618; Tue, 6 Jul 93 13:33:20 -0700
Received: from localhost by nag.cs.Colorado.EDU with SMTP id AA13755
  (5.65c/IDA-1.4.4 for <riscy@pyramid.com>); Tue, 6 Jul 1993 14:33:07 -0600
Message-Id: <199307062033.AA13755@nag.cs.Colorado.EDU>
To: riscy@pyramid.com
Subject: Re: Should we have SCSI? 
In-Reply-To: Your message of "Tue, 06 Jul 1993 14:57:52 MDT."
             <9307061957.AA10528@ubitrex.mb.ca> 
Date: Tue, 06 Jul 1993 14:33:07 -0600
From: Drew Eckhardt <drew@nag.cs.Colorado.EDU>
Sender: riscy-request@pyramid.com
Reply-To: riscy@pyramid.com
Status: RO
X-Status: 


--------

    
    > SCSI on the motherboard:
    
    > For those wanting SCSI, the motherboard wins hands down.
    
    I definitely agree.  And the proliferation of CD-ROM is
    going to make SCSI more desirable through the next 3-5 years.
    
    The part could be socketed, though PLCC sockets drive me up
    the wall around here with unreliable connections.  
    
    There are a number of SCSI controller chips worth looking
    at, including the NCR 5380 and derivatives (including the
    NatSemi DP8490) and the Western Digital 33c93 and derivatives.

    I would tend towards the 5380 type, as they are multiple sourced
    (cheap), well-known (Seagate ST-01), and can provide reasonable
    performance.  

Nope.

I wrote the Linux Seagate ST-01 (ie, with the Nat. Semi part) driver, and 
I would avoid this part like the plague.  The designer obviously failed his 
introductory digital circuits class where they taught FSM design, because
the chip can't tell you when the SCSI REQ line transitioned low->high,
only when it's currently high.  Faster SCSI devices work OK if you 
just sit in a tight loop and blindly write to the port, with slower
SCSI devices, you have to loop, wait for it to go low then high, and 
be sure to timeout because you'll miss it a lot.  The chip can only 
generate an interrupt when the SCSI SEL line is raised, so most of the 
time you sit arround spinning your wheels in a loop (and you can't 
DMA to it because you need to make sure the status registers have the 
right values)

IMHO, the 53cf90 series is the best choice I've seen so far, I don't
know about the Western Digital chip, but it rings a bell as the chip
used in the WD7000 controller.

--------


 
From tim@ubitrex.mb.ca  Ukn Jul  6 17:19:06 1993
Received: from gossip.pyramid.com by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA20034; Tue, 6 Jul 93 17:19:05 EDT
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA04813; Tue, 6 Jul 93 14:17:56 -0700
Received: by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA04674; Tue, 6 Jul 93 14:17:41 -0700
Received: from goss.pyramid.com
	by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA04662; Tue, 6 Jul 93 14:17:39 -0700
Received: from ubiserver.ubitrex.mb.ca 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA04809; Tue, 6 Jul 93 14:17:44 -0700
Received: from ska.ubitrex.mb.ca ([192.75.16.23]) by ubitrex.mb.ca (4.1/SMI-4.1)
	id AA10815; Tue, 6 Jul 93 16:16:34 CDT
Date: Tue, 6 Jul 93 16:16:34 CDT
From: tim@ubitrex.mb.ca (Tim Braun)
Message-Id: <9307062116.AA10815@ubitrex.mb.ca>
To: riscy@pyramid.com
Subject: Re: Should we have SCSI?
Sender: riscy-request@pyramid.com
Reply-To: riscy@pyramid.com
Status: RO
X-Status: 


> 
>     I would tend towards the 5380 type, as they are multiple sourced
>     (cheap), well-known (Seagate ST-01), and can provide reasonable
>     performance.  
> 
> Nope.
> 
> I wrote the Linux Seagate ST-01 (ie, with the Nat. Semi part) driver, and 
> I would avoid this part like the plague.  ...

I had heard the DP5380 had bugs, but didn't know the extent of them.  It's
precisely this experience that's useful to us, though.  The derivative chips 
(dp8490, 53c80, 53c90) chips were more what I meant, and the similarity can 
be useful.  Choosing a chip that's a complete unknown leads to a repeat 
performance (return of the plague).

> IMHO, the 53cf90 series is the best choice I've seen so far, I don't
> know about the Western Digital chip, but it rings a bell as the chip
> used in the WD7000 controller.

NCR 53cf90 sounds fine, assuming the pricing holds (US$19.45).  I assume
the "cf" is faster than the "c"?  The data you gave didn't say why one would
pay the US$1.85 more for the "cf".

You know, I've been watching/collecting data on chips like this for years, 
and do you think I can find any of it in my office now?  Not a chance.
________________________________________________________________
Tim Braun                          |
Ubitrex Corporation                | Voice: 204-942-2992 ext 228
1900-155 Carlton St                | FAX:   204-942-3001
Winnipeg, Manitoba, Canada R3C 3H8 | Email: tim@ubitrex.mb.ca

 
From drew@caesar.cs.Colorado.EDU  Ukn Jul  6 17:27:27 1993
Received: from gossip.pyramid.com by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA20606; Tue, 6 Jul 93 17:27:25 EDT
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA05453; Tue, 6 Jul 93 14:26:12 -0700
Received: by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA06260; Tue, 6 Jul 93 14:25:57 -0700
Received: from goss.pyramid.com
	by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA06251; Tue, 6 Jul 93 14:25:55 -0700
Received: from caesar.cs.colorado.edu 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA05449; Tue, 6 Jul 93 14:26:02 -0700
Received: from localhost by caesar.cs.Colorado.EDU with SMTP id AA02574
  (5.65c/IDA-1.4.4 for <riscy@pyramid.com>); Tue, 6 Jul 1993 15:25:32 -0600
Message-Id: <199307062125.AA02574@caesar.cs.Colorado.EDU>
To: riscy@pyramid.com
Subject: Re: Should we have SCSI? 
In-Reply-To: Your message of "Tue, 06 Jul 1993 16:16:34 CDT."
             <9307062116.AA10815@ubitrex.mb.ca> 
Date: Tue, 06 Jul 1993 15:25:30 -0600
From: Drew Eckhardt <drew@caesar.cs.Colorado.EDU>
Sender: riscy-request@pyramid.com
Reply-To: riscy@pyramid.com
Status: RO
X-Status: 


    > IMHO, the 53cf90 series is the best choice I've seen so far, I don't
    > know about the Western Digital chip, but it rings a bell as the chip
    > used in the WD7000 controller.
    
    NCR 53cf90 sounds fine, assuming the pricing holds (US$19.45).  I assume
    the "cf" is faster than the "c"?  The data you gave didn't say why one woul
   d
    pay the US$1.85 more for the "cf".

The 53cf90/53c94 chips are fast-SCSI II versions of  the 53c90/53c94.
    
--------


 
From caret@pyramid.com  Ukn Jul  6 19:20:57 1993
Received: from gossip.pyramid.com by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA24798; Tue, 6 Jul 93 19:20:56 EDT
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA16062; Tue, 6 Jul 93 16:19:57 -0700
Received: by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA28516; Tue, 6 Jul 93 16:19:42 -0700
Received: by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA28506; Tue, 6 Jul 93 16:19:40 -0700
From: caret@pyramid.com (Neil Russell)
Message-Id: <9307062319.AA28506@sword.eng.pyramid.com>
Subject: Re: Should we have SCSI?
To: riscy@pyramid.com
Date: Tue, 6 Jul 93 16:19:39 PDT
In-Reply-To: <9307061735.AA04097@luna.think.com>; from "Jerry Callen" at Jul 6, 93 1:35 pm
X-Mailer: ELM [version 2.3 PL11]
Sender: riscy-request@pyramid.com
Reply-To: riscy@pyramid.com
Status: RO
X-Status: 

>    From: caret@pyramid.com (Neil Russell)
>    Date: Sat, 3 Jul 93 21:27:23 PDT
> 
>    In my opinion, Adaptec have one of the best SCSI controllers that you
>    can get for a PC like machine.  Why is it good?  It provides master
>    DMA, scatter/gather modes, etc.
> 
> Sounds great, but you forgot to mention the part number! :-)
> 
> Seriously:
> 
> - What part are you thinking of?
> 
> - Is it easy to get in small quantities?
> 
> - How easy is it to glue to a 3051 memory bus? I looked at the NCR 53C710 a
>   while back, and it would have been a pain to glue to the bus.
> 
> - Got a number for Adaptec so I can get the doc?

I was refering to the 1542, 1742, etc boards.  I called Adaptec and was told
that they did not sell any of their chips; just boards.  Despite this
I have seen Adaptec chips on disk drives, and maybe one or two other
places, but the point still stands: we cannot use Adaptec at all.

-- 
Neil Russell		(The wizard from OZ)
Pyramid Technology			Email:  caret@pyramid.com
3860 N. First Street			Voice:  (408) 428-7302
San Jose, CA 95134-1702			  FAX:  (408) 428-8845

 
From aki@akix.cts.com  Ukn Jul  6 22:20:12 1993
Received: from gossip.pyramid.com by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA29565; Tue, 6 Jul 93 22:20:11 EDT
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA00453; Tue, 6 Jul 93 19:19:13 -0700
Received: by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA24779; Tue, 6 Jul 93 19:18:56 -0700
Received: from goss.pyramid.com
	by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA24772; Tue, 6 Jul 93 19:18:53 -0700
Received: from crash.cts.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA00446; Tue, 6 Jul 93 19:19:03 -0700
Received: from akix by crash.cts.com with uucp
	(Smail3.1.28.1 #15) id m0oDP5p-0000CjC; Tue, 6 Jul 93 19:18 PDT
Received: by akix.uucp (/\==/\ Smail3.1.21.1 #21.2)
	id <m0oDOlh-0002HoC@akix.uucp>; Tue, 6 Jul 93 18:57 PDT
Message-Id: <m0oDOlh-0002HoC@akix.uucp>
Date: Tue, 6 Jul 93 18:57 PDT
From: aki@akix.cts.com (Aki Atoji)
To: riscy@pyramid.com
Subject: Re: Should we have SCSI?
In-Reply-To: <9307062116.AA10815@ubitrex.mb.ca>
References: <9307062116.AA10815@ubitrex.mb.ca>
Sender: riscy-request@pyramid.com
Reply-To: riscy@pyramid.com
Status: RO
X-Status: 



   >     I would tend towards the 5380 type, as they are multiple sourced
   >     (cheap), well-known (Seagate ST-01), and can provide reasonable
   >     performance.  
   > 
   > Nope.
   > 
   > I wrote the Linux Seagate ST-01 (ie, with the Nat. Semi part) driver, and 
   > I would avoid this part like the plague.  ...

I've programmed 5380 in embedded SCSI arena as well.  I concur with
avoiding the chip.  The problem with this chip is that it's not much
more than a SCSI 'latch', and that this chip requires a lot of hand
holding by the driver code which tends to be a lot of work.  The
driver must step through each of the command phases, pretty much
monitoring the SCSI bus directly through the chip.  While this is an
excellent way to learn the actual SCSI bus operation (you start to
memorize waveforms this way), it takes a lot of code and uses up CPU
time.

The 53C9x series are definitely not this way and is recommended.  I
believe one of the PD OS's (BSD Net-2 tape?  Mach?) even had driver
for 53C94 in it.

---------------------------------------------------------------------------
Aki Atoji              Unix, X, Networking and Embedded Realtime Consulting
           aki@akix.cts.com                   crash!akix!aki@trout.nosc.mil
---------------------------------------------------------------------------

 
