NetNews Offline 2/NetNews Offline Volume 2.iso/news/comp/sys/amiga/misc-part1/9467

Internet Message Format | 1996-08-05 | 19.9 KB

Path: news.restena.lu!usenet
From: [e-mail] (Manou BILLA)
Newsgroups: comp.sys.amiga.misc
Subject: Re: Haynie joins AT team
Date: 24 Mar 1996 00:01:33 GMT
Organization: Not organized
Message-ID: <[e-mail]>
References: <[e-mail]>
Reply-To: [e-mail]
NNTP-Posting-Host: slip1.restena.lu
X-Newsreader: THOR 2.22 (Amiga;TCP/IP)


On 20-Mar-96 13:43:03, Paul Mellors ([e-mail]) wrote to
All () about Re: Haynie joins AT team
(<[e-mail]>):

Hi Paul

>It's not just a rumour;

>I personally heard it from Stefan Doymeyer's (AT President) own mouth.

>Dave Haynie is ON the AT dev team and is relocating to Bensheim.


>STUPID QUESTION, AND I HOPE I DONT GET FLAMED FOR ASKING THIS BUT
>WHO IS DAVE HAYNIE?? (CRINGE WAITING FOR A SLAP)


Dave is THE AMIGA creator! Nearly all big box AMIGAs are his creations!
A2000, A3000, A4000, SCSI II controllers, GFX cards, turbo cards and other
things , of which most never did appear in the shops! Like the A3000+ or the
A5000 AAA! :-( The A3000+ was nearly finished, even AT&T did have a prototype
to run their own voice controlled OS! which should have been integrated into
the AMIGA OS! :-((((

Check his homepage at:

        http://www.iam.com/people/dave.html

He now works for SCALA (the creators of SCALA MM 200, 300, 400, ...)

Here's some info from his homepage:



       Cool (kewl, adj.)

"An unreachable goal for some, a way of life for others. The obvious example
of this latter category being Dave Haynie, pictured
above"

--from the glossary of The Zorro III Bus Specification

Dave Haynie

Engineer, Producer and Director, Author

[e-mail]

Dave was a Senior Hardware Engineer at Commodore-Amiga, Inc. thoughout the
days of the A2500, A3000, and A4000. He is the producer and
director of a documentary about the last years at Commodore, including details
of management folly and of how billiant and subversive engineers dealt
with same. He is also the author of DiskSalv 3. Both are published by
Intangible Assets Manufacturing. Dave is now a Senior Engineer at Scala, Inc.
as well as continuing to work on Amiga projects for Intangible Assets
Manufacturing.

Comming soon: Articles Dave wrote for a newsletter that never happened, and
maybe even a Dave Haynie Fan Club (perhaps as a mailing list .which
would include random thought by Dave or Dave and Dale and offers to get
authographed merchandise and stuff).



Dave Haynie has been involved in the Amiga community since the dawn of the
Amiga. He was an engineer on the Commodore C128 at the time Commodore
bought Amiga. He started using the Amiga in 1985, as soon as he could get
his hands on one. He bought one in 1986, and has been programming it ever
since.

As an Systems/Hardware Engineer at Commodore, Dave was the chief engineer
of the Amiga 2000, Amiga 2630, the Zorro III Bus Specification, Amiga
@{"3000+" link A3000Plus}, Amiga 4091, and the @{"Nyx prototype" link Nyx},
and a leading member of the
engineering team on the Amiga 3000 and Amiga 4000. Dave had a number of
@{"really cool projects" link Cool.Projects} in the works when Commodore went
under.

Independently of Commodore, Dave has been involved in a number of Amiga
projects. As well as DiskSalv, Dave wrote SetCPU, the popular 68030 MMU
tool, and several other example programs. Dave has written extensively, for
magazines including AmigaWorld, Amazing Computing, Amiga Sentry, .Info,
AmigaWorld Technical Journal, Compute!, Amiga Shopper (UK), and most
recently, Amiga Format.

In less technical times, Dave resides in the great state of South Jersey,
with his wife Liz, kids Sean (3) and Kira (1), dog Auryn (Borzi), and cat
Iggy (Black & White, footwarmer). When he's not on the computer, Dave has
too many other interests. He practices Aikido regularly, he's into
photography, video (see his first film, @{"the Deathbed Vigil" link
Deathbed.Vigil}), cycling,
canoing, swimming, woodworking, Japanese knives, writing (technical and
songs), modern Rock music, and good beer. Since the demise of Commodore
("well, at least I can have a life now"), Dave's been working at Scala,
Inc., and learning to play keyboards.
@endnode

@node A3000Plus
@title "A3000+, the first AGA System"
In 1990, Commodore was nearing completion of the first major upgrade to the
Amiga chip set. Code named "Pandora", and later dubbed "AA" (because of the
@{"AAA" link Nyx} project, already underway), this chip set would boot the
basic graphics
capabilities of the Amiga considerably, while retaining full register-level
compatibility with the ECS and original Amiga chip sets.

Ultimately, Commodore needed a test system for these new chips, and so they
naturally assigned Dave Haynie to the project. Not satisfied to just build
an "AA3000", Dave looking into building an all-around better system which
included the new chipset.

In February of 1991, the first A3000+ booted up Workbench. The new chips
ran the existing AmigaOS almost without incident. In addition to the AA
chips, this first 3000+ had a digital signal processor, the AT&T DSP3210,
built in as a local bus coprocessor. The DSP3210 and the Amiga were a match
made in heaven. The 3210 was a local bus master, allowing for DSP systems
to be built without the expensive and limiting SRAM of earlier designs. The
3210, at 50MHz, crunched 32-bit floating point at up to 25MFLOPS, five
times faster than the 68040. AT&T has a full fledged, multitasking,
multiprocessing DSP operating system for the 3210, which used an arbitrary
general purpose OS as a host. The Amiga's low overhead, near realtime OS
was a perfect mate for AT&T's VCOS.

Of course, prototypes will be prototyes, and the DSP never worked on the
Rev 0 edition of the A3000+. But everything else did. For Rev 1, a very
extensive DSP audio system was put into place, including hardware CODECs
for 16-bit stereo I/O at up to 48kHz and phase-correcting
telecommunications, for V32 modems.

After Rev 1, Commodore Engineering management was changed by the then
president of Commmodore International, @{"Mehdi Ali" link Deathbed.Vigil}. The
new VP of
engineering, Bill Sydnes, was opposed to the A3000+, and virtually every
other project underway at the time -- no use making the previous
administration (the folks who brought you the A500, A2000, and A3000) look
good. The final revision of the Amiga 3000+ was a scaled down version, as
mandated by the administration. A flaw in some custom DSP support logic,
built into the new A3000-architecture DMAC chip, made the DSP a problem.

The DSP lived on for awhile, despite management. Dave Haynie worked on his
own time to get systems reworked, and work out any additional bugs, in the
DSP hardware. Jeff Porter, onetime Director of New Product Development and
the other driving force behind the DSP, managed to keep the software
development funded. Eric Lavitsky, DSP expert, consultant, and longtime
Amiga supporter, did the actual VCOS port. This port was, in fact finished.

And the DSP had a kind of afterlife. After Bill Sydnes was fired, Lew
Eggebrecht took over Commodore Engineering. While not an amazing leader,
Lew did turn a number of projects on that were floundering as skunkworks
efforts necessarily hidden from Sydnes. Dave had proposed a DSP board be
made for Zorro III, and Lew put two engineers on it full time. Although
Commodore never built the resulting board, the design was nearly complete,
and it was build by a company that licensed the design before Commodore
went under.
@endnode

@node Nyx
@title "Nyx, The AAA Prototype"
The most advanced project ever attempted at Commodore was the creation of
the Advanced Amiga Architecture, or AAA. Started in the late 80's, AAA was
an effort to build a new Amiga architecture that was once again head and
shoulders above the mainstream.  AAA was a major advance. It would deliver
32 and 64-bit systems, using DRAM or VRAM. Many new graphics modes were
supported, including 24-bit, HAM10, and compresses 8 and 24-bit modes. The
blitter and copper were fully 32-bit, and the copper could feed the
blitter. Graphic resolution went up to 1280x1024 noninterlaced, and the
pixel clock could change on a line-by-line basis, to support hardware
promotion of older screen modes.  Audio was extended to 8 channels, with
sampling rates up to 100kHz at 16bits/sample. The floppy disk interface was
fast enough for 4MB floppies, 150KB/s CD-ROMs, Digital Radio, and other
serial streams, and decoding could be done on-chip or in software.

In 1992 Dave Haynie designed the "Nyx" prototype, which was the first home
for the AAA chips. Based on the A3000 architecture and lots of programmable
logic, three working systems were built. When Commodore stopped funding the
AAA project, critical chip revisions had been released to tape, but not yet
made. The existing AAA chips delivered 24 bit graphics at high resolution,
demos set up the copper feeding the blitter, doing CPU-less animations. The
next rev was supposed be enough to boot the AmigaOS.
@endnode


@node DeathBed.Vigil
@title "The Deathbed Vigil"
The Deathbed Vigil and other tales of digital angst
                by Dave Haynie

Set the way-back machine for April, 1994. Everyone was worried about the
continued existence of Commodore. I had been away, interviewing for new
jobs in Texas, so I came in, first time that week, on Tuesday, April 26,
1994. Rumors were running rampant about a bigtime layoff happening the next
day. We in Engineering had already been on a major league skeleton crew
since the summer of 1993, so this was clearly a sign of the beginning of
the end.

So, when I woke up Wednesday, not knowing with any certainty if I'd have a
job to go to tomorrow, I thought about videotaping Commodore. After all,
this whole Amiga thing, which ran far beyond Commodore, the Amiga
community, nearly all my active interpersonal relationships, and in some
sense, the last vestige of the real small computer industry; once full of
excitement and new ideas, but by this time more concerned with perpetuating
and recreating obsolete, "best of the 70s, as long as UNIX isn't
considered" computing. It was one of those ideas you think about, say "hey,
wouldn't thi be cool", but then dismiss as soon as there's an obsticle.

So I set out to do some taping. Fortunately, I had three batteries charged
for my Sony TR-7 8mm camcorder; recharged after my trip to Texas. And
fortunately, K-Mart had blank 8mm tapes for sale. So I went to Commodore,
and proceeded to do a walk-around of the Commodore building.

Pretty early on, it was clear that the layoff was happening. All but about
30 people were layed off; I was one of the "still employed", it was less
clear who the lucky ones were. At lunchtime, we went to our Mexican place,
Margarita's, for the last big layoff party. There, many things were said
about the Commodore management, some of it on-camera.

At the layoff party, Randell Jesup told me of a "Deathbed Vigil" party that
he and Bryce Nesbitt were throwing, on Saturday. When Commodore bought
Amiga, the good folks at the original Amiga company in Los Gatos, CA, held
an "Amiga Wake" party. This proved premature, if in retrospect technically
correct; Randell didn't want to make the same mistake. So I filmed the
party, where all kinds of cool things took place: interviews, tales of the
golden and not-so-golden years of Commmodore-Amiga, a burning of the L.B.M.
effigy (some associate this with ex-President of Commodore, Mehdi Ali),
smashing of keyboards, the "Chicken Lips Blues" song (performed by Mike
Rivers, written by attendees), and other great events. Some strange
happenings, post-party, were also filmed at Commodore.

Once done, I had to figure out what to do with this 4-5 hours of video. I
decided to make a real, for sale videotape, and to try to tell a bit of the
story of What Went Wrong, along with the antics, anger, info, and catharsis
of this time. I realized I wanted to have on tape some small piece of this
amazing thing I had been involved with for 10 years, and I figured fans of
the Amiga might want a look too. So I set out to really make my first film.

And along the way, I decided I wanted to know: is Desktop Video real. I
never did any video stuff at Commodore. So I wanted to know, could a novice
videomake sit down with a consumer camcorder and deck, an Amiga, some plug
ins and the right software, and actually make a good video. So I put
together my system, including:

        @{"Amiga 3000+" link A3000plus} prototype
        Scala MM300 authoring system (provided by Scala)
        Scala EE100 LANC/IR controller (provided by Scala)
        SuperGen 2000 (borrowed)
        GVP TBCplus (borrowed)
        JVC HR6900 SVHS deck (paid in cash)
        Deluxe Paint IV

Over the course of four months, I put together the video. I added various
bits of information gleaned from conversations with past and present
Commodore employees, and other folks "in the know". Mike Rivers provided me
with some original music, and I wrote lyrics to one of his songs, which I
affectionately entitled "F.Y.M.". We drank beer, mixed it, and I became a
Rock Icon. NOT. Anyway, if you like the Amiga, and want some idea of what
went wrong, you can still order "The Deathbed Vigil and other tales of
digital angst" from @{"Intangible Assets Manufacturing" link IAM}.
@endnode

@node Cool.Projects
@title "Other Really Cool Projects"
There seems to be a general feeling in the Amiga community that Commodore's
engineering teams spent years developing Really Amazing Things, only to
have Marketing deep-six them on the verge of production. This does
occasionally happen; the @{"A3000+" link A3000Plus} is a good example of this.
What was more
often the case, though, were projects done on the side, "skunkworks"
projects, if you will, that were often cool, but didn't get very far, due
to varying management support, available funding, and the amount of
"copious spare time" available to the project's owner.

A few of Dave Haynie's less successful projects, over the years, included:

        A2630           This one actually made it out the door. For six
                        months, it was a funded skunkworks project. Nearly
                        overnight, it became A Real Product.

        BIGRAM          A 16MB board for the A2630. Two were built. Hey, in
                        1988, 16MB was lots of memory.

        FASTRAM         An 8MB board with fast page support for the A2630.
                        As it turned out, this was a bit too complex to do
                        in PAL logic at the required speeds, but it was a
                        good design exercise.

        BIGRAMZ3        This was done in about two weeks, start to finish,
                        as a Zorro III design example for the 1991 DevCons.
                        This is a 64MB Zorro III memory card that supports
                        Zorro III burst. It benchmarkst at about 80% the
                        speed of local bus memory, a fairly impressive
                        accomplishment given the less-than-ideal Zorro III
                        interface of the A3000 architecture. About four of
                        these exist.

        A2631           After the A3000 went out, Commodore was still,
                        strangely enough, shipping lots of A2500/30s.
                        Certain niches wanted the larger box of the A2000.
                        Every A2500 got an A2630 and A2091 board. One
                        Friday, over beer and Mexican food, Dave Haynie and
                        Greg Berlin got the idea that this was stupid, in
                        the light of the A3000 architecture. So the next
                        week, Dave cranked out a replacement, based on the
                        A3000 architecture, which we called the A2631.
                        This was an A2000 CPU socket board with Buster,
                        RAMSEY, the DMAC and SCSI chip, 68030, and 68882.
                        It cost less than the A2630, delivered high
                        performance SCSI, and could take 16MB of RAM.
                        Management wasn't interested, even though it would
                        have saved money. Two prototype boards were built.

        Gemini          This was a multiprocessing board, designed to test
                        and stress the features of the Level II Buster chip
                        (Rev 8 and beyond). The problem with inventing your
                        own expansion bus is that you have to build
                        everything for it. So for fun, this board did
                        something interesting; it had two 68030s, each with
                        4MB of RAM and independent Zorro III access. Two of
                        these were built, but the project resources were
                        pulled before it was debugged. Had it been
                        developed fully, this could easily have helped to
                        debug the Buster chip before the A4000 shipped.

        Acutiator       Only a paper design, Acutiator was an effort to
                        specify a whole new system-level architecture,
                        replacing the A3000 architecture used in all A3000
                        and A4000 systems. The goal was a cost efficient,
                        high performance architecture that could deliver
                        anything from midrange systems (about midway
                        between A4000 and A1200) on up to fully
                        professional Amiga system heretofor nonexistent.
                        Haynie originally designed a new "Amiga Modular
                        Interconnect" bus, but adopted the fairly similar
                        PCI bus once it was announced. The main idea was to
                        make "highly modular" Amiga systems, wherein the
                        system board design was independent of CPU or
                        graphics subsystems. A small amount of design work
                        had been done on this, but it was largely ignored
                        by management.

        SCARAB          The SCARAB board, the last thing Haynie worked on
                        at Commodore, was an effort to build a high
                        performance graphics card based on off-the-shelf
                        SVGA chips. The card ran a PCI bus locally, with
                        bridged to Zorro III and to the video slot. With
                        the video slot interface, Amiga chip graphics could
                        be converted, in realtime, to PCI cycles,
                        which wrote the SVGA graphic memory, in a window
                        controlled by SCARAB registers.  In essence, this
                        was a programable "flickerFixer" that could handle
                        any scan rate. The board could also support
                        "hybrid" graphics modes, where in the Amiga chips
                        were still used, but went into a very slow scan
                        mode, so they could put out 1024x768 at 8 bits in
                        slowscan, which would be converted to 72Hz
                        noninterlaced by SCARAB (this is somewhat like
                        "Hedley hires", an easy addition to the AmigaOS).
                        RTG drivers would ultimately hit the board,
                        directly, over Zorro III. Lots of design work went
                        into this, but it became pretty clear there was no
                        money left to actually build any of it.
@endnode



03-AUG-95 / Intangible Assets Manufacturing / Dale L. Larson / [e-mail]



--
Bye Manou
------------------------_----------------------------------------------
Manou BILLA       |  _ //  Connect your AMIGAs...
4, Ave. Nic Kreins|  \X/             ... A1000 / A2500 / A3000 ...
L-9536 WILTZ      |      ------ Member Team AMIGA Luxembourg ------
                  | email [e-mail]      FIDO 2:2455/560.8
-----------------------------------------------------------------------
           [PGP public key available on request]

...  Oregano: the ancient Italian art of pizza folding.