mroussel@alchemy.chem.utoronto.ca (Marc Roussel) (05/02/91)
Recently, I posted a message to comp.sys.cbm indicating my
disappointment with Commodore's marketing effort. I have a second
anecdote to relay, this one with a happy conclusion.
In Algorithm, no. 2.2 (March 1991), Walter Stricker mentions the
Amiga's hardware "blitter". The following is a quote from his letter to
the editor:
A blit is a rapid transfer of blocks of memory from one
place to another. [...]
Because it can combine up to three blocks of memory using
any one of the 256 possible logical functions, it is
(conceptually) a massively parallel universal one bit computer.
This makes it quite well suited to computing CA's. [...]
PopLife written by Olaf Siebert computes 13 generations per
second of Conway's Life CA on a 320 by 200 cell field.
The good news is, I now know what I can do with an Amiga that I couldn't
do with most other microcomputers. The bad news is, this department
(which does a lot of CA work) doesn't have an Amiga and as far as I
know, no one's looking at them. We are looking at far more expensive
hardware which may give us this sort of performance on CA's (and will
blow the Amiga out of the water in other areas, I'm sure). Given the
numbers quoted above however, the Amiga would have an excellent
price/performance ratio on certain types of problems.
What I'd like to know from Commodore is why didn't we find out
about things like this when we were trying to decide what hardware to
buy? Do those guys even know that this sort of thing is hot current
research and that numbers like those described above are significant to
people like us? Where are your sales reps, Commodore?
On a more technical side, does anyone know if these blitter operations are
available on lower end Amiga's? This would make the Amiga's even more
respectable since we could then use A1000's (or whatever) as development
workstations and use the higher end machines for production work. Are
these blitter operations accessible from a high-level language
(preferably some form of Fortran)?
I realize that the cross-post to comp.sys.cbm is probably not
entirely appropriate, but in view of the recent discussion there, I
thought it might be interesting to some of the people there. In any
event, I have redirected followups to comp.sys.amiga.misc. As I don't know
much about the Amiga newsgroups, I trust someone will redirect the
conversation to a more appropriate group if I've messed up. Just be
clear about where you're sending it if you do so I can subscribe to the
appropriate newsgroups and thus keep abreast of this discussion.
Marc R. Roussel
mroussel@alchemy.chem.utoronto.caalfalfa@milton.u.washington.edu (Corey Lawson) (05/03/91)
Hmm...Well, any amiga should be good for what you want to do. I hate to do this, but the CA programs for the NeXT usually are written to take advantage of the DSP chip that is on it (Motorola DSP56000) (this is kind of an inad- vertent plug for the NeXT...) Other than that...Plus, if you get a NeXT, someone makes a board with 5 DSPs on it, plus Motorola has made a newer version of it, called the DSP96000 or something, that is truly awesome, but I'm not sure if anyone has released it in any product for NeXTs, Amigas, PCs, or whatever... -corey lawson alfalfa@milton.u.washington.edu
jdickson@jato.jpl.nasa.gov (Jeff Dickson) (05/03/91)
In article <1991May2.184530.20781@milton.u.washington.edu> alfalfa@milton.u.washington.edu (Corey Lawson) writes: >Hmm...Well, any amiga should be good for what you want to do. I hate to do >this, but the CA programs for the NeXT usually are written to take advantage >of the DSP chip that is on it (Motorola DSP56000) (this is kind of an inad- >vertent plug for the NeXT...) > >Other than that...Plus, if you get a NeXT, someone makes a board with 5 DSPs >on it, plus Motorola has made a newer version of it, called the DSP96000 or >something, that is truly awesome, but I'm not sure if anyone has released it >in any product for NeXTs, Amigas, PCs, or whatever... > >-corey lawson >alfalfa@milton.u.washington.edu ANSWER: What good in any computer.
jdickson@jato.jpl.nasa.gov (Jeff Dickson) (05/03/91)
>In article <1991May2.184530.20781@milton.u.washington.edu> alfalfa@milton.u.washington.edu (Corey Lawson) writes: >Hmm...Well, any amiga should be good for what you want to do. I hate to do >this, but the CA programs for the NeXT usually are written to take advantage >of the DSP chip that is on it (Motorola DSP56000) (this is kind of an inad- >vertent plug for the NeXT...) > >Other than that...Plus, if you get a NeXT, someone makes a board with 5 DSPs >on it, plus Motorola has made a newer version of it, called the DSP96000 or >something, that is truly awesome, but I'm not sure if anyone has released it >in any product for NeXTs, Amigas, PCs, or whatever... > >-corey lawson >alfalfa@milton.u.washington.edu ANSWER: What good is any computer.
peterk@cbmger.UUCP (Peter Kittel GERMANY) (05/03/91)
In article <1991May2.143334.23026@alchemy.chem.utoronto.ca> mroussel@alchemy.chem.utoronto.ca (Marc Roussel) writes: > > [ The Blitter ] > This makes it quite well suited to computing CA's. [...] > PopLife written by Olaf Siebert computes 13 generations per > second of Conway's Life CA on a 320 by 200 cell field. > > What I'd like to know from Commodore is why didn't we find out >about things like this when we were trying to decide what hardware to >buy? Do those guys even know that this sort of thing is hot current >research and that numbers like those described above are significant to >people like us? Where are your sales reps, Commodore? Well, many people DID find out. But you see, Life is a subject only for a few scientists, the average sales person or dealer is simply bored by such stuff, they really don't know the interesting sides of life (Life?). I don't see any way to use this as a marketing argument for broader audience. And the scientists normally know very quickly about such possibilities, don't know why this didn't make it up to your place. The mentioned software PopLife really runs on *640* by 200 (or 256 PAL or 512 PAL interlaced) screens! It's REALLY amazing. I keep it in my demo drawer to impress people like you with Amiga. > On a more technical side, does anyone know if these blitter operations are >available on lower end Amiga's? This would make the Amiga's even more >respectable since we could then use A1000's (or whatever) as development >workstations and use the higher end machines for production work. Are >these blitter operations accessible from a high-level language >(preferably some form of Fortran)? To all questions a clear YES. -- Best regards, Dr. Peter Kittel // E-Mail to \\ Only my personal opinions... Commodore Frankfurt, Germany \X/ {uunet|pyramid|rutgers}!cbmvax!cbmger!peterk
mroussel@alchemy.chem.utoronto.ca (Marc Roussel) (05/04/91)
In article <1197@cbmger.UUCP> peterk@cbmger.UUCP (Peter Kittel GERMANY) writes: >In article <1991May2.143334.23026@alchemy.chem.utoronto.ca> >mroussel@alchemy.chem.utoronto.ca (Marc Roussel) writes: >> [ The Blitter ] >> This makes it quite well suited to computing CA's. [...] >> >> What I'd like to know from Commodore is why didn't we find out >>about things like this when we were trying to decide what hardware to >>buy? Do those guys even know that this sort of thing is hot current >>research and that numbers like those described above are significant to >>people like us? Where are your sales reps, Commodore? > >Well, many people DID find out. But you see, Life is a subject only >for a few scientists, the average sales person or dealer is simply >bored by such stuff, they really don't know the interesting sides of >life (Life?). I don't see any way to use this as a marketing argument >for broader audience. And the scientists normally know very quickly >about such possibilities, don't know why this didn't make it up to >your place. I'm not asking for Commodore to use CA's as part of their general marketing campaign, I'm suggesting that they should hire sales reps with experience in scientific computation and send them to science departments. Apple and IBM do this. As to your comment that "scientists normally know very quickly", this is certainly not the case here. Many companies actually appear to be interested in selling us machines, and we let them come to us. We don't have time to look under rocks for interesting new hardware. If Commodore doesn't send sales reps to our department, we won't find out about their machines except serendipitously, as I just did. It's really that simple. Marc R. Roussel mroussel@alchemy.chem.utoronto.ca
elg@elgamy.RAIDERNET.COM (Eric Lee Green) (05/08/91)
From article <1991May2.184530.20781@milton.u.washington.edu>, by alfalfa@milton.u.washington.edu (Corey Lawson): > Hmm...Well, any amiga should be good for what you want to do. I hate to do > this, but the CA programs for the NeXT usually are written to take advantage > of the DSP chip that is on it (Motorola DSP56000) (this is kind of an inad- > vertent plug for the NeXT...) The NeXT doesn't cost $500, though. You're right, though. The Amiga blitter runs at the same rate no matter what processor is in the machine... so a low-cost Amiga 500 is as good as the top-of-the-line 25mhz Amiga 3000 for that particular application. (Actually, the A3000 could do it faster with the CPU than the A500 could do it with the blitter, but that's another tale). Now, if Commodore ever releases a high-speed chipset, that story would change radically... -- Eric Lee Green (318) 984-1820 P.O. Box 92191 Lafayette, LA 70509 elg@elgamy.RAIDERNET.COM uunet!mjbtn!raider!elgamy!elg Looking for a job... tips, leads appreciated... inquire within...