rumbo@impch.imp.com (Peter Kunz) (08/30/90)
hello there there have been some rumors in fido about an a3500 coming out next year and an a4000 with 68040 sometime in the future. does anyone know more about this? is the 68040 card for the a2000 and a3000 available yet. would it be worth plugging the 68040 into a 16 bit a2000 or has c= come up with some ingenious gadget whicxh turns the a2000 into a 32 bit machine? bcnu pete
6600dan@ucsbuxa.ucsb.edu (Dan Zerkle) (09/01/90)
In article <2231@impch.imp.com> rumbo@impch.imp.com (Peter Kunz) writes:
there have been some rumors in fido about an a3500 coming out next year and
an a4000 with 68040 sometime in the future. does anyone know more
about this?
is the 68040 card for the a2000 and a3000 available yet. would it be worth
plugging the 68040 into a 16 bit a2000 or has c= come up with some ingenious
gadget whicxh turns the a2000 into a 32 bit machine?
Wild rumors with no known basis in fact say:
Commodore is rumored to be working on an Amiga 3500 which is an Amiga
3000 in a tower case with lots of slots (3000 has 4 zorro inline with
2 ISA and one video). THis is more than reasonable because the only
reason the 3000 has few slots is because the case isn't any taller.
Commodore is rumored to be working on a 68040 machine, which would
presumably be the Amiga 4000. This rumor is substantiated somewhat in
that Motorola mentions Commodore as one of the manufacturers that is
working with their new '040 chip.
If Commodore is not working on these, they should be.
However, NOBODY actually has a commercial release of a 68040-based
machine. This is because Motorola does not have everything perfect on
the chip. There is still testing going on. Apparently, the math unit
was causing fits the last I heard.
HOWEVER, NeXT, Inc. has announced that they are going to announce a
68040 machine in the middle of September. This would point to a very
near release date for the 040. What's more, Supra is working on a
card that goes in the CPU slot of the 3000 that carries an 040. My
guess is that the latter has a lot of development left to do, and that
NeXT is going to have something ready to go very soon.
As to the 16/32 bit dilemma.... When you get a board with a 68030
accellerator, you typically get to put 32-bit RAM on that board.
Thus, you have a partially 32-bit system. Unfortunately, the Zorro II
bus is permanently 16-bit. This means that RAM there is SLOW (to your
68030). Most other devices on the bus will not cause problems at 16
bits. You will be much better off with 32-bit SCSI if you want to run
some extremely disk-intensive application like file serving on a
network, or swapping and paging on UNIX. A very high-speed network
(is thick ethernet 32-bit?) could also conceivably benefit from 32-bit
bus access. For most single-user applications, though, you'll be fine
at 16 bits.
--
Dan Zerkle 6600dan@ucsbuxa.ucsb.edu 6600dan@ucsbuxa.bitnet (805) 968-4683
Amiga.... Because life is too short for boring computers.dalka@cbnewsc.att.com (kenneth.j.dalka) (09/04/90)
This is pure speculation on my part but..... I was reading an article that said that the 68040 was a single chip that contained both the Math coprocessor and the "regular" processor in one unit. Because of this, the 68040 actually costs less to make and will probably cost less to buy from motorola that a 68030 and the matching math chip. Therefore CBM could save money on the A3000 by just replacing the 68030 and math chip on the motherboard with a 68040 arrangement. I've also heard that the A3000 motherboard has hardware on it that only a 68040 knows how to deal with. The rest is a marketing decision. -- Ken Dalka (Bell Labs) att!ihlpf!dalka IHC 1U-228 (708) 305-4449
navas@cory.Berkeley.EDU (David C. Navas) (09/05/90)
In article <1990Sep4.134613.11940@cbnewsc.att.com> dalka@cbnewsc.att.com (kenneth.j.dalka) writes: >This is pure speculation on my part but..... >I was reading an article that said that the 68040 was a single chip >that contained both the Math coprocessor and the "regular" processor >in one unit. Because of this, the 68040 actually costs less to make >and will probably cost less to buy from motorola that a 68030 and the >matching math chip. I'm sure that in the short run, this will not be true. >The rest is a marketing decision. Speaking of marketing decisions, how about a slew of better naming conventions? The numbering scheme is getting slightly awkward... For example, for that new 24 bit color machine, with hi-res monochrome display capability (for that programmer in us all...) how about the Amiga Ranger :) Just my opinion. David Navas navas@sim.berkeley.edu "Excuse my ignorance, but I've been run over by my train of thought." -me
mt87692@tut.fi (Mikko Tsokkinen) (09/05/90)
> >This is pure speculation on my part but..... > >I was reading an article that said that the 68040 was a single chip > >that contained both the Math coprocessor and the "regular" processor > >in one unit. Because of this, the 68040 actually costs less to make > >and will probably cost less to buy from motorola that a 68030 and the > >matching math chip. > I'm sure that in the short run, this will not be true. > >The rest is a marketing decision. > Speaking of marketing decisions, how about a slew of better naming conventions? > The numbering scheme is getting slightly awkward... For example, for that new > 24 bit color machine, with hi-res monochrome display capability (for that > programmer in us all...) how about the Amiga Ranger :) > Just my opinion. Excuse me but why do you have to have 68040 in your machine when most of people are quite happy with 68000. Especially if you allready have A3000. What to do you with more speed? Compared to price you pay for it (68040 $800). Just my $0.02. MIT -- Mikko "Assembler rules OK!" Tsokkinen Internet mt87692@tut.fi : UUCP tut!mt87692 : Bitnet mt87692@fintut
amc4919@cec2.wustl.edu (Adam Michael Costello) (09/11/90)
In article <MT87692.90Sep5160456@korppi.tut.fi> mt87692@tut.fi (Mikko Tsokkinen) writes: > Excuse me but why do you have to have 68040 in your machine when most of >people are quite happy with 68000. Especially if you allready have A3000. >What to do you with more speed? Compared to price you pay for it (68040 $800). >Just my $0.02. I'm sure everyone has different reasons, but people like me mainly want to have a computer that kicks everybody else's computer's butt (haven't you ever heard anyone call a particular computer a "kick-ass machine"?). Also, given _any_ amount of computing power, we will find a way to use it. AMC (a soon-to-be A3000 owner, already a debtor!)
murphy@maxwell.physics.purdue.edu (William J. Murphy) (09/11/90)
In article <MT87692.90Sep5160456@korppi.tut.fi> mt87692@tut.fi (Mikko Tsokkinen) writes: > >> >This is pure speculation on my part but..... >> >I was reading an article that said that the 68040 was a single chip > > Excuse me but why do you have to have 68040 in your machine when most of >people are quite happy with 68000. Especially if you allready have A3000. >What to do you with more speed? Compared to price you pay for it (68040 $800). Speaking just for myself, If I had the choice of crunching the data I have now at speed x versus speed y*x where y>1, I will choose y*x. I can't wait for NeXT to produce their 68040 upgrade so that I can start processing at y*x. I hate having to calculate how many days it will take to finish filtering and averaging my data. I never knew how slow a 25 MHz 68030 could be. 8-) So, if C= comes out with a 68040 Amiga before I scrounge the $$$$ to purchase an A3000, I will buy that instead. Bill Murphy | Turtle Schmurtle, I'm the murphy@gibbs.physics.purdue.edu | Rabbit. Amiga, it's not a religion, it's a lifestyle | B. Bunny