saj@chinet.chi.il.us (Stephen Jacobs) (12/08/88)
I'm just stirring it up here. If you read electronics ads, you've noticed that the price premium for faster rather than slower memory chips isn't what it used to be. Similarly, I can easily remember when an 8 MHz 68000 used to fetch more than the price of a 16 MHz 68000 today. My understanding of the guts of an ST isn't what it might be, but it seems that it would essentially require faster versions of the custom chips and a relatively small change in the handling of video refresh to make a top-to-bottom 16 MHz ST that Atari could make a profit selling at a price I'd consider paying (yeah, I'm glossing over FCC approval--how much of a problem is it? I dunno). Maybe the Atari contingent might consider this a suggestion. The ST seems to match up in raw crunch power roughly even with an AT-class machine with 20-30% faster clock speed. The magazines suggest that the up-and-coming home computer is a ~20 MHz AT clone (I wouldn't want one, but so what?). A faster ST would line up nicely there, with the (if and when) 68030 box leaning heavily on the 80386-based office class of machines. I'm greedy. I don't wanna lose that cpu speed to wait states ANYWHERE.
hyc@math.lsa.umich.edu (Howard Chu) (12/09/88)
In article <7107@chinet.chi.il.us> saj@chinet.chi.il.us (Stephen Jacobs) writes:
%I'm just stirring it up here. If you read electronics ads, you've noticed
%that the price premium for faster rather than slower memory chips isn't what
%it used to be. Similarly, I can easily remember when an 8 MHz 68000 used to
%fetch more than the price of a 16 MHz 68000 today. My understanding of the
%guts of an ST isn't what it might be, but it seems that it would essentially
%require faster versions of the custom chips and a relatively small change in
%the handling of video refresh to make a top-to-bottom 16 MHz ST that Atari
%could make a profit selling at a price I'd consider paying (yeah, I'm glossing
%over FCC approval--how much of a problem is it? I dunno).
% Maybe the Atari contingent might consider this a suggestion. The ST seems
%to match up in raw crunch power roughly even with an AT-class machine with
%20-30% faster clock speed. The magazines suggest that the up-and-coming home
%computer is a ~20 MHz AT clone (I wouldn't want one, but so what?). A faster
%ST would line up nicely there, with the (if and when) 68030 box leaning heavily
%on the 80386-based office class of machines.
% I'm greedy. I don't wanna lose that cpu speed to wait states ANYWHERE.
Well, Megabyte Computers in Texas has yet to deliver on their promised Turbo16
upgrade. Now, two months after their expected release, I have to believe that
it's not as easy as all that. Unfortunate. I was really looking forward to it.
(Y'know, a Mega-4 with a 68000 running at 16 MHz, with 68881 [or 68882] math
coprocessor, would probably outclass a Sun 3/50. Throw in the 19 inch mono
screen and it would be a very very slick workstation.)
I think you're being too conservative re: the 68030. Why restrict 'em to office
machines? A slower-clocked 68030 would make a pretty decent home workhorse, eh?
(And to think, a few months ago I would've given anything for a 25Mhz 68020...)
Lotsa nice dreams.
I heard, don't remember where, that an ST with a 16 MHz CPU could not be used
with the blitter chip. (I'd really like to remember where I heard this, 'cause
I'd like to know who's got an ST working with a 16 MHz CPU!!) That'd be awful
depressing, if true. (Of course, SoftTrek's TurboST would still work. Gee,
200% CPU speedup, oughta mean 400% faster textblits! Wow! Maybe I shouldn't
have sold my 1040 after all...)
I heartily agree with that last line. I wanna run flat out, top speed, no
obstacles... Gimme faster chips! (Yeehah!)
--
/
/_ , ,_. Howard Chu
/ /(_/(__ University of Michigan
/ Computing Center College of LS&A
' Unix Project Information Systemsjafischer@spurge.waterloo.edu (Jonathan A. Fischer) (12/10/88)
As a related question, I'm really curious whether _anyone_ has seen the 'Turbo' 16MHz upgrade that was first advertised almost a year ago. I believe they're still advertising in STart as if it were delivering, but so far I've yet to see one mention of anyone actually seeing it, let alone using it. Last I heard, it was going into production, and this was quite some time ago. What's the current status? Anyone out there know? -- -Jonathan Fischer Pet peeve #256: The pulltop on V8 juice cans.
saj@chinet.chi.il.us (Stephen Jacobs) (12/10/88)
In the referenced article, Howard Chu cited as evidence that it would be harder than I suggested to make a 16 MHz ST the experience of Megabyte Computers, whose 16MHz accelerator board never reached production. I'm suggesting something different. The story I heard about why the speedup flopped was that the MSI-type logic to connect the cpu to the slower video circuitry was designed around a chip that was later found to be inadequate. You make the change that video gets every FOURTH memory cycle instead of every second memory cycle (the cpu still couldn't use them, since it's a design principle of the ST to use memory faster than the cpu, but maybe DMA perform- ance could benefit from the newly available cycles). You also make those custom chips in the box faster. It's not an add-on, it's a new machine, but the engineering changes would seem to be of a kind that could be done properly by a small number of people in a short time (=affordable). If anyone from Megabyte is reading this, I hope you're back at it after the disappointment.
hyc@math.lsa.umich.edu (Howard Chu) (12/12/88)
In article <10338@watdragon.waterloo.edu> jafischer@spurge.waterloo.edu (Jonathan A. Fischer) writes: > > As a related question, I'm really curious whether _anyone_ has >seen the 'Turbo' 16MHz upgrade that was first advertised almost a year >ago. I believe they're still advertising in STart as if it were >delivering, but so far I've yet to see one mention of anyone actually >seeing it, let alone using it. Last I heard, it was going into >production, and this was quite some time ago. What's the current >status? Anyone out there know? >-- > -Jonathan Fischer >Pet peeve #256: The pulltop on V8 juice cans. When I called Megabyte in September, they said they were being forced to redesign the board because one of the parts they used was no longer available on the market. They mentioned that they might have the new design ready by October. Also, they were accepting orders for the board, but only by mail. (i.e., no phoned in credit card orders...) I called again last week, but I guess it was a little too early Texas time, so none of their support staff were in yet. The guy answering the phone did say, however, that the board was not yet available. I was going to call again later but have succeeded in putting it off so far... -- / /_ , ,_. Howard Chu / /(_/(__ University of Michigan / Computing Center College of LS&A ' Unix Project Information Systems