bruceh@pnet06.cts.com (Bruce Henderson) (07/03/88)
The Motorola 88000 is pretty much THE processor of
thefuture. After going to the Motorola seminars and
studyingthe docs on it, I feel that the power offered
here is something that hopefully Apple won't pass up.
I think the optimum design would be to use to 88100s. One
for graphics only and the other for everything else. If both
of the processor modules used 2 88200 CMMU's on the
instruction side of the path (32K) then it would be possible
to store 1 segment in the CMMU at all times. [ more on why
this is importiant later ] Other system memory could be
configured to operate in burst mode. So as the 88100 needed
another segment, it could read it in a "BURST". [for those
of you not so hardware intensive hacks burst mode allows the
CPU to fetch a large amount of data from memory at the
highest possible rate]. The 2 processor sets could be
connected via the 88000 p bus so that the interrupts could
occur 1 cycle. Yes, this idea doesn't have a lot of
refinement, but I have been doing a great deal of thinking
about ways that a 88K Mac could exist, achitecture and such.
And I have decided to appoint myself the independent
Mac88000 evangelist. The power is just too great to pass
up.
Wait, you say. The 88000 isn't software compatible
with the 68000! Ah, this is where the ROM/Toolbox guys at
Apple will have to earn thier pay. The trick is to make the
Mac88000 source code compatable with the rest of the Mac
line. If every last one of the Mac ROM calls can be
implemented on the Mac88000, then the trick is just to allow
MPW to generate 88000 code as well as 68000 code. So to
make an Mac88000 version of a current peice of software
would only requore a recompile with the {+88000} flag set.
This may seem like too much to ask for, but the truth of the
matter is is that if Apple were to go ahead with this now,
while the hardware hacks were making the box, the toolbox
critters could be already creating the code for this things
ROMS, thanks to the Tektronix 88000 MacII board. So
really, the ROM development team doesn't have to wait until
the hardware guys had somthing that works. So I know that
all of you LightSpeed guys are howling out there. I think
that Apple should make sure that the major compiler writers
are seeded as soon as they have anything that works!
Well, there is a lot more to my Ideas, If anyone is
really interested they can contact me, or just wait for the
next time I feel like having an 88000 revival meeting.....
Bruce Henderson
-AsmI
UUCP: hodge.cts.com!pnet06!bruceh
ARPA: hodge!pnet06!bruceh@crash nosc.mil
INET: bruceh@pnet06.cts.comSTORKEL@RICE.BITNET (Scott Storkel) (07/07/88)
The Motorloa 88000 the processor of the future? For Motorola maybe. But hey, the Sun 4 with 10 MIPS performance is here NOW, using the SPARC architecture. And other companies are supporting SPARC, too. Last I heard, the 88000 was still a pipe dream. So why not the MacSPARC? Or while we're at it, why not let the guys at Apple finish designing their own RISC chip. Heck, we could have Quickdraw in micro-code! Scott Storkel Macintosh Software Development Rice University
spector@vx2.GBA.NYU.EDU (David HM Spector) (07/09/88)
Pipe dream? Huh? Tektronix is developing MacintoshII NuBus cards around
it... they're as expensive as bmw's but they exist.... what one needs to
do is see how fast one can implement a 68020 emulator for the 88000, and if
it will seriously kill the performance..
DHMS
-------------------------------------------------------------------------------
David HM Spector New York University
Senior Systems Programmer Graduate School of Business
Arpa: SPECTOR@GBA.NYU.EDU Academic Computing Center
UUCP:...!{allegra,rocky,harvard}!cmcl2!spector 90 Trinity Place, Rm C-4
HamRadio: N2BCA MCIMail: DSpector New York, New York 10006
AppleLink: D1161 CompuServe: 71260,1410 (212) 285-6080
"What computer puts out work like this?" "Hire us and we'll tell you."
XYZZYGLORPfnf@fishpond.UUCP (Fred Fish) (07/10/88)
In article <370STORKEL@RICE> STORKEL@RICE.BITNET (Scott Storkel) writes: >The Motorloa 88000 the processor of the future? For Motorola maybe. But hey, >the Sun 4 with 10 MIPS performance is here NOW, using the SPARC architecture. >And other companies are supporting SPARC, too. Last I heard, the 88000 was >still a pipe dream. So why not the MacSPARC? Excuse me. A pipe dream? Tek has had an M88000 based coprocessor card running in Mac-II's for months. Yes, it's pricey. Yes, the price will probably drop drastically once chips are available in volume. My impression is that the companies that have announced support for the M88000 are far more serious about producing real, working hardware than the ones that have announced support for SPARC (except Sun of course :-). -Fred (Formerly part of the M88000 software team) -- # Fred Fish, 1346 West 10th Place, Tempe, AZ 85281, USA # noao!nud!fishpond!fnf (602) 921-1113
walter@garth.UUCP (Walter Bays) (07/12/88)
In article <370STORKEL@RICE> STORKEL@RICE.BITNET (Scott Storkel) writes: >The Motorloa 88000 the processor of the future? For Motorola maybe. But hey, >the Sun 4 with 10 MIPS performance is here NOW, using the SPARC architecture. >And other companies are supporting SPARC, too. Last I heard, the 88000 was >still a pipe dream. So why not the MacSPARC? Tektronix can fit the 88000 on a NuBus card because the cache and MMU is integrated on the module. A SPARC couldn't fit, though I think Sun is working on cache and MMU support now. Of course Clipper originated the module with integrated CAMMU, and it's shipping in volume. (No commercial: You can't buy a Clipper Mac board - only PC-AT, UNIX, and bad old MS-DOS :-) Sigh :-( -- ------------------------------------------------------------------------------ My opinions are my own. Objects in mirror are closer than they appear. E-Mail route: ...!pyramid!garth!walter (415) 852-2384 USPS: Intergraph APD, 2400 Geng Road, Palo Alto, California 94303 ------------------------------------------------------------------------------
dsc@izimbra.CSS.GOV (manic pop thrill) (07/12/88)
In article <370STORKEL@RICE> STORKEL@RICE.BITNET (Scott Storkel) writes: >the Sun 4 with 10 MIPS performance is here NOW, using the SPARC architecture. i believe this is the number claimed by sun for certain integer-intensive benchmarks. check out some of the numbers posted every so often in comp.arch to get the feel of a sun-4's performance on a more general (and perhaps more realistic) job mix. the processor often turns out to more like 6 to 7 mips rather than 10. dsc
tim@amdcad.UUCP (07/13/88)
Expires: Sender: Followup-To: In article <925@garth.UUCP> walter@garth.UUCP (Walter Bays) writes: | In article <370STORKEL@RICE> STORKEL@RICE.BITNET (Scott Storkel) writes: | >The Motorloa 88000 the processor of the future? For Motorola maybe. But hey, | >the Sun 4 with 10 MIPS performance is here NOW, using the SPARC architecture. | >And other companies are supporting SPARC, too. Last I heard, the 88000 was | >still a pipe dream. So why not the MacSPARC? | | Tektronix can fit the 88000 on a NuBus card because the cache and MMU is | integrated on the module. A SPARC couldn't fit, though I think Sun is | working on cache and MMU support now. Of course Clipper originated the | module with integrated CAMMU, and it's shipping in volume. (No commercial: | You can't buy a Clipper Mac board - only PC-AT, UNIX, and bad old MS-DOS :-) | Sigh :-( YARC (gee, what does that spell, backwards? ;-) is making a Mac II coprocessor board utilizing the Am29000 RISC processor and the Am29027 Arithmetic Accelerator. Performance is ~30,000 Dhrystones without requiring external caches (it uses burst mode accesses with interleaved static-column DRAMS). -- -- Tim Olson Advanced Micro Devices (tim@delirun.amd.com)
kaufman@polya.Stanford.EDU (Marc T. Kaufman) (07/13/88)
In article <22334@amdcad.AMD.COM> tim@delirun.amd.com (Tim Olson) writes: >YARC (gee, what does that spell, backwards? ;-) Yet Another RISC Computer
domo@riddle.UUCP (Dominic Dunlop) (07/13/88)
[Cross-posting to comp.arc as it seems relevant; followup to comp.sys.mac.] In article <370STORKEL@RICE> STORKEL@RICE.BITNET (Scott Storkel) writes: >And other companies are supporting [Sun's] SPARC, too. Last I heard, the >88000 was still a pipe dream. So why not the MacSPARC? A good business reason is that Sun has got to be a major competitor for Apple from 1989 onwards, and it's not seen as a smart commercial idea to make oneself reliant on a technology perceived as being controlled by a competitor -- even where that competitor has an avowed policy of ``throwing technology over the wall'' in the hope that the resulting growth in the market for standard products will far offset any consequent reduction in the developer's market share. Look how long it took for other hardware suppliers to feel really good about promoting Sun's NFS as their main vehicle for transparent Local-Area file-sharing. Me, I'd love to see MacSPARC, as it would add more momentum to the bandwagon that's promoting SPARC as a standard UN*X hardware architecture. If that bandwaggon doesn't roll, we'll have another five years of having to accommodate multiple architectures for no good reason, and, boy, am I tired of doing that after the last ten years. But I fear it won't happen. Apple's got to come out with a RISC machine to stay in the MIPS race with, among others, Sun. Relying entirely on guesswork, I'd currently put the shortest odds on a Motorola solution. Thoughts, anyone? (See other postings for discussion of whether the 88000 is real or not.) -- Dominic Dunlop domo@sphinx.co.uk domo@riddle.uucp
steve@hpiacla.HP.COM (Steve Witten) (07/13/88)
/ hpiacla:comp.sys.mac / STORKEL@RICE.BITNET (Scott Storkel) / 2:08 pm Jul 6, 1988 / >...why not >let the guys at Apple finish designing their own RISC chip. Heck, we could >have Quickdraw in micro-code! >Scott Storkel >Macintosh Software Development >Rice University >---------- Microcode for a RISC chip? Come on, that would defeat all the advantages of a RISC chip!!!! The ROM approach would probably be better but it would have to be bigger... Can you imagine a Mac with 1Mb ROMs??? =============================================================================== Steve Witten steve%hpiacla@hplabs.HP.COM Industrial Applications Center {ucbvax, hplabs}!hpda!hpdsla!hpiacla!steve Hewlett-Packard Co. steve@hpiacla "...I'm no fool! Nosirree!..." -- J. Cricket
walter@garth.UUCP (Walter Bays) (07/13/88)
Someone: 88000 card for Mac II Scott Storkel: Wish there were a SPARC card. Me: Wish there were a Clipper card. Tim Olson: YARC is making a 29000 card. You sure can do a lot with a NuBus. Now what's the news on the MicroChannel? Did I hear someone was making a Z80 card? :-) -- ------------------------------------------------------------------------------ My opinions are my own. Objects in mirror are closer than they appear. E-Mail route: ...!pyramid!garth!walter (415) 852-2384 USPS: Intergraph APD, 2400 Geng Road, Palo Alto, California 94303 ------------------------------------------------------------------------------
hsu@pitstop.UUCP (David Hsu) (07/14/88)
In article <925@garth.UUCP> walter@garth.UUCP (Walter Bays) writes: >In article <370STORKEL@RICE> STORKEL@RICE.BITNET (Scott Storkel) writes: >>The Motorloa 88000 the processor of the future? For Motorola maybe. But hey, >>the Sun 4 with 10 MIPS performance is here NOW, using the SPARC architecture. >>And other companies are supporting SPARC, too. Last I heard, the 88000 was >>still a pipe dream. So why not the MacSPARC? > >Tektronix can fit the 88000 on a NuBus card because the cache and MMU is >integrated on the module. A SPARC couldn't fit, though I think Sun is >working on cache and MMU support now. I presume you mean "a well-supported SPARC couldn't fit". It seems to me that a simple SPARC board would fit...isn't the guy that designed the Tek 88k board the same one who designed the Definicon SPARC board for the IBM-PC? Of course, there was no cache (just static-column twiddling) and I don't think there was an MMU. -dave hsu, lowly consultant dhsu@sun.com My opinions have nothing at all to do with Sun Microsystems; I just post from here.
bruceh@pnet06.cts.com (Bruce Henderson) (07/14/88)
Actually the original post was mine. and it was more along the
lines of a suggestion that Apple should really consider this
chipset in any new computers they may be planning. I am familiar
with the Tek 88000 board, and I can't wait until the Green Hills
Pascal is up on it. I have been experementing with the 88000
and tring to define possible ways to make such things as the
toolbox work. The thrust of my work right now is Color quick draw.
I can't really go any deeper in depth than that because I am afraid
that Apple will jump down my throat because I have been rewriting
such things as the "bottleneck procedure" in 88000 code as well as some
of their proprietary region stuff. No, NOT for SALE, I want to see
just how hard to would be to make an 88000 Mac. Well, I'll post more
later to let you guys know how the experements go.
Bruce Henderson
asm...
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
SNAIL: 9311 Eton, Chatswoth CA 91311
VOICE: (818) 700 8854
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
UUCP: {crash uunet}!pnet06!bruceh : \I/
ARPA: crash!pnet06!bruceh@nosc.mil : -*-
INET: bruceh@pnet06.cts.com : /I\
UUCP: hodge.cts.com!pnet06!bruceh
ARPA: hodge!pnet06!bruceh@crash nosc.mil
INET: bruceh@pnet06.cts.comdre%ember@Sun.COM (David Emberson) (07/14/88)
At this point, I would think that the high cost of the 88000 chipset would be prohibitive for a low cost machine. Also, the Harvard architecture, while allowing reduced CPI in the absence of an on-chip I cache, requires more external components. If Apple were to do a RISC machine now, the only processors that would meet the cost constraint would probably be the LSI Logic SPARC or Mips cpus which will be available as macros, the Fujitsu SPARC, or possibly the 29K. My guess is that they will stay with 68030/40/50 for some time. Dave Emberson (dre@sun.com) p.s. The only thing that Sun controls about SPARC is the trademark. Many licensees are already designing SPARCs of all flavors on their own. Obviously we have some clout, because at the moment we are the biggest consumer of SPARC chips. But we don't (and couldn't possibly) "control" the SPARC vendors.
ccasths@pyr.gatech.EDU (Scott Hinckley) (07/16/88)
In article <4860014@hpiacla.HP.COM> steve@hpiacla.HP.COM (Steve Witten) writes: >to be bigger... Can you imagine a Mac with 1Mb ROMs??? Yes. +=======================================================================+ |Scott Hinckley - OCS User Assistant AKA - Galaxy's End | |Georgia Insitute of Technology, Atlanta Georgia, 30332 | |uucp: ...!gatech!pyr!ccasths | |ARPA: ccasths@pyr.gatech.edu | +=======================================================================+
rusty@pnet06.cts.com (Mr. Rusty Hodge) (07/16/88)
steve@hpiacla.HP.COM (Steve Witten) writes: >/ hpiacla:comp.sys.mac / STORKEL@RICE.BITNET (Scott Storkel) / 2:08 pm Jul 6, 1988 / >>...why not >>let the guys at Apple finish designing their own RISC chip. Heck, we could >>have Quickdraw in micro-code! > >Microcode for a RISC chip? Come on, that would defeat all the advantages of >a RISC chip!!!! The ROM approach would probably be better but it would have >to be bigger... Can you imagine a Mac with 1Mb ROMs??? Apple should probably look to something like the AMD Bit Slice chips to microcode QuickDraw. It would easily out-perform the 88000. And it would be basically a QuickDraw HLL engine- the desired 'Silicon QuickDraw' everyone has been lusting for. Rusty Hodge, HCR Inc, 1588 N. Batavia St. Orange, CA 92667 (714) 974-6300 rusty@hodge.cts.com [uunet vdelta crash]!hodge!rusty FAX (714) 921-8038
kgeisel@nfsun.UUCP (kurt geisel) (07/16/88)
In article <949@garth.UUCP> walter@garth.UUCP (Walter Bays) writes: >Someone: 88000 card for Mac II >Scott Storkel: Wish there were a SPARC card. >Me: Wish there were a Clipper card. >Tim Olson: YARC is making a 29000 card. > >You sure can do a lot with a NuBus. Now what's the news on the >MicroChannel? Did I hear someone was making a Z80 card? :-) Hey, don't laugh. I remember something about someone making a 6502 card which runs Apple II software. How about a $10K machine which runs Appleworks? +--------------------------------------------------------------------------+ | Kurt Geisel, Intelligent Technology Group, Inc. | | Bix: kgeisel | | ARPA: kgeisel%nfsun@uunet.uu.net US Snail: | | UUCP: uunet!nfsun!kgeisel 65 Lambeth Dr. | | Pittsburgh, PA 15241 | | If a rule fires and no one sees it, did it really fire? | +--------------------------------------------------------------------------+
chris@softway.oz (Chris Maltby) (07/18/88)
In article <607@riddle.UUCP> domo@riddle.UUCP (Dominic Dunlop) writes: > Me, I'd love to see MacSPARC, as it would add more momentum to the > bandwagon that's promoting SPARC as a standard UN*X hardware architecture. > If that bandwaggon doesn't roll, we'll have another five years of having to > accommodate multiple architectures for no good reason, and, boy, am I > tired of doing that after the last ten years. But I fear it won't happen. Then UNIX would be another VMS... A major reason you can buy your UNIX computer so cheaply is that there is competition between hardware and application software builders because they don't have to design YAPOS (yet another proprietary operating system) or invest heaps in porting the application to YAPOS. A hardware standard would leave us locked in with that. Roll on IBM - I want to standardise on the 360 (actually 1401) architecture - its RISC and CISC together!
Brita_CC_Meng@cup.portal.com (07/18/88)
Actually, Dave, I think that the guy that did the Definicon SPARC board for the PC was the one that did the AMD29000 board for the Mac; Trevor Marshall of YARC. (YARC, by the way, stands for "Yet Another Ruddy Coprocessor". Trevor's from Australia.) Brita Meng I have no signature.
jbn@glacier.STANFORD.EDU (John B. Nagle) (07/20/88)
In article <441@softway.oz> chris@softway.oz (Chris Maltby) writes: >Roll on IBM - I want to standardise on the 360 >(actually 1401) architecture - its RISC and CISC together! When IBM came out with the PC/RT, ballyhooing their "reduced instruction set machine", I observed that the PC/RT instruction set was larger than that of the IBM System/360. This was a bit amusing. Incidentally, the architecture of the IBM 1401, a popular business computer of the early 1960s, was totally different from that of the System/360. The IBM 1401 was a variable-number-length machine with decimal arithmetic. Each character of memory held 6 bits. Memory addresses were decimal. The IBM System/360 was the first machine with what we today consider standard computer architecture: byte-addressable memory (the word "byte" was coined at IBM and first appeared outside IBM in the System/360 product announcement.), 32-bit words, 16-bit halfwords, 64-bit longwords, 8-bit characters, and a 16-megabyte address space potentially expandable to 2^32 bytes. It represented a clean break with IBM's two previous product lines; the business machines, with decimal arithmetic and memory, and the scientific machines, the 701-704-709-7040-7090-7094 line, with 36-bit words, binary arithmetic, and a 64K word address space. IBM 1401 emulation hardware was available for the smaller System/360 machines as an option, but this was not part of the 360 architecture. Sorry for the digression, but I wanted to set the record straight. The IBM System/360 was a major breakthrough, and the hardware architecture was excellent. (The same cannot be said for the software architecture; design mistakes made in OS/360 still haunt us.) John Nagle
levin@bbn.com (Joel B Levin) (07/20/88)
In article <17573@glacier.STANFORD.EDU> jbn@glacier.UUCP (John B. Nagle) writes:
(The IBM 1401 was a variable-number-length machine with decimal arithmetic.
(Each character of memory held 6 bits.
8, if you include the parity bit and (fanfare) the word mark.
...
([The 360] represented a clean break with IBM's two previous product lines; the
(business machines, with decimal arithmetic and memory, and the scientific
(machines, the 701-704-709-7040-7090-7094 line, with 36-bit words,
(binary arithmetic, and a 64K word address space.
Plus some other interesting machines like the 1620 (about which I know
little) and the 7070 with its 10,000 words of 10 decimal digits and 99
index registers.
The 1401 (actually 1440) was my very first machine. Easy: read a card
by issuing a Move (MLC) instruction. What a shock to go to a PDP-5
where each individual character had to be read by hand!
( Sorry for the digression, but I wanted to set the record straight.
Ah, nostalgia ...
( John Nagle
/JBL
UUCP: {backbone}!bbn!levin USPS: BBN Communications Corporation
ARPA: levin@bbn.com 150 Cambridgepark Drive
POTS: (617) 873-3463 Cambridge, MA 02140