$CSB211@LSUVM.BITNET (mark orr) (05/06/90)
i sense that you guys weren't entirely thrilled with my
multiple 68k approach - though i can't understand why - it
is both doable and inexpensive...
anyway, the title of this article is correct. thirty-two bit
6502's already exist (and have existed since late '87)
the chip that i'm referring to is the arm - acorn risc machine.
it is a risc chip with an instruction set deliberately designed
around the 6502 - actually the instruction set is a superset of
the 6502. it was designed for the archimedes a310 computer in
england, which was previewed in the october 1987 issue of byte.
the archimedes was designed to be a high-performance replacement
of the acorn computer (bbc micro) which uses the 6502 as its main
processor. it is also operating system compatible with the bbc micro.
how does it perform? well, here's the specs:
(I'D ENCOURAGE YOU TO FIND A COPY OF THE OCT '87 BYTE AND
SEE THIS MACHINE FOR YOURSELF...THE ARTICLE RUNS FROM PAGES
125 TO 130)
THE ARCHIMEDES A310 COMES STANDARD WITH
- 1 MEGABYTE OF RAM
- 1 3.5" 800K DISK DRIVE
- 512K ROM
- GRAPHICS: 640 BY 512, 256 COLORS OUT OF A
PALETTE OF 4096
- SOUND: EIGHT VOICES, STEREO
- PARALLEL PRINTER INTERFACE, RS-432 (RS-422?) SERIAL INTERFACE
ACCORDING TO BYTE LAB'S TESTS:
ARCHIMEDES COMPAQ DESKPRO MAC SE W/
386 (16 MHZ) HYPERCHARGER
80287 (10 MHZ)
FIBBONACCI 52.4 SEC 53.1 71.6
SIEVE 5.7 SEC 6.0 14.9
SORT 10.0 SEC 5.6 20.6
SAVAGE 91.2 SEC 21.5 8.8
DHRYSTONES/ 4901 3748 2176
SECOND ----
ALL BENCHMARKS WERE PERFORMED IN C. THE ONLY DISAPPOINTING
MEASURE WAS IN THE SAVAGE BENCHMARK (MANY NESTED TRIG. FUNCTIONS,
LOGS, ANTILOGS, ETC.) WHEN THESE TEST WERE PERFORMED IN THE
INCLUDED BASIC LANGUAGE, SAVAGE TIME WAS CUT TO 32.8 SECS.
SO THIS SYSTEM MUST COST A SMALL FORTUNE RIGHT??
< EL WRONGO >
TRY $1600 US (ABOUT 1000 POUNDS).
THE A300 LINE IS, ACORDING TO THE ARTICLE, TARGETED FOR THE
HOME AND EDUCATION MARKET. THE A400 LINE WHICH INCLUDES MORE
MEMORY, FAST HARD DRIVES AND EXPANSION SLOTS, IS TARGETED MORE
FOR BUSINESS AND PROFESSIONAL USE.
(YOU REALLY SHOULD SEE THIS ARTICLE - I'M NOT KIDDING)
AND FOR YOU COPROCESSOR HATER'S OUT THERE - IT HASNT GOT ANY.
ALL GRAPHICS, SOUND AND I/O ARE PERFOMED STRICTLY BY "ARM POWER"
(WITH A LITTLE HELP FROM GS TYPE GLU'S)
AND AS FOR THE ARM - "IT'S DESIGN DELIBERATELY RESEMBLES A
6502 BROUGHT UP TO DATE WITH SHORT, FAST INSTRUCTIONS AND A
SUPERFAST INTERRUPT RESPONSE TIME" THE ARM CAN ALSO ADDRESS 64
MEGABYTES OF RAM.
DROOL OVER THAT FOR A WHILE.
MARK ORR | "YOU AREN'T GOING TO
| BE HAPPY UNTIL YOU
$CSB211@LSUVM.SNCC.LSU.EDU | CUT YOURSELF WITH IT!!!"
@LSUVM (BITNET) |
|jm7e+@andrew.cmu.edu (Jeremy G. Mereness) (05/07/90)
> Excerpts from netnews.comp.sys.apple2: 6-May-90 why wait for wdc or > asic? 3.. mark orr@LSUVM.BITNET (3050) > the chip that i'm referring to is the arm - acorn risc machine. > it is a risc chip with an instruction set deliberately designed > around the 6502 - actually the instruction set is a superset of > the 6502. it was designed for the archimedes a310 computer in > england, which was previewed in the october 1987 issue of byte. various spoo about how this thing outperforms a Compaq '386 and a Mac SE with a Hypercharger (accelerator?).... and a price of about $1600 loaded and there is a souped up model made for businesses..... > Excerpts from netnews.comp.sys.apple2: 6-May-90 why wait for wdc or > asic? 3.. mark orr@LSUVM.BITNET (3050) > AND AS FOR THE ARM - "IT'S DESIGN DELIBERATELY RESEMBLES A > 6502 BROUGHT UP TO DATE WITH SHORT, FAST INSTRUCTIONS AND A > SUPERFAST INTERRUPT RESPONSE TIME" THE ARM CAN ALSO ADDRESS 64 > MEGABYTES OF RAM. Yow! Is this thing still produced? Can anyone get a look at the thing? Any specs about registers, clock speeds, data busses, that kind of thing??? Would 65816 binaries completely bomb with this thing? ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ ^^^ |Jeremy Mereness | Support | Ye Olde Disclaimer: | |jm7e+@andrew.cmu.edu (internet) | Free | The above represent my| |a700jm7e@cmccvb (Vax... bitnet) | Software | opinions, alone. | |staff/student@Carnegie Mellon U. | | Ya Gotta Love It. | ------------------------------------------------------------------------- --
ggray@wpi.wpi.edu (Gary P Gray) (05/07/90)
In article <9005061648.AA00952@apple.com> $CSB211@LSUVM.BITNET (mark orr) writes: >anyway, the title of this article is correct. thirty-two bit >6502's already exist (and have existed since late '87) > >the chip that i'm referring to is the arm - acorn risc machine. >it is a risc chip with an instruction set deliberately designed >around the 6502 - actually the instruction set is a superset of >the 6502. it was designed for the archimedes a310 computer in >england, which was previewed in the october 1987 issue of byte. I know two hardware hackers who had gotten ahold of the specs for this thing. They were mighty impressed(a 32 bit risc processor for ~$35) Now, had Apple used this (or is going to use this) for a new //, watch out! I don't know offhand what the MIPS for this chip are, but it is certainly fast enough for any application a low to mid level machine would do. As for it's 6502ish design, I'd say they took inspiration and not much else from the chip (actually, one of the aforementioned hardware hackers jokingly referred to the 6502 as a slow RISC machine with few registers :). The memonics are not taken directly from the 6502 to the best of my recollection (they are your generic low level assembly commands, load, store, and, or etc.) There is a few weird things, like each instruction has a conditional test in it. From what I understand, you have instructions like "transfer R1 to R2 if z bit is set." It's a 32 bit processor, but it's address range is 28 bits (oh gee, *how* would I ever fit my programs in ) UN*X has been developed for this system, so there are development tools. Now, should Apple choose to develop with this and build up a simple and straight forward hardware base... Anybody have info on the Acorn computer that uses this? -- -- WARNING!!! The above opinions may be HAZARDOUS or FATAL if swallowed!!! -- "My mind is going... I can feel it... | Gary Gray -- ggray@wpi.wpi.edu I can feel it..." | GEnie: GGRAY6 (This space unintentionally left blank)
north@Apple.COM (Don North) (05/08/90)
In article <9005061648.AA00952@apple.com> $CSB211@LSUVM.BITNET (mark orr) writes: > >anyway, the title of this article is correct. thirty-two bit >6502's already exist (and have existed since late '87) > >the chip that i'm referring to is the arm - acorn risc machine. >it is a risc chip with an instruction set deliberately designed >around the 6502 - actually the instruction set is a superset of >the 6502. it was designed for the archimedes a310 computer in >england, which was previewed in the october 1987 issue of byte. > >the archimedes was designed to be a high-performance replacement >of the acorn computer (bbc micro) which uses the 6502 as its main >processor. it is also operating system compatible with the bbc micro. > >AND AS FOR THE ARM - "IT'S DESIGN DELIBERATELY RESEMBLES A >6502 BROUGHT UP TO DATE WITH SHORT, FAST INSTRUCTIONS AND A >SUPERFAST INTERRUPT RESPONSE TIME" THE ARM CAN ALSO ADDRESS 64 >MEGABYTES OF RAM. I think you've mistaken 'cultural' with 'binary' compatibility. In no way is the ARM a binary compatible superset of the 6502/65816 processor series (ie, like the 65816 is to the 6502). It's architecture does reflect some interesting optimizations/tradeoffs in the instruction set design to achieve very good performance (ie, >1K dhrystones/MHz clock rate). Radius uses this chip in their MacII/NuBus graphics accelerator to good effect. The Acorn Archimedes machine (for which the CPU was originally designed) is also a pretty hot performing box. VLSI Technology sources the chip in the USA under the part number VL86C010. Even more interesting is the follow-on part VL86C020; this part principally adds an on-chip 4K byte cache to allow building reasonable cost 20MHz+ systems, without requiring sub-50ns DRAMs. -- Don North ----- Apple Computer, Inc. ----- Advanced Technology Group UUCP: ...!{voder,nsc,decwrl,sun}!apple!north CSNET: north@Apple.COM {{ Facts are facts, but any opinions expressed are my own, and *do not* }} {{ represent any viewpoint, official or otherwise, of Apple Computer, Inc.}}