AppleCrate II: Apple II-Based Parallel Computer 126
sproketboy noted that many years ago
Michael J. Mahon built the AppleCrate — a parallel stack of Apple IIs — for no good reason. Recently he came back with the AppleCrate II, which more than doubles the number of motherboards, and at least triples the awesomeness.
yes but... (Score:1)
will it run quake?
Re: (Score:3)
I'm even wondering if this thing runs faster than my cellphone...
Re: (Score:2)
I'm even wondering if this thing runs faster than my cellphone...
1MHz * 17?
Nope.
Re: (Score:2)
I think not. But I wouldn't be surprised if it'll update your facebook status.
Re: (Score:2)
If you have the same T-mobile HTC that my girlfriend has... the answer is "hell yes".
Re: (Score:2)
I'm even wondering if this thing runs faster than my cellphone...
Even? The original Quake ran OK, but not great on a P60. On a P133, it ran well. On a P133 with a Riva128 (yeah, so I couldn't justify a Voodoo) it ran great. Have you any idea how much faster your cellphone is?
Re: (Score:2)
Imagine a Beowulf cluster of these....
Re: (Score:2)
Imagine a Beowulf cluster of these....
Hehe.
Re: (Score:2)
The Apple IIGS ran Wolfenstein 3D, so... maybe.
Re: (Score:1)
My name is finally appropriate (Score:3, Funny)
Apple II == 6502 CPU from Commodore Semiconductor
I'd sooner have an Apple IIgs stack however (with its 16 bit 65000). Same ease-of-use as the original 8 bit computer, but operates about six times faster, and has a Mac-style OS.
Re: (Score:3)
16-bit 65000...is that like, 3000 less than the Motorola 68000? Did Apple get a discount, or just not use the extra transistors?
And if you were going for M68000-based machines, why not the Atari ST or Commodore Amiga? (Nothing like dredging up decades-old flame wars. :-)
Re:My name is finally appropriate (Score:4, Informative)
It was actually the 65c816, which was a 16-bit version of the 6502, completely unrelated to the m68k, and binary-compatible with the 6502 so the IIgs could still run old Apple II software.
Re: (Score:2)
Ah! I had totally forgotten about that beast. That was so long ago my memory of the 8-bit era is a bit pixelated.
Re: (Score:1)
Re: (Score:2)
>>>16-bit 65000...is that like, 3000 less than the Motorola 68000? Did Apple get a discount?
Funny, but your joke is actually truthful. 6500 processors originated at Motorola as a cheap alternative to the 6800 ($25 versus $300), but when Motorola tried to kill the project, the engineers spun-off their own company. In fact before today's "everything intel" or everything x86 world, multiple lines of CPUs existed:
Commodore's 6502/8502 series (used in 90% of the 8 bit computers/game consoles)
Zilog's
Re: (Score:2)
AAAAGHHH! (Score:3)
The 6800 came from Motorola. The 6502 was the successor the the 6501, both of which came from MOSTEK. Commodore may have had cpus contracted out, but the 6500 came nearly a decade before commodore hit the bigtime. MOSTEK was one of the leaders at the time because they offered experimenters a $20 "kit" that included a manual AND a 6501 CPU chip. This price was phenomenal at the time. The 6502 was the cpu that powered the KIM and SYM microcomputer kits, which were also very capable and very affordable - at $2
Re: (Score:3)
Re: (Score:2)
>>>The 6502 was the successor the the 6501, both of which came from MOSTEK. Commodore may have had cpus contracted out, but the 6500 came nearly a decade before commodore hit the bigtime.
Stop spreading incorrect information.
- The 6501 was developed by Motorola engineers.
- Motorola didn't want to sell it, so the engineers created a spinoff company: MOS.
- MOS became Commodore Semiconductor in 1976 (not 86 as you incorrectly implied).
Re: (Score:2)
The Commodore and Zilog lines never went higher than 16 bit.
There was the Zilog Z8000 which was as 32bit as the Motorola 68000. It had a 16-bit ALU with ability to merge adjacent registers into 32bit or 64bit values. The Z80000 (aka Z320) was a true 32-bit processor. The Motorola 68000 had 32 bit registers, but a 16-bit ALU.
Assigning the 6502 to Commodore is a misnomer. The original 6502 was made by MOS Technology and licensed to Rockwell and Synertek. MOS Technologies was later purchased by Commodore. The 65C02, 65C802 and the 65C816 were designed by Weste
Re: (Score:1)
The 68000 had a 16-bit ALU, but from the 68020 onward the series had a 32-bit ALU
http://en.wikipedia.org/wiki/68000_family [wikipedia.org]
Re: (Score:2)
Of course, the Z80 had a 4-bit ALU, IIRC.
And, the CMOS 65xx variants were designed and (I believe) manufactured by Western Design Center, not Western Digital Corporation, as well as manufactured by various second sources. (Well, that's not completely true - the W65C816S was designed by Sanyo, because the original W65C816 sucked ass, and Sanyo had to redesign it to fix the bugs, when they needed a 65816 for something they were designing. But, this was very late into the IIGS's life, and too late for it to be
Re: (Score:2)
>>>Assigning the 6502 to Commodore is a misnomer.
Not really.
Why do you think Commodore could sell their C64s for just $200?
Answer: Because they were able to get the 65XXs and 85XXs for FREE. That gave them a competitive advantage over atari, apple, et cetera who were forced to pay full price. It's a key reason why the C64 sold 30 million units: because it was cheaper than the others.
Re: (Score:2)
Re:My name is finally appropriate (Score:5, Informative)
No, the 6502 was designed by Motorola and manufactured by MOS Tech. Besides, he's using enhanced 2e's which sported a 65c02 which was designed by Western Design Center...
Re:My name is finally appropriate (Score:4, Informative)
More correctly, the 6502 was designed by Chuck Peddle and Bill Mensch. Both engineers were working for Motorola, but the 6502 was an underground project. Management didn't approve of their efforts to create a cheaper version of the 6800. They left, and started working for MOS to finish the design.
Re: (Score:1)
Re: (Score:2)
No, the 6502 was designed by Motorola and manufactured by MOS Tech.
Almost, but not quite [wikipedia.org]. The 6501 was designed by a team at Motorola but Motorola management was uninterested so the entire design team resigned en masse, shopped the design around and finally MOS Technology produced it.
The 6501 was pin-compatible with the Motorola 6800, Motorola sued immediately so the not-pin-compatible, but otherwise identical, 6502 was designed and produced.
And the rest is history.
Re: (Score:2)
No your wrong. The 6502 was designed by Chuck Peddle and Bill Mench who went to MOS they came up with the 6502that after Motorla sued them for making the 6800 pin compatible 6501. MOS was later bought by Commodore in 1976.
So yes for most of Apples life they where using a Commodore CPU.
Re: (Score:2)
Re: (Score:2)
But it is still a MOS 6502. AKA a Commodore CPU. Just because it was licensed doesn't change that.
Re: (Score:2)
Re: (Score:2)
And, actually, Commodore Business Machines didn't make any computers at all until they bought MOS, and their computers before the Amiga were all descendants of MOS designs, so it could be argued that everything from the KIM-1 (which was pre-CBM), to the PET (the first CBM computer), to the C128 were all MOS machines.
(The Amiga was bought from outside of Commodore, so...)
Re: (Score:2)
>>>MOS Tech
A division of Commodore Semiconductor Group in Norristown, PA since 1975.
>>>Western Design Center.
A licensee of Commodore Semiconductor, who owned the 6502 IP. The arrangement was similar to how AMD licensed the 8088 and 80286 from intel. i.e. Second sourced.
Re: (Score:1)
Re:My name is finally appropriate (Score:4, Interesting)
Western Design Center WAS Bill Mench. I worked with him. He had the rights to the 6502 as well as the follow on 16C6502. Bright guy but a disaster to work with. The world revolved around Bill. Did not matter that all of the world semiconductor fabs had their own design rules for how THEY manufactured semiconductors. Bill designed things based upon what he thought they SHOULD be using. Made for very fun times when you had to do silly things like DRC and timing analysis. In particular race condition analysis because you.... oh.... ran your CLOCK lines in poly (aka resistor) instead of metal. Can you say "Race Condition?" or how about "Lets beat the clock!"...............
gad I thought I forgot those days....................
Re: (Score:2)
I meant the 65C816. Been too long. That chip was a MotherF*@#&*#&*@# to get working........
Re: (Score:2)
Back when I was younger I wanted to see if I could get my hands on some "new old stock" 65C802s (an '816 with external buses scaled down to a 65C02-compatible package) for a personal project. So I emailed their sales address to see if they might know of a supplier who still had some or if they perhaps had some inventory to unload.
I got a reply from Bill Mensch himself. That was an interesting conversation that finally ended with me saying something like "I *really* only want like 5 chips and the project i
Re: (Score:2)
Commodore owned MOS technologies but I have never heard it called that before
Re: (Score:2)
Blasphemy. My limited edition "Woz" IIgs has a 65816 in it.
http://en.wikipedia.org/wiki/WDC_65816/65802 [wikipedia.org]
Enjoy,
Re: (Score:2)
I think you're missing the point. If he wanted performance, he could have used a machine built in this century. Here, the goal is clearly to show what can be done with underpowered machines, so using a more powerful machine would somewhat defeat the point. Plus, using a popular machine for its day makes the experiment resonate more with more people than choosing something as obscure as the IIgs would have done. Not to mention the fact that it's got to be a whole lot easier to find old Apple IIs, since s
Re: (Score:1)
>>>the goal is clearly to show what can be done with underpowered machines, so using a more powerful machine would somewhat defeat the point.
You mean like if I bought a 45hp Lupo to prove my point that... um... not sure? I guess that I can get to work with only 1/6 gallon of fuel. (just joking)
I like to write word processing documents on my C64 just to prove that "yes these old computers did actual work". I guess this is guy is trying to do the same thing.
>>>using a popular machine for
Re: (Score:2)
Hardly "popular". Apple IIe sold what? 500,000 units?
Don't know about the IIe specifically, but the II series (of which the IIe was the most popular and longest-lived) overall sold between 5 and 6 million. The 8-bit Atari machines (which were considered fairly successful) only sold about 2 million. Moreover, the Apple II was the only machine to be successful in both the home and small business markets until the IBM PC came along. In fact, until the PC, the Apple II was the most popular microcomputer for businesses, outselling the whole CP/M family combined
Re: (Score:2)
It's estimated that there were 5-6 million Apple IIs sold in its entire run.
As of the //e's launch, there were 750,000 units sold, so subtract that to remove the ][ and ][ Plus.
That leaves the //c, which was considered only a moderate success, and the IIGS, which was gimped to keep it from competing against the Mac, and was most often used as a faster //e anyway. (And, in fact, the //e was sold for a year AFTER the IIGS was discontinued.) The //e was considered a wild success, too...
Re: (Score:2)
Hardly "popular". Apple IIe sold what? 500,000 units? Both Atari and Commodore and even the ugly green monochrome IBM PC outsold it.
Troll, troll, troll. ~1,000,000 units in 1984 alone.
Re: (Score:2)
Hey, everyone is telling me that cpu speed doesn't matter anymore, it's all about the cores, so the 8 cpu AppleCrate should be twice as fast as my 4-core desktop machine. The 17 CPU Applecrate II must be blazingly fast! I'd love to see some Javascript benchmarks for that.
So what if my desktop has 2.5GHz CPU and the Apple IIe runs at .001Ghz, more cores is what matters!
Re: (Score:1)
Apple II == 6502 CPU from Commodore Semiconductor
No, Apple ][ == 6502 CPU from MOS Technologies
Re: (Score:2)
>>>No, Apple ][ == 6502 CPU from MOS Technologies
which was purchased by Commodore in 1976.
i.e. Apple ][ == 6502 CPU from CSG.
i.e. Atari 800== 6502 CPU from CSG.
i.e. Atari VCS==6507 CPU from CSG
i.e. Nintendo ES==6502 core from CSG
"And now you know..... the rest of the story."
Re: (Score:2)
You suck. It is the 65C816 by WDC which was at least 2x faster than an 8088 and > 4x faster than a 6502 from MOS. Of course that's stock numbers because the latest 65C816s run closer to 20MHz and faster than the original 2.8MHz (2.6MHz under load). Still have my IIGS.
What's the point though now an Android phone can emulate it n-times faster.
20% cooler (Score:2)
oblig (Score:1)
Can you imagine a Beowulf cluster of these?
News for Nerds... (Score:1)
Re: (Score:1)
It's meant to be interpreted as || rather than &&. Short-circuit evaluation saves a lot of time when validating article submissions.
Re: (Score:2)
Supercomputer? (Score:1)
Isn't this like running a program on a Cyber CDC 6600, a state of the art supercomputer when it was developed, instead of a generic PC which would do the same job in less than 1/10 of the time.
Re: (Score:2)
Re: (Score:3)
Of course, nowadays the generic PC is a cluster computer. How many single-core machines do you see these days? Factor in the GPU, and you pretty much are hacking on a mid-80s vector computer.
Re: (Score:2)
Nice, but (Score:2)
I would be more impressed if he built a 6502 processor with DIY chemical processing.
Needs a single chip (Score:1)
This is nice, but I would have been more impressed by a bunch of Apple 2s made in Verilog and integrated into a single chip.
Cool but... (Score:2)
This is really cool and there are a lot of really interesting ideas but wouldn't using PC motherboards have worked just as well. Seems a waste to hack AppleIIe boards for this when they could be used to keep Apple IIs running.
I like the idea of using the offsets to create the system and NADA net is also a very interesting hack but a stack of cheap PC mother boards from EBay would be alot more powerful and would leave the AppleII boards free to be used to keep the old classics alive.
Re: (Score:2)
If you really wanted cheap and simple, a network of AVRs would be approximately what he has here. Meanwhile, I can use the Apple IIe boards to restore more Apple IIe's...
Re: (Score:2)
Except you can buy an Apple IIe motherboard that IS based on an ASIC and actually runs BETTER than the original AND uses like a tenth the power. It's also more reliable because it has, basically, one motherboard chip, one cpu, two memory chips, and not much else.
Re: (Score:1)
Out of curiosity, do you get much demand for restored Apple IIe's?
Re: (Score:2)
From the way it looks the boards were not modified at least so they could be returned to normal service someday.
If he did modify them permanently, he should be drawn and quartered.
Re:Cool but... (Score:4, Informative)
If you look at this actual website, it becomes more obvious WHY he's using Apple II boards.
#1) He's using discrete components and actual wire and solder to cobble the boards into a single computer. He uses perf-board and socketed chips to build his extra peripherals. YOU CANNOT DO THAT with modern Intel-based mobos. They are all surface mount and pretty much unhackable unless you've got some elYte equipment.
#2) The Apple II board was built by a hacker, for hackers. That makes it the obvious choice in a nutshell. This is exactly the kind of stuff Woz wanted people to do with his creation.
#3) There's no challenge to doing parallel computing with an intel mobo -- they are already coming off the shelf with 8 cores. What's the fun in that?
Re: (Score:2)
"#1) He's using discrete components and actual wire and solder to cobble the boards into a single computer. He uses perf-board and socketed chips to build his extra peripherals. YOU CANNOT DO THAT with modern Intel-based mobos. They are all surface mount and pretty much unhackable unless you've got some elYte equipment."
Depends on the motherboard. If you find some old ones that have printer ports you can use them for all sorts of bit banged IO.
Or you could us a pic or avr interfaced to the USB port if you w
Re: (Score:2)
Never seen a USB network adaptor. Now worse than an Apple.
Re:Cool but... (Score:5, Insightful)
#2) The Apple II board was built by a hacker, for hackers. That makes it the obvious choice in a nutshell. This is exactly the kind of stuff Woz wanted people to do with his creation.
Wow, kind of a reminder of how much they've changed. These days it's, "You need an Apple technician to replace the battery / hard drive / casing / logo..."
Re: (Score:1)
Wow, kind of a reminder of how much they've changed. These days it's, "You need an Apple technician to replace the battery / hard drive / casing / logo..."
Not much has changed; that was always Jobs's mantra -- it was just Woz's crew that prized hackability. That's why Jobs was in charge of sales. Hackability went out with the Lisa.
Re: (Score:1)
Wow, kind of a reminder of how much they've changed. These days it's, "You need an Apple technician to replace the battery / hard drive / casing / logo..."
That is why I haven't owned an Apple Computer since my old and faithful Apple ][e. I used that computer from around 6th grade until I was in my first year of graduate school, when I finally bought a 386DX PC. Crazy, I know.
And finally, a story where my user name makes sense...
Re: (Score:1)
But only Apple.... (Score:2)
but only Apple is going at great lengths to stop more savvy users to experiment...
Re: (Score:2)
He's using discrete components and actual wire and solder to cobble the boards into a single computer. He uses perf-board and socketed chips to build his extra peripherals. YOU CANNOT DO THAT with modern Intel-based mobos. They are all surface mount and pretty much unhackable unless you've got some elYte equipment.
His extra peripherals are some hardware tied off the serial ports and an audio mixer that takes sixteen TTL signals (in the Applecrate II) and mixes them. He also rewrote part of the boot code, you can do that on PCs too. So there's nothing you couldn't do with PC motherboards here except stack them without some kind of fancy active cooling.
There's no challenge to doing parallel computing with an intel mobo -- they are already coming off the shelf with 8 cores. What's the fun in that?
More to the point, they already come with at least one GigE interface.
Re: (Score:2)
Well, that makes plenty of sense. As a former Amigan on a budget I have a bunch of experience with PLIP, which is pretty much the same story except again on much fancier hardware and with a port which was meant for I/O which was supplied with the hardware.
Maybe I have this totally confused with some other computer, but IIRC the way you get audio out of an Apple II from Applesoft BASIC is to either PEEK or POKE the correct address, implying that it's just an address line someplace that's tied to the audio am
Re: (Score:1)
Re: (Score:2)
I sit corrected. It's nice that there were some user-addressable bits in there someplace that you could adapt to the purpose.
Re: (Score:2)
"He's using discrete components and actual wire and solder to cobble the boards into a single computer. He uses perf-board and socketed chips to build his extra peripherals. ... The Apple II board was built by a hacker, for hackers. That makes it the obvious choice in a nutshell."
That's how stuff was done back them. Processors and RAM were dropped into sockets. Boards had discrete parts. "Upgrades" meant filling in the empty RAM sockets so you had 48K of RAM and not just 16K.
Re: (Score:2)
Yes, I remember upgrading the RAM on my first computer by inserting ICs into sockets. But what I continue to find really amazing is how cheap absolutely mind-boggling amounts of memory are now. I recently upgraded a machine to 8GB of RAM for about $100; it seems like only yesterday that I could have spent several times as much money to upgrade a machine to 8MB. Likewise, I've still got some old 160kB 5.25" floppy disks hanging around... now I could easily carry many gigs of storage around on a keychain, and
Re: (Score:2)
"... it seems like only yesterday that I could have spent several times as much money to upgrade a machine to 8MB."
Several times? I once upgraded my first Macintosh II to 8MB of RAM, and I swear I paid at least $1,1000 to do so...
C64 stack (Score:3)
Would it even be possible to make a stack with Commodore 64's?
Re: (Score:2)
Wouldn't exactly be 64-bit anymore then, would it?
Re: (Score:2)
Why not? They're communicating via serial so the same kind of thing should be pretty easy to do. On the other hand, it might be niftier to utilize the expansion bus for communications to leave the serial ports free.
Re: (Score:2)
(useful technical response deleted)
This right here is why Slashdot is still great. Cowardly or not, keep it up.
Re: (Score:1)
(It is 10baseT, not based)
Space-saving (Score:1)
Well, what are you waiting for? Slap a touch screen on that puppy, and make it a 17-processor tablet computer!
Re: (Score:1)
Anybody else notice a similarity (Score:2)
love this topic (Score:2)
Typo (Score:1)
Should be: "AppleCrate I (at the time I didn't realize that it was number "I" ;-)) was great fun".
crap, now I forget the html code for sarcasm...
Imagine, if you can... (Score:1)
It makes one wonder... (Score:2)
What kind of impact would this have had if people were doing this back in 70's?
Granted, this guy is just using it mostly for audio processing. (Impressively done, though... especially if you ever experimented with audio sampling on an Apple II using self-designed software and custom-built I/O interfaces)
What I'm curious about, is whether the video output from each of thee boards could be combined into either a single high resolution display matrix approaching VGA at a low depth, or layered atop each other t
Re: (Score:2)
The boards are powered by a PC AT power supply. The average power consumed by an Apple //e board is about 4.2 watts, so the whole 17-board crate consumes only about 70 watts in total, and both the AppleCrate and the power supply run only a few degrees above ambient temperature.