Hacker News new | past | comments | ask | show | jobs | submit login
AppleCrate II: A New Apple II-Based Parallel Computer (2015) (michaeljmahon.com)
111 points by aresant on Nov 10, 2020 | hide | past | favorite | 34 comments



Amazing experimentation here. I sometimes wonder what computers from parallel timelines look like, and this is a fun what-if possibility.


What if Apple did not invent the Macintosh? They would just parallel Apple II systems together for more speed?


No. New cheap microprocessors were being released constantly that were much faster than what was in Apple II, it was clear to everyone that the future was in new machines with one of the better chips.

The one big mistake Woz made when designing Apple II was picking the MOS 6502 CPU. It made sense at the time, because for the price it was the technically best chip available. The problem was that the company that made it was in serious trouble for multiple reasons.[0] Soon after the Apple II was released, MOS had to accept a low bid buyout offer from Commodore to not go bankrupt, and as Commodore CEO Jack Tramiel absolutely loathed to spend money on literally anything, this basically ended future development of the line.

[0](Mainly, terrible IP practices. MOS was founded because a Motorola engineer recognized a huge market niche that Motorola was poorly serving, and despite repeated pleas to the management, they chose not to do anything. So he set out with a few co-conspirators to found a new chip company to serve that market. However, he chose to design his chip to be way too similar to 6800 that Motorola was already making, and immediately after release Motorola sued them for stealing the design. During discovery, it was found that at MOS there were internal technical manuals related to the Motorola chip that were Motorola property and that the founders took with them when they left. Motorola won, took basically all of the profit that MOS had made, but allowed the company to continue operating.)


> The one big mistake Woz made when designing Apple II was picking the MOS 6502 CPU

It was the cheapest chip on the market. It was faster than its competition, and also used by Atari, Commodore and, later across the Atlantic, on the Acorn/BBC computers. It's one of the most popular 8-bit architectures. While Commodore never invested much in upgrading the 6502 into 16 and 32 bits, they invested a lot in other developments. Too little and too late, the 65CE02-derived chip in the C65 was quite impressive for an 8-bit CPU.


It was the cheapest chip, and way faster than the competition, but the company making it had already been sued and had a very dubious future. In a pure short term technical sense, it was the correct pick. But considering the bigger picture it was dubious.

> While Commodore never invested much in upgrading the 6502 into 16 and 32 bits, they invested a lot in other developments. Too little and too late, the 65CE02-derived chip in the C65 was quite impressive for an 8-bit CPU.

Not during Tramiel's reign they didn't. The man was famously averse to spending a single cent he didn't have to. He didn't believe in the concept of budgets, so he preferred to have to personally okay every single expenditure. If that caused work to stall when he wasn't available to do so, well, that was just the cost of doing business...

When he bought MOS, with it he got perhaps the world's best silicon design team. He promptly chased them away by denying them a bonus they felt they deserved, and after that the only development on the CPU line was for cost-cutting. It was only after he stormed off in 1984 that Commodore got back into CPU development.


Not only that, but it was very easy to program. With experience, you could read a hex dump.


True. I still remember LDA from memory is AD, NOP is EA and JMP is 4C...


No, they would have invented something faster that wasn't a Mac. Perhaps an evolution of the //GS.


The IIgs came out in 1988, four years after the Macintosh. As such, GS/OS was pretty heavily influenced by the Mac user interface; it was practically a demake. It still had some unique quirks, though -- for instance, it used proportional scrollbars, which wouldn't show up on the Macintosh until 1997 (in Mac OS 8).


Begging pardon, 1986, two years after the Mac. It also beat the Mac to market with a color GUI and ADB keyboards and mice. But yup, hardware-wise it was kneecapped from the start (1MHz IO bus, 65816 cpu would never rival m68k, low cpu clock speed) entirely for marketing reasons.


Didn't it only slow down for slot-based IO? Or did it slow down to 1MHz for the built-in IO devices too?


Yup, the built-in graphics RAM was 1MHz as well; It couldn't redraw its entire screen inside a single refresh (although by remapping the 65816 stack to screen RAM, it got close enough to do most of it). The integrated disk controller (IWM) also ran at 1MHz.


Such a shame. Video RAM could be mapped to dual-port memory and run at any speed on both sides. :-(

I really don't like the IIgs's design.


Yes, if the Mac never existed then the //GS and GS/OS would have looked quite different but they still would have existed in some form. With no Mac upgrade path, it's likely that all the different Apple II form factors would have been upgraded with the 65816 to give the platform more life. Then a next generation "Apple IV" system might have combined a 68K and 65816 to offer 32-bit performance and Apple II backwards compatibility like the PS2 having a PS1 built in. (Apologies rbanffy, I wrote this before seeing that you posted almost the same idea earlier.)


Not only 4 years after, but it had a slower 65816 to make sure it wasn't faster than a Mac.


We know that Apple was investigating the successor to the 65816 as used in the then current //GS https://mirrors.apple2.org.za/apple2.caltech.edu/miscinfo/65...


I often wonder what a 65832 (or 65864) would look like. For instance, I'd add instructions treating all registers as floating point (fp32 and fp64). And have full 32/64 bits data buses.


Wonder no more! Here's a preliminary datasheet.

http://archive.6502.org/datasheets/wdc_w65c832_preliminary_m...

The part was never produced, in large part because Apple wasn't interested in continuing the IIgs line.


In this case, it's better to imagine what could have than to see what it would have been. That's basically the 65816 documentation but 16 is find/replaced with 32. Some of time. A, X, and Y are 32-bit but the data bus is still 8-bit, address space is still 24-bit, stack, direct page, and program counter are still 16-bit.


One of the first things I'd throw out is binary compatibility. Making it source-compatible at the assembly level would be enough and would free a lot of possibilities that being attached to old code wouldn't.


They'd probably make a cut-down Lisa (which the Mac, conceptually, at least, is), or an Apple IV (as in "Apple IV, The Apology"), compatible with both II and III, but offering the full III hardware for the Apple II side. To effectively use parallel processors they'd need an SMP-capable multitasking OS, which was not trivial back then.


I mean, that's not doing anything that IBM wasn't already doing. And they chose to gamble on Steve Jobs' vision than make a "me-too product".

That's how you make the small things, just, unforgettable.


Mac is totally change the world. Ibm is the me-too product of Apple II. Not sure about what you said.


The PC, with the successor of the 8085 inside running the successor of CP/M (in the form of PC-DOS) was a me-too of the S-100 CP/M desktop computer, not the Apple II.


The reference for the PC team at IBM was the Apple II running CP/M on a Microsoft SoftCard Z80 board as this was the most popular small business computer in early 1980. It is likely that this was an influence on the PC's expansion cards that look like larger Apple II cards more than like S-100 cards. It certainly was the reason why they thought they could license CP/M from Bill Gates.


> The reference for the PC team at IBM was the Apple II running CP/M on a Microsoft SoftCard Z80

Then they learned nothing.


CP/M from Gary Kildall.


The Bill Gates we needed.



I never understood how the Mac got off the ground. The first ones were awful.

In 1995 I was given the task of making some posters (or something like that) on a Macintosh. I think it was an SE. The thing would constantly pop up error messages with codes: Error 14345. And nothing else.

I had a nightmare in which popups would open faster than I could close them.

I had used Fleet Street Editor on a BBC Micro. It was slow and laborious, but much more solid to use.


Which software did you use for that?

We also had an SE in the early 90s, it didn't seem more prone to crashes than any other computer I've ever used.

I still think they should've named the 2nd gen iPhone SE the SE/30, BTW.


I don't recall. It was too long ago and it was a one-off job.


Incredible. Woz would be proud!


He probably is.




Consider applying for YC's W25 batch! Applications are open till Nov 12.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: