4 ms·

Had a number of things gone differently Commodore might have been Apple. Both the C64 and the Amiga were way ahead of their time both in terms of raw performanc

by api 1y ago

Had a number of things gone differently Commodore might have been Apple. Both the C64 and the Amiga were way ahead of their time both in terms of raw performance and, for the C64, price/performance at least when it first came out. I learned to program on a C64 and still fondly remember it as an amazing gateway machine into computing. Was great for games too, better than most consoles of the day. Unfortunately the 64, like all those 8-bit machines, was a technical dead end, and by the time the Amiga got momentum PC clones were eating the entire industry. PC clones killed everything but Apple, which barely clung to life through the 90s, and some Unix workstations in the high end market. It just wasn’t possible to compete with the price cuts and CPU performance gains that came with volume and scaling. (I remember in the early 90s a lot of doubts about whether x86 could be made as fast as Sparc or Alpha or other things, but Intel and later AMD did it… especially when it came to price/performance.) In retrospect Amiga might have competed there had it gone higher end and been a Unix-like OS underneath.

LoganDark 1y ago
> (I remember in the early 90s a lot of doubts about whether x86 could be made as fast as Sparc or Alpha or other things, but Intel and later AMD did it… especially when it came to price/performance.) And now Apple is in the process of beating x86 to a pulp with ARM.
- kevin_thibedeau 1y ago
  Going full circle using a team with Alpha heritage.
- api 1y ago
  The thing that made x86 take over the world was the immense scale and volume it got from the PC clone and small PC-based server (NT and early Linux) market. It allowed huge R&D investments and let x86 eat more expensive workstation and server CPUs from the bottom up, eventually conquering the server and then cloud. ARM got the same thing from the mobile phone market, allowing it to eat x86 from the bottom up in exactly the same way. It went from phones and IoT, to laptops and small servers, and now to large cloud servers. ARM does have a small advantage from its ISA being much easier to decode, especially with a wide decoder, but x86 engineers have shown that they can stay competitive.
II2II 1y ago
> Unfortunately the 64, like all those 8-bit machines, was a technical dead end I suspect that most of them were dead-ends for business reasons rather than technical reasons. More often than not, they were in a race to the bottom to try to sell to home users. Compatibility between generations, when it did exist, was often an afterthought. Commodore owning MOS probably didn't help either. A 16-bit successor to the 6502 was eventually introduced, in the same year as the 386, and the 65816 didn't even come from Commodore/MOS. I'm not sure what the story behind the other popular 8-bit processor is, i.e. the Z80. It has a 16-bit successor in 1979, but I've rarely heard of it's use.
mrandish 1y ago
> Had a number of things gone differently Commodore might have been Apple. As a die-hard Amiga owner and fan, I've pondered this for many years but have come to the sad conclusion the mythical "handful of things" which, if done differently, would have meaningfully changed the fate of the Amiga simply never existed. The key reason is that many of the same things which made the Amiga so extraordinary in 1985, like custom graphics co-processors and designing the system around analog video standards also made the Amiga uniquely problematic to migrate to a RISC architecture like Power PC in the 90s. At the end, Commodore's unannounced (but in progress) plan was to migrate the Amiga platform to an HP RISC CPU, abandon analog video timing for 1280 x 960 digital RGB displays and run old Amiga software via emulation. I'm highly skeptical any such emulation would have worked sufficiently well to be considered "an Amiga." Even today, the authors of the incredibly advanced WinUAE Amiga emulator are still discovering subtle timing issues after 20 years of development. If the "new Amiga" platform couldn't run at least 90% of 68K Amiga games at 100% speed then it's not really an Amiga and instead a new and different computer platform. And if it's pretty much a new platform anyway, why not consider the PC or Mac? Apple only barely managed to survive the 90s transition and the Mac was nearly a 'plain vanilla' platform (no custom co-processors, no brilliant analog timing hacks, etc) so it was much easier to emulate on a RISC platform than the Amiga would have been. In 1993 a 68020-based Amiga could emulate a 68000-based Mac entirely in software and run PageMaker at nearly 100% speed, whereas a 1992 68020 Mac emulating an Amiga in software would have been unimaginable. So, we're left with a Catch 22. If we go back in time to the 80s and change the things about the Amiga platform necessary for it to survive the 90s industry transitions, we're also undoing the things that made the Amiga so special when it launched. Surviving the apocalypse of the 90s required surviving three separate 'extinction-level' asteroid impacts: 1. CISC -> RISC, 2. Analog video -> Digital video, and 3. Vertically integrated computer company -> Ecosystem of component vendors (mid-90s PCs had multiple vendors competing to be the CPU, operating system, graphics card and sound card). Neither Commodore nor Atari could have survived all three of those 'planet killers'. Even Apple WITH Steve Jobs, more money, a bigger, better brand and an emergency bankruptcy-preventing $150M loan from Bill Gates - barely survived. The first asteroid also took Motorola themselves out of CPUs (even Intel almost didn't get RISC-emulating-x86 CISC microcode working in time) and the third asteroid changed the entire playing field, causing even IBM to retreat from desktop. (and some would argue Apple only existed on life-support with the candy iMacs keeping the lights on until the iPod/iPhone really put the company back on its feet.)
rasz 1y ago
> It just wasn’t possible to compete with the price cuts and CPU performance gains that came with volume and scaling. Commodore could if it tried, sadly Commodore was all about milking it. They were buying 68000 CPUs from Motorola at something ridiculous like $2 in mid eighties! Very slim margin over manufacturing cost thanks to good negotiations (owning foundry means you know how much it really cost to make chips, multiple 3rd party sources available). By 1991 Commodore was selling $100 BOM A500 for $300 retail. Amiga was printing money just like C64 before it, all with close to zero R&D cost. And then suddenly it stopped in 1992 https://cdn.arstechnica.net/wp-content/uploads/2017/01/commodore-sales-modified.png https://cdn.arstechnica.net/wp-content/uploads/2017/01/commo... 1992 was the year of market being flooded with cheap 386/VGA clones running Win3.1 thanks to Intel lost AMD Am386 lawsuit forcing wide price drops at the low end. i386SX-25 went from Q1 1990 $184 to Q4 1992 $59, AM386DX-40 Q2 1991 $231 to Q1 1993 $51. Brand name 386@40 + cache, 4MB ram, 170MB HDD, dual floppy, SVGA + Monitor, Win 3.1, keyboard + mouse systems were ~$1500 while 386SX mom&pop builds started below $1000. Turns out there are limits to how long you can sell same low end computer by just dropping price. C64 lasted impressive !12! years with majority of sales between 1983-1990, $600 down to $150 (afair at the end manufacturing cost was something silly like $30). Amiga 6 years with no upgrades $1000 to $300 + 2 more bumbling with fail half upgrades until bankruptcy. Btw in 1990 68000 CPUs were finally $3 closing in on crazy Commodore negotiated price from 1985. Enabled SEGA to sell Genesis at $100-150 retail. Prices according to "Semiconductor procurement worldwide" 1991-1993 by Dataquest.
- icedchai 1y ago
  Yep. Coincidentally I got my first 386 clone in 1992, and eventually started playing around with early Linux distros. I still had my Amiga, but it was obvious the writing was on the wall when you could get a decent 386SX for the price of an A1200.

by api 1y ago

LoganDark 1y ago

> (I remember in the early 90s a lot of doubts about whether x86 could be made as fast as Sparc or Alpha or other things, but Intel and later AMD did it… especially when it came to price/performance.) And now Apple is in the process of beating x86 to a pulp with ARM.

kevin_thibedeau 1y ago
Going full circle using a team with Alpha heritage.
api 1y ago
The thing that made x86 take over the world was the immense scale and volume it got from the PC clone and small PC-based server (NT and early Linux) market. It allowed huge R&D investments and let x86 eat more expensive workstation and server CPUs from the bottom up, eventually conquering the server and then cloud. ARM got the same thing from the mobile phone market, allowing it to eat x86 from the bottom up in exactly the same way. It went from phones and IoT, to laptops and small servers, and now to large cloud servers. ARM does have a small advantage from its ISA being much easier to decode, especially with a wide decoder, but x86 engineers have shown that they can stay competitive.

II2II 1y ago

> Unfortunately the 64, like all those 8-bit machines, was a technical dead end I suspect that most of them were dead-ends for business reasons rather than technical reasons. More often than not, they were in a race to the bottom to try to sell to home users. Compatibility between generations, when it did exist, was often an afterthought. Commodore owning MOS probably didn't help either. A 16-bit successor to the 6502 was eventually introduced, in the same year as the 386, and the 65816 didn't even come from Commodore/MOS. I'm not sure what the story behind the other popular 8-bit processor is, i.e. the Z80. It has a 16-bit successor in 1979, but I've rarely heard of it's use.

mrandish 1y ago

> Had a number of things gone differently Commodore might have been Apple. As a die-hard Amiga owner and fan, I've pondered this for many years but have come to the sad conclusion the mythical "handful of things" which, if done differently, would have meaningfully changed the fate of the Amiga simply never existed. The key reason is that many of the same things which made the Amiga so extraordinary in 1985, like custom graphics co-processors and designing the system around analog video standards also made the Amiga uniquely problematic to migrate to a RISC architecture like Power PC in the 90s. At the end, Commodore's unannounced (but in progress) plan was to migrate the Amiga platform to an HP RISC CPU, abandon analog video timing for 1280 x 960 digital RGB displays and run old Amiga software via emulation. I'm highly skeptical any such emulation would have worked sufficiently well to be considered "an Amiga." Even today, the authors of the incredibly advanced WinUAE Amiga emulator are still discovering subtle timing issues after 20 years of development. If the "new Amiga" platform couldn't run at least 90% of 68K Amiga games at 100% speed then it's not really an Amiga and instead a new and different computer platform. And if it's pretty much a new platform anyway, why not consider the PC or Mac? Apple only barely managed to survive the 90s transition and the Mac was nearly a 'plain vanilla' platform (no custom co-processors, no brilliant analog timing hacks, etc) so it was much easier to emulate on a RISC platform than the Amiga would have been. In 1993 a 68020-based Amiga could emulate a 68000-based Mac entirely in software and run PageMaker at nearly 100% speed, whereas a 1992 68020 Mac emulating an Amiga in software would have been unimaginable. So, we're left with a Catch 22. If we go back in time to the 80s and change the things about the Amiga platform necessary for it to survive the 90s industry transitions, we're also undoing the things that made the Amiga so special when it launched. Surviving the apocalypse of the 90s required surviving three separate 'extinction-level' asteroid impacts: 1. CISC -> RISC, 2. Analog video -> Digital video, and 3. Vertically integrated computer company -> Ecosystem of component vendors (mid-90s PCs had multiple vendors competing to be the CPU, operating system, graphics card and sound card). Neither Commodore nor Atari could have survived all three of those 'planet killers'. Even Apple WITH Steve Jobs, more money, a bigger, better brand and an emergency bankruptcy-preventing $150M loan from Bill Gates - barely survived. The first asteroid also took Motorola themselves out of CPUs (even Intel almost didn't get RISC-emulating-x86 CISC microcode working in time) and the third asteroid changed the entire playing field, causing even IBM to retreat from desktop. (and some would argue Apple only existed on life-support with the candy iMacs keeping the lights on until the iPod/iPhone really put the company back on its feet.)

rasz 1y ago

> It just wasn’t possible to compete with the price cuts and CPU performance gains that came with volume and scaling. Commodore could if it tried, sadly Commodore was all about milking it. They were buying 68000 CPUs from Motorola at something ridiculous like $2 in mid eighties! Very slim margin over manufacturing cost thanks to good negotiations (owning foundry means you know how much it really cost to make chips, multiple 3rd party sources available). By 1991 Commodore was selling $100 BOM A500 for $300 retail. Amiga was printing money just like C64 before it, all with close to zero R&D cost. And then suddenly it stopped in 1992 https://cdn.arstechnica.net/wp-content/uploads/2017/01/commodore-sales-modified.png https://cdn.arstechnica.net/wp-content/uploads/2017/01/commo... 1992 was the year of market being flooded with cheap 386/VGA clones running Win3.1 thanks to Intel lost AMD Am386 lawsuit forcing wide price drops at the low end. i386SX-25 went from Q1 1990 $184 to Q4 1992 $59, AM386DX-40 Q2 1991 $231 to Q1 1993 $51. Brand name 386@40 + cache, 4MB ram, 170MB HDD, dual floppy, SVGA + Monitor, Win 3.1, keyboard + mouse systems were ~$1500 while 386SX mom&pop builds started below $1000. Turns out there are limits to how long you can sell same low end computer by just dropping price. C64 lasted impressive !12! years with majority of sales between 1983-1990, $600 down to $150 (afair at the end manufacturing cost was something silly like $30). Amiga 6 years with no upgrades $1000 to $300 + 2 more bumbling with fail half upgrades until bankruptcy. Btw in 1990 68000 CPUs were finally $3 closing in on crazy Commodore negotiated price from 1985. Enabled SEGA to sell Genesis at $100-150 retail. Prices according to "Semiconductor procurement worldwide" 1991-1993 by Dataquest.

icedchai 1y ago
Yep. Coincidentally I got my first 386 clone in 1992, and eventually started playing around with early Linux distros. I still had my Amiga, but it was obvious the writing was on the wall when you could get a decent 386SX for the price of an A1200.