The Amiga Computer Community Portal Website (2025)

The majority of A500 games kick the OS into a single-task game like a game console.

Quote:

The Amiga had all the stuff in ROM as an default HD was deemed to expensive and loading the full OS from floppy would have been way to slow

Please, read what you wrote before posting...

The first Atari ST loaded the GEM/TOS from a floppy, and Atari understand that was a big mistake, and put it into some eproms... But let me guess : you are going to rewrite history, and saying I'm a liar, right ?

Last edited by CosmosUnivers on 21-Feb-2025 at 06:26 PM.
Last edited by CosmosUnivers on 21-Feb-2025 at 05:43 PM.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

Without a physical Kickstart, the A500 don't start at all : the more complete, the better !

Lack of guaranteed 68K MMU stalled baseline AmigaOS into frozen-in-time unprotected memory OS era. Amiga's 32-bit OS is half-baked.

Last edited by Hammer on 21-Feb-2025 at 08:41 PM.

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

Putting most of the OS into ROM enabled to have "full OS" without an HD or extreme boot times from floppies. And it also saved some RAM.

Original A3000 has Kickstart 1.4 ROMS and these ROMs load Kickstart software from either floppy or hard drive into RAM and then use MAPROM.

My A3000 version had Kickstart 2.04 in ROM that was later swapped out for Kickstart 3.1 ROM.

PC IDE started in 1986 and it took a while for Commodore to implement IDE PIO for A500 class machines e.g. A300 (A600). Henri Rubin focused on a few 100,000-scale A2000 markets and a GVP-like add-on business model.

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

This thread is about 040 and 060. Most A4000-040's were used for 'serious' applications, not games.

Your mindset doesn't reflect the PC's MPC Level 2 (486SX-25, 4MB RAM) and MPC Level 3 (Pentium 75, 8MB RAM). https://en.wikipedia.org/wiki/Multimedia_PC

This topic covers PC's MPC Level 2 and MPC Level 3 equivalent for the Amiga.

Before PiStorm32 and Emu68, I have TF1260 with 68060 rev1 @ 62.5Mhz for my A1200 rev 1D1 and my primary purpose for my A1200 was not for "serious" applications i.e. play Doom-type games and MP3 music.

Quote:

People don't realize the amount of talent and hard work that goes into making a good game. Take Doom for example. If they had stuck to the original design specs it would have sucked.

EA Bull Frog's Magic Carpet targeted MPC Level 2 PC.

Core Design's Tomb Raider DOS targeted MPC Level 3 PC and it was in parallel development with PS1 and Saturn in 1994 for the 1996 release.

PC had a large install base for texture-mapped 2.5D/3D from 1992.

IndyCar Racing for PC was also released in 1993. https://en.wikipedia.org/wiki/IndyCar_Racing

From https://www.intel.fr/content/dam/doc/report/history-1994-annual-report.pdf
Intel reported the following
1. In 1994's fourth quarter, Pentium unit sales accounted for 23 percent of Intel's desktop processor volume.
2. Millions of Pentiums were shipped.
3. During Q4 1993 and 1994, a typical PC purchase was a computer featuring the Intel 486 chip.
4. Net 1994 revenue reached $11.5 billion.
5. Net 1993 revenue reached $8.7 billion.
6. Growing demand and production for Intel 486 resulted in a sharp decline in sales for Intel 386 from 1992 to 1993.
7. Sales of the Intel 486 family comprised the majority of Intel's revenue during 1992, 1993, and 1994.
8. Intel reached its 6 to 7 million Pentiums shipped goal during 1994. This is only 23 percent unit volume.
9. On revenue, the USA is Intel's home market.

By the end of 1994, Intel's Pentium PC install base crushed the entire Amiga install base of 4 to 5 million units!

From the start of PPC Macs in 1994 to Jan 1995, Apple's Mac customers purchased 1.2 million PowerMacs.

By the numbers, Intel's unified X86 PC platform is a monster when compared to the Amiga i.e. it mirrored the USA superpower military and economic might against the smaller German military and economy during WW2.

The strength of the Japanese game platforms is based on the 1980s and 1990s strength of Japan as the world's 2nd largest economy.

EU plan is about economies of scale. The same for Australia-New Zealand's Closer Economic Relations/Trans-Tasman Travel Arrangement at a smaller scale. The good old "might is right".

Last edited by Hammer on 21-Feb-2025 at 10:08 PM.
Last edited by Hammer on 21-Feb-2025 at 10:02 PM.
Last edited by Hammer on 21-Feb-2025 at 09:43 PM.

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

A1000 had a Kickstart disk.

Which meant it shipped with 256k extra RAM that did cost real money in 85 and was cumbersome to use.
Neither suitable for a A500 class computer.

Quote:

Original A3000 has Kickstart 1.4 ROMS and these ROMs load Kickstart software from either floppy or hard drive into RAM and then use MAPROM.

The A3000 had an HD as standard and the 1.4 was a "useable" OS just like any other 1.x version.

Quote:

PC IDE started in 1986 and it took a while for Commodore to implement IDE PIO for A500 class machines e.g. A300 (A600). Henri Rubin focused on a few 100,000-scale A2000 markets and a GVP-like add-on business model.

Relevant how?
An HD was pretty much standard in PCs long before IDE came along, while it never was standard in the A2000.
Third party add ons and 1st party bundles (A2500) doesn't change that.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

Hammer wrote:
@bhabbott

Quote:

This thread is about 040 and 060. Most A4000-040's were used for 'serious' applications, not games.

Quote:

Before PiStorm32 and Emu68, I have TF1260 with 68060 rev1 @ 62.5Mhz for my A1200 rev 1D1 and my primary purpose for my A1200 was not for "serious" applications i.e. play Doom-type games and MP3 music.

Anyway, thanks for the headcount. I too have a system that meets the criteria, so that's 2. But we're not going to get an accurate measure of the market like this. OneTimer1's method is a good one because it counts the number of people who have a suitable system and are actively using it.

As another example - using the Internet Archive's Wayback Machine I looked at the number of downloads for '68040/060-Loadlib fix for SysInfo4' on Aminet, uploaded on Feb 12 2020. The results were:-

Dec 19,2020 284

Jun 19 2021 562

May 28 2023 1115

Oct 8, 2024 1431

Feb 22, 2025 1572

Assuming that every download represents another 040/060 owner that means there should be at least 1500 active users. Heimdall wants to sell at least 500 copies of his game, which looks doable given that's only ~30% of the user base or less.

Quote:

Quote:

People don't realize the amount of talent and hard work that goes into making a good game. Take Doom for example. If they had stuck to the original design specs it would have sucked.

EA Bull Frog's Magic Carpet targeted MPC Level 2 PC.

Core Design's Tomb Raider DOS targeted MPC Level 3 PC and it was in parallel development with PS1 and Saturn in 1994 for the 1996 release.

Magic Carpet was crap.

Tomb Raider was an excellent game that was not part of the 1992/3 'flood' you talk about. If someone ports it to the Amiga I'll be the first to buy it whatever the price. I have run the PC demo on both my A1200 with 50MHz 030 and my A600 with Vampire V2, via emulation in PC Task. It's fast enough to be playable (just) on the Vampire! On the A1200 it looks gorgeous in AGA but plays at ~0.5 fps.

A lot of hard work and talent went into that game, and it shows.
Quote:

The team kept the project deliberately simple and comparatively modest in scope, with its development budget being approximately £440,000. The production atmosphere was fairly informal. Development began in 1994 and lasted eighteen months. The team endured excessive overtime and crunch during the last stages....

A third-person 3D action-adventure like Tomb Raider was unprecedented at the time, and the development team took several months to find a way to make Gard's vision for the game work on the hardware of the time... It is Core Design's contention that, prior to the development of Tomb Raider, they were "struggling somewhat" with 32-bit development.

Quote:

Kronos wrote:
@Hammer

Quote:

A1000 had a Kickstart disk.

Which meant it shipped with 256k extra RAM that did cost real money in 85 and was cumbersome to use.
Neither suitable for a A500 class computer.

Quote:

Original A3000 has Kickstart 1.4 ROMS and these ROMs load Kickstart software from either floppy or hard drive into RAM and then use MAPROM.

The A3000 had an HD as standard and the 1.4 was a "useable" OS just like any other 1.x version.

Whether loaded from disk or in ROM, the OS was still ROM-based in operation. The only reason for doing it differently in the A1000 and A3000 was that in both cases the final ROMs weren't ready for distribution. That's how Commodore was able to produce stable ROM versions without delaying model introduction. having the OS in ROM was a key feature that helped make the Amiga cheaper and more reliable.

Quote:

An HD was pretty much standard in PCs long before IDE came along, while it never was standard in the A2000.
Third party add ons and 1st party bundles (A2500) doesn't change that.

bhabbott wrote:

Actually it does. The vast majority of A2000s were either sold with a hard drive or soon had one added.

Define "vast majority" and "soon".

But in reality the A2000(A) was an A1000 with slots and the OS in ROM and when that turned out to be too bugged it was replaced with an A500 with slots (A2000B).

It did ship without an HD, without an HD controller on board, without a HD suitable filesystem in ROM and without a non hacky way to boot directly from HD once you added it (read Kick 1.2).

Putting half the OS into ROM was a means to get away with less RAM but it was in no way the reason for the A500s success.
C= could have just shipped a minimal ROM just looking for a bootable floppy or one containing a basic BASIC from which you would have typed "LOAD ...." C64 style.
Sales numbers in 88-91 would have been about the same as they were.

Last edited by Kronos on 22-Feb-2025 at 01:53 PM.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

That's right. Not 030.

Shapeshifter's MacOS 68K covered MS Office, Netscape, and Adobe requirements.

Quote:

As another example - using the Internet Archive's Wayback Machine I looked at the number of downloads for '68040/060-Loadlib fix for SysInfo4' on Aminet, uploaded on Feb 12 2020. The results were:-

Dec 19,2020 284

Jun 19 2021 562

May 28 2023 1115

Oct 8, 2024 1431

Feb 22, 2025 1572

Assuming that every download represents another 040/060 owner that means there should be at least 1500 active users. Heimdall wants to sell at least 500 copies of his game, which looks doable given that's only ~30% of the user base or less.

Quote:

There was a flood of crap coming out on the PC that was instantly forgettable and not worth playing.

Magic Carpet was crap.

**Later open source port to the Amiga. Most listed PC games would run on 68LC040 Amiga AGA with bitter assist C2P. 68LC040 is 80486SX equivalent.

A small minority of game/ex-demo scene programmers have ignored Commodore's official warning on slow software C2P and discovered a game usable bitter assist C2P workaround. John Carmack's 1994 statement repeated Commodore's official view on this issue.

Jeff Porter's multi-media group created this PR problem to promote the group's CD32 C2P hardware. Jeff Porter's counter-corporate politics to wreak Jeff Frank's A1200 work and the negative PR shows on Commodore's SDK documentation. The lesson: C= should hire a competent game programmer to create a bitter assist C2P workaround and make it the official position.

https://bigbookofamigahardware.com/bboah/product.aspx?id=1604

The “chunky to planar” logic was thought out in a lunchtime conversation between Beth Richard (system chip design), Chris Coley (board design), and Ken Dyke (software) over Subway sandwiches on a picnic table in a nearby park one day, because Ken was telling us how much of a pain it was to shuffle bits in software to port games from other platforms to the Amiga planar system. We took the idea to Hedley Davis, who was the system chip team manager and lead engineer on Akiko and he said we could go ahead with it. I showed him the “napkin sketch” of how I thought the logic would work and was planning on getting to it the next day as it was already late afternoon by that point. I came in the next morning and Hedley had completed it already, just from the sketch!

Commodore-driven negative PR in official Amiga 3.1 SDK documentation was later proven false with multiple Amiga AGA Doom ports. Commodore's internal politics has spilled into the public arena and published in official documentation. John Carmack's 1994 statement is not the root problem.

Commodore - The Final Years book includes negative comments on Jeff Frank's management performance.

Last edited by Hammer on 23-Feb-2025 at 02:01 AM.
Last edited by Hammer on 23-Feb-2025 at 01:51 AM.
Last edited by Hammer on 23-Feb-2025 at 12:42 AM.
Last edited by Hammer on 22-Feb-2025 at 11:37 PM.
Last edited by Hammer on 22-Feb-2025 at 11:33 PM.

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

bhabbott Quote:

As another example - using the Internet Archive's Wayback Machine I looked at the number of downloads for '68040/060-Loadlib fix for SysInfo4' on Aminet, uploaded on Feb 12 2020. The results were:-

Dec 19,2020 284

Jun 19 2021 562

May 28 2023 1115

Oct 8, 2024 1431

Feb 22, 2025 1572

Assuming that every download represents another 040/060 owner that means there should be at least 1500 active users. Heimdall wants to sell at least 500 copies of his game, which looks doable given that's only ~30% of the user base or less.

Good idea, but the dl info is available for searches and it would be better to look at a sample of programs.

name | version | date | dls
ScummVM_RTG_060.lha 2.5.1.05 2024-11-17 3515
ScummVM_AGA_060.lha 2.3.0.04 2024-07-02 9326
68060-19101999.lha 46.7 2015-03-27 7605

ScummVM for the 68060 is a good example for games. There are 3515 dls of the RTG version since November of 2024 and 9326 dls of the AGA version since July of 2024. Some of the downloaders are no doubt using emulation or FPGA hardware but still have 68060+ systems as ScummVM requires CPU performance. There have been 7605 dls of the Phase 5 68060.library since 2015 which is just for the most popular brand of older 68060 accelerators and suggests quite a few 68060 accelerator owners.

bhabbott Quote:

1.4 was buggy and not very usable, but that's not important.

Whether loaded from disk or in ROM, the OS was still ROM-based in operation. The only reason for doing it differently in the A1000 and A3000 was that in both cases the final ROMs weren't ready for distribution. That's how Commodore was able to produce stable ROM versions without delaying model introduction. having the OS in ROM was a key feature that helped make the Amiga cheaper and more reliable.

If ROM was enough cheaper than memory, it would be expected that Commodore would have been more aggressive about increasing the ROM size to save memory. The Amiga ROMs were 256kiB, 512kiB and 1MiB while some later hardware could support 2MiB ROMs which Commodore never used? The Acorn Archimedes quickly moved to 2MiB ROMs for RISC OS and there were even 4MiB ROMs despite data compression in ROM which the 68k Amiga did not use as far as I know. It is claimed that RISC OS was assembly optimized which is more likely to increases code size on ARM. Even when optimizing for size, ARM code was ~72% larger than 68k code in Vince Weaver's code density results. RISC OS did include more of the OS in ROM for quick startup to a more usable state where LoadWB was not included in ROM on the 68k to do the same. Oddly prefs settings were later saved in ENV: which used valuable memory but could have been in a recoverable drive like a RAD: drive with early settings available for LoadWB before the hard drive could spin up. The 68k Amiga could have easily been one of the early instant on systems. A few small utilities could have been added to a recoverable drive as well.

Dave Haynie said that Commodore was experimenting with flash memory which was first commercially available from Toshiba for NAND in 1987 and for NOR from Intel in 1988. The 68k Amiga used ROM like later MCUs use NOR flash which supports execution of code unlike NAND. Both NAND and NOR flash became cheap but how fast I am not aware. Flash memory was used in Phase 5 accelerators though so it was likely economical by at least the late 1990s if not the mid or early 1990s. NAND flash became very cheap, allows dense on chip memory storage and is low power which is why it is used for SSDs today even though it has special block read and write requirements instead of random access capabilities like the less dense, higher power and not scaling below about a 45 nm chip process NOR flash. The RP2350 MCU variants get around the lack of scaling below 45nm by stack a NOR die on top of the MCU die. As an alternative, I wonder if it would be possible to have a small ROM with simple decompression code that would read from NAND flash and decompress to memory which would use more memory but offers several advantages. The NAND flash is nonvolatile so bug fixed Amiga modules could be replaced. The libraries could already be made so no need for MakeLibrary() which saves startup time and all of the library could be a single memory area that allows PC relative addressing. Add a new shorter 68k (d32,PC) addressing mode for which the encoding is available and this would be efficient even for larger libraries. The library base would no longer need to be in the A6 register when calling a new style library as the library could always access itself PC relative. Maybe this would be preferable to using off die NOR flash Amiga ROMs on modern 68k Amigas? I guess the bigger question is does anyone care about Amiga development instead of cutting out the heart here and the soul there to create a Frankenstein style monstrosity out of the once elegant 68k Amiga design?

bhabbott Quote:

Actually it does. The vast majority of A2000s were either sold with a hard drive or soon had one added. It's one of the reasons you bought an A2000 rather than an A500. Many were also upgraded with an accelerator card too, typical a 25MHz 030 but in later years up to a 40MHz 040. After AGA came out RTG cards also became popular. This setup meets the requirements of the OP!

Most Amiga 2000s have hard drives and many have accelerators but most are 68030 accelerators. Many Toaster Amiga 2000 systems were upgraded to 68040 accelerators and likely many have survived. Most Amiga 2000 ZorroII RTG cards are very old and ZorroII bus bandwidth is likely too low for fast paced RTG games. Any ZorroII RTG hardware that could count as 68040+ hardware would be a bonus.

Last edited by matthey on 22-Feb-2025 at 11:31 PM.

ScummVM for the 68060 is a good example for games. There are 3515 dls of the RTG version since November of 2024 and 9326 dls of the AGA version since July of 2024. Some of the downloaders are no doubt using emulation or FPGA hardware but still have 68060+ systems as ScummVM requires CPU performance. There have been 7605 dls of the Phase 5 68060.library since 2015 which is just for the most popular brand of older 68060 accelerators and suggests quite a few 68060 accelerator owners.

Aminet download stats for 68060-V44_3.lha has 11,101.

"More than 10,000" shipped for Vampire V2 (AC68080 V2) when I was in the queue during the COVID-19 lockdown.

80,000+ for THEA500mini.

https://tooomm.github.io/github-release-stats/?username=michalsc&repository=Emu68
Emu68 v1.0.4 has 2,522 downloads

Emu68-tools 2024-07-30 has 3,185 downloads
Emu68-tools 2022-02-04 has 6,018 downloads

I use pre-made for Emu68 CoffeineOS and I wouldn't be in Github's Emu68 v1.0.4 download stats.

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

Relevant how?
An HD was pretty much standard in PCs long before IDE came along, while it never was standard in the A2000.
Third party add ons and 1st party bundles (A2500) doesn't change that.

For A500, 3rd party ICD offered low-cost AdIDE controller from 1990.

Production-ready ECS A1000Jr had IDE in 1991 and all Commodore national subsidiaries rejected it due to missing 256 color display.

https://youtu.be/S_ZEWvokp3o?t=3459
286-16 with fast VGA running A500 class 2D games.
At 2:02, running Lotus 3,
At 5:45, running Speed Ball 2,
At 14:51, running Great Courts 2,
At :57:39, running Crazy Cars 3,
PC exclusives
At 11:32 , Wolfenstein 3D (1992),
At 1:11:46 The Catacomb Abyss 3D (1992),

A300 (A600) project had extra PCMCIA-related work which delayed its release into 1992.

Commodore Germany wanted a 256-color display capable AA500 with hard disk capability ASAP.

Commodore UK's A300 for C64C replacement and Commodore Germany's hard disk capability shouldn't have been mixed up.

If there wasn't a mix-up and without anti-GVP PCMCIA, an A500 revision would have IDE.

AA3000plus's AGA R&D was frozen in June 1991 with an incomplete beta state that was restarted in Feb 1992.

Based on the published A1200's R&D time scale from Feb 1992 to May 1992 (four months with AA Gayle and Budgie R&D), AA1000plus and AA3000plus would be released in Q4 1991.

AA Gayle R&D is not required for PCMCIA less Amigas.

Budgie displaced Bridgette which in turn displaced A3000's four PLL bridge chips (A600 has two PLL bridge chips).
Budgie includes a 32-bit Ramsey memory controller function.
Budgie includes a mini-Buster function for Zorro II address space PCMCIA and A1200 CPU local edge connector.

Fat Gary + Bridgette + Ramsey is enough to get 32-bit AGA Amiga operational for AA500.

A300 (A600) project was a product of a mix-up of different requirements.

Jeff Frank's AGA R&D time scale is dead wrong!

The original AA1000plus would have one or two Zorro II slots (in place of PCMCIA) and one local CPU slot.

Last edited by Hammer on 23-Feb-2025 at 01:44 AM.

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

Hammer wrote:
@Kronos

Quote:

Relevant how?
An HD was pretty much standard in PCs long before IDE came along, while it never was standard in the A2000.
Third party add ons and 1st party bundles (A2500) doesn't change that.

Fall 1984 – Bill Frank of WD develops the initial IDE concept and it is shown to IBM, DEC, and Compaq that winter and spring
May 1985 - Compaq Deskpro 286 announced, apparently containing a conventional AT controller and ST506 drive.
July 1985 - WD IDE business plan includes a “40 pin, single cable connection for IBM PC compatible card” and states that WD “is negotiating a custom project with Compaq Computer. The product will be a 20 megabyte Integrated Drive possibly using either a Microscience or a Tandon head/disk controller.”
February 1986 - Compaq Portable II announced, containing an drive comprising a WD controller mounted on a 3½-inch drive from MiniScribe
September-October 1986 - At Buscon EastKen Hallam of WD presented a paper "A CASE FOR DIRECT PC-BUS CONNECTION: SIDE STEPPING THE TRICKY INTERFACES" describing direct connection to AT bus extension with 40 pin connector.
Late 1986 - Compaq Deskpro 386 ships, containing an drive comprising a WD controller and a 5¼-inch CDC drive. At this time, it is likely the same drive was incorporated into an enhancement to the Deskpro 286, originally shipped in May 1985.

Here's a photo of the drive. You can see the ST506 controller board mounted on top of the hard drive controller board, with the usual ribbon cables (at the rear) connecting between them. This was effectively a standard ISA bus ST506 controller card, except with a connector for a ribbon cable going to the motherboard instead of plugging into an ISA bus slot.

Why does it matter? IDE didn't become a real standard until many years later. Meanwhile SCSI was a real standard. The Apple Macintosh Plus used it in 1986. SCSI interface chips and drives were readily available for anyone to use anywhere, unlike IDE which initially was restricted to Compaq computers.

But ST506 drives had been the 'standard' on PCs since 1982 when IBM used it on the PC-XT. By 1987 there were plenty of them around, so people started looking at ways to use them on the Amiga. The first hard drive controllers for the Amiga used SCSI to ST506 adapters. But this was an expensive way to do it, so simple ISA bus adapters were developed to use an ST506 controller, typically an OMPTI 5520 or 5527. By 1988 a number of these were being sold for the A1000/A500 and A2000. In 1989 I reverse engineered one and wrote my own driver for it, using an OMPTI card and Miniscribe 20MB drive that I bought second-hand for NZ$300. Then I made a 'bridgeboard' for my friend's A2000 so he could have the same setup. It worked very well.

So we were doing the same thing that Bill Frank of WD did. The ISA bus interface was very cheap and easy to make, consisting of nothing more than 3 standard TTL logic chips and an 8-bit ISA slot connector. This was cheaper than a SCSI interface and drive, and easier to obtain.

IDE took all the fun out of it. No more low-level formatting using a program written in BASIC, and entering bad blocks from the list printed on the drive (perhaps including some you had added if more bad blocks were found on test). I miss those days. You understood the hardware at a low level and had a personal relationship with it, and when you got it all working well there was a real sense of achievement.

Quote:

Production-ready ECS A1000Jr had IDE in 1991 and all Commodore national subsidiaries rejected it due to missing 256 color display.

Quote:

A300 (A600) project had extra PCMCIA-related work which delayed its release into 1992.

Quote:

Commodore Germany wanted a 256-color display capable AA500 with hard disk capability ASAP.

Quote:

Commodore UK's A300 for C64C replacement and Commodore Germany's hard disk capability shouldn't have been mixed up.

Quote:

If there wasn't a mix-up and without anti-GVP PCMCIA, an A500 revision would have IDE.

Quote:

AA3000plus's AGA R&D was frozen in June 1991 with an incomplete beta state that was restarted in Feb 1992.

Quote:

Based on the published A1200's R&D time scale from Feb 1992 to May 1992 (four months with AA Gayle and Budgie R&D), AA1000plus and AA3000plus would be released in Q4 1991.

Quote:

AA Gayle R&D is not required for PCMCIA less Amigas.

I love the way you think you know everything about chip design, and would have done a much better job with your 33 years of hindsight. Why don't you hop in your time machine and go back to 1991 and fix it for us, eh? Just one thing - if you screw up and Commodore doesn't produce the A1200, I'll be pissed.

Why does it matter? IDE didn't become a real standard until many years later. Meanwhile SCSI was a real standard. The Apple Macintosh Plus used it in 1986. SCSI interface chips and drives were readily available for anyone to use anywhere, unlike IDE which initially was restricted to Compaq computers.

1989 Acer 915V has IDE https://theretroweb.com/motherboards/s/acer-915v

1989 Commodore PC-40 III (based on Western Digital FE3400-10 chipset) has an IDE.

AA version A1000Plus with IDE in the 1990 time scale context.
From Commodore - The Final Years

Jeff Porter had initially conceived an A1000 Plus in early November
1990 with a $350.42 bill of materials using a 16-bit 14 MHz 68000
processor and two Zoro II slots. However, Augenbraun countered his
proposal with a more powerful 32-bit 16 MHz 68EC020 and a single
slot for $351.96. “My original concept was it was going to have one
expansion card so it was supposed to just have an edge card
connector on the end of the board, not the male one that an A500
had where it needed special peripherals,” explains Augenbraun. “A
female one so it could just take a regular A2000 (Zorro II) card. And
it would have a hard drive interface built in.”

However, Porter favored two Zorro II slots, based on the feedback
he received in Europe. “It was really supposed to have one card but
it grew to two cards,” says Augenbraun. “Jeff really felt strongly that
it needed two cards and I was resistant because of cost. We did a
riser card which annoyed me because it was another 20 bucks in
cost.”

The A1000 Plus would reside in a “pizza box” form factor case,
similar to the one used by the original A1000, minus the keyboard
garage. The likely inspiration came from the Macintosh LC, which
also used the pizza box form factor (which in turn might have been
inspired by the Amiga 1000). When Commodore’s regional offices
heard of the new system, they began waiting in anticipation. “The
idea was it was a baby A3000 at a delta price point over an A500,”
explains Augenbraun.

(skip)

Although Commodore placed numerous engineers on its high-end
machines, such as the A3000, the money-making low-end machines
had barely any engineers assigned to them. The A1000 Plus would
have only Augenbraun to design the PCB and gate array chip. “It
was me. That's the way Commodore worked,” says Augenbraun. “I
did the chip for it.”

However, Augenbraun was ideally suited to the task. “I am the rare
engineer who loves cost reduction,” he says. “For me it was all about
cost. One of the big battles was if the hard drive should be SCSI or
ATA (IDE)]. ATA was cheaper, it saves a chip. But there was just this
fetish for SCSI. Performance for SCSI is actually lower. But there was
sort of a brand with the Amiga at the time on SCSI.”

Quote:

IDE took all the fun out of it. No more low-level formatting using a program written in BASIC, and entering bad blocks from the list printed on the drive (perhaps including some you had added if more bad blocks were found on test). I miss those days. You understood the hardware at a low level and had a personal relationship with it, and when you got it all working well there was a real sense of achievement.

Quote:

And they got the A1200 'ASAP' (as soon as it was 'possible').

Quote:

The A3000 Plus should never have been used for anything other than a testbed for AA. The A3000 was a failure and another one wouldn't be any better.

ECS A3400 with Bridgette is A3000 without SCSI, Amber/framebuffer, and four PLL bridge chips.

A3000, CD32, A1200, and A4000 have the same 32bit CPU 7.1 MB/s max 32bit Chip RAM access behvior. CD32, A1200, and A4000 are all based on AA3000+'s AA.

2. AGA beta testing was frozen around June 1991.

3. Bill Sydnes' ECS A1000Jr is a failure since all Commodore national subsidiaries have rejected it.

4. Commodore produced 50 working A3000+.

Quote:

Dave Haynie mucked around trying to put a DSP chip in it for iffy reasons, when he could have been helping design the A1200 instead.

A1200 is Jeff Frank's managed project after Mehdi Ali directly commanded Bill Sydnes / Jeff Frank to build an AA500 class machine in Feb 1992 after Bill Sydnes / Jeff Frank resisted AGA.

Dave Haynie already stated AA3000plus's AGA R&D was frozen for "more than six months" and AA1000plus was cancelled. Try again.

ECS A1200 = Bill Sydnes' / Jeff Frank's ECS A600 and ECS A1000Jr's A2200 with 68020 projects.

AA600 = A1200 i.e. A600 with AA Galye modification + AA3000's drop-in AA + Budgie (includes A3400's Bridgette+Ramsey+mini-Buster). AA600 reverted to the cancelled AA1000plus's 68EC020 CPU and AA selection. A3400's Bridgette covers A3000's four PLL bridge chips and the A600 version has two PLL bridge chips.

Most of A1200's AA + 68EC020 componet selection is covered by the canceled AA1000plus. A1200's Feb 1992 R&D restart is back to around June 1991 state.

AA1000plus has one or two 16-bit Zorro II slots vs AA600/A1200 has one 16-bit ISA-based PCMCIA.

2MB RAM target covers 1991 MPC Level 1.

https://amiga.net.au/files/Tech_Amiga/Commodore_A4000_Pilot_Production_Release_300.pdf
A4000's pilot production release #300. A document distributed on August 28, 1992

Quote:

I love the way you think you know everything about chip design, and would have done a much better job with your 33 years of hindsight. Why don't you hop in your time machine and go back to 1991 and fix it for us, eh? Just one thing - if you screw up and Commodore doesn't produce the A1200, I'll be pissed.

Commodore The Final Year book shows the timeline for A1200's project starting from Feb 1992 to May 1992 which is four months that included AA Gayle and Budgie R&D.

Dave Haynie already stated AA3000plus's AGA R&D was frozen for "more than six months" and AA1000plus was canceled.

Quote:

An A500 'revision' did get IDE. It was called the A600.

A600 is not A500 with IDE since it's missing Zorro I and a keypad (for spreadsheets).
A600 solidified Commodore being a toy company. LOL

Last edited by Hammer on 24-Feb-2025 at 02:59 PM.
Last edited by Hammer on 24-Feb-2025 at 02:55 PM.
Last edited by Hammer on 24-Feb-2025 at 01:30 PM.

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

matthey wrote:

Good idea, but the dl info is available for searches and it would be better to look at a sample of programs.

name | version | date | dls
ScummVM_RTG_060.lha 2.5.1.05 2024-11-17 3515
ScummVM_AGA_060.lha 2.3.0.04 2024-07-02 9326
68060-19101999.lha 46.7 2015-03-27 7605

ScummVM for the 68060 is a good example for games. There are 3515 dls of the RTG version since November of 2024 and 9326 dls of the AGA version since July of 2024.

As of today there have been 9336 cumulative downloads of ScummVM_AGA. To get some idea of what that consists of we can use the Wayback Machine. On October 13 2017 it shows version 1.5.0.007 with 4089 downloads. On September 28 2020 it has version 1.9.0.01 with 5736 downloads. July 18 2024 has version 2.3.0.03 with 8674 downloads, and August 28 2024 has version 2.3.0.04 (the current version) with 8992 downloads.

So version 2.3.0.04 was available for download sometime between July 18 2024 and August 28 2024, and has been downloaded 334 times since then. That means only a small fraction of the total count represents 'new' users. We can only speculate on the number of actual users and how many are still active, but it's unlikely to be anywhere near 9336. I'm guessing it's less than 1,000.

Having said that, there could be a large number of Amiga fans who never bothered to download ScummVM because they aren't sufficiently interested in it. I myself only downloaded it once to try out Day of the Tentacle, and haven't used it since. That doesn't mean we aren't interested in other games though, especially new ones.

Yes, there are not that many users. The OS4 game with best sales I got still below 200 copies sold. Another reason why I always support both ways, OS4 and 68k.

Also, if it was a native build, it might entice PPC users to splurge on a 3D game that was built natively for their beloved HW, since - as I am being told - there actually aren't any PPC-native 3D games at all!

Quote:

BTW there is a second way how you can try out OS4 - there is a software called QEmu, which you can run on a PC or Mac, which emulates a PowerPC and runs OS4. Of course it is not the same, but then no need to buy an expensive AmigaOne.

All I really need, is something like WinUAE, where I can Alt-Tab, after creating an executable and running it, right there on my PC. Emulator is the best way for that.

Quote:

OS4 on PC/Mac hardware has a few disadvantages:

- Unless you give it an own partition, it will be slow if your Amiga HD is bigger than 2 GB (it uses file image then)
- It does not have 3D Hardware support (Appearently in works)
- It of course does not feel the same, if you run it like an emulator

- transferring files is a bit annoying, you basically go through ftp or Samba

3D HW Support - do you mean Warp3D here ? Or some different API ? Didn't PPC ecosystem have 3DFX / Glide API too ?

Quote:

On my 12th generation i7 system I get similar speed like with an AmigaOne x1000 with QEmu (and I can have OS4 on a Laptop this way ^^).

Quote:

Speed comparision with Heretic 2 OS4 in 640x480 software renderer (of course if I run it with 3D HW Renderer on the x1000 I get more than twice the fps rate than on QEmu with software renderer). Wazp3D is not really fast enough on QEmu, BTW (maybe on a i9 ?)

I guess, when I get back to U.S. in 2 weeks, I'll focus on updating my Benchmark build (say, for a week or two) and try to get it tested on these forums. This way, while it's being tested, I'll get some real-world numbers from the PiStorm and will have enough time to adjust some details/features before June...

The day I got my AmigaOne 800 MHz and finished the transfer of data from my poor old A4000 (which was partially broken, you never knew if you managed to switch it on again once you switched it off, sometimes it took me 1h to get it on) I switched off the A4000 one final time, never to look back again. Asides from "not broken" the more modern system and that OS4 just is much nicer had to do with it.

I'm assuming it can run 68k code [directly or indirectly] in some sort of JIT (hopefully) ?

What's the MIPS from SysInfo on that thing?

If your libraries cam support different resolutions

Quote:

If your libraries cam support different color depths you will have the possibility to support more hardware (320x200x8 ... 1024x768x24) it all depends on the level of abstraction you GFX engine does provide

But, as it currently looks that we don't really need to care for anything below Vampire, I'm afraid it'd be a wasted effort for systems like PiStorm/A600GS which has plenty bandwidth anyway...

Quote:

(f.e. world coordinates in floating point).

I even ported it to a 4 MHz Atari Lynx (released a short 3D game for one compo there).

But, Atari is stuck in the past. Amiga is the only retro platform that is marching forward (PPC, PiStorm, Vampire, A600GS, ...)

Quote:

I'm not so much the assembler fan because it will bind you to one platform, but you can still decide to do uncritical elements in a high level language like 'C' for faster development.

I have Z80 (for Spectrum Next), 6502, 68000, 68040, RISC DSP/GPU targets implemented. This is why I say that supporting PPC is just a matter of writing a compiler backend for PPC. Well, `just` ...

Quote:

If you are alone on your project, keep it small, you can still publish expansions later.

Last edited by Heimdall on 27-Feb-2025 at 09:42 AM.