When was the first 68040 accelerator for Amiga on sale?

[CCCP alert]

So I was looking up some stuff and found the first 486 PC was an Apricot server (VX FT?) in sept 1989. Looks like the first 040 chip was about half a year later.

As the 486 is the first real decent 32bit x86 chip I was wondering what was the first Amiga accelerator card with 040 and 32 bit RAM.

Is the A4000/040 actually the earliest you could have an 040 Amiga or was there a CPU slot card for A2000/3000 that was out earlier.

[TCD]

http://amiga.resource.cx/manual/Progressive040_2000.pdf
From 1991 for the A2000 (not sure if this is the first though).

[cgugl]

Not for sale, but the first cpu board with the 68040 seems to be this one
https://bigbookofamigahardware.com/b...ct.aspx?id=221

[CCCP alert]

Hmm that 040 proto for CPU slot shows Commodore engineers knew the importance of the 040 CPU in the coming desktop wars.

Quite a grey area finding the first proper 486 PC too, it's all about the Compaq server of 89 and then not much else is talked about.

I would imagine even an expensive 040 card for an expensive A3000 and a 24bit Z2/Z3 video card would be cheaper than a branded 25mhz 486 desktop PC around the time of the launch of the 040 processor in early 1990 though. A 286 cost twice as much or more than an A1000 back in my discussions with my computer science teacher who wanted one for the lab. By late 1992 cobbled together 25mhz 486s via big adverts in Computer Shopper were half the price of the A4000/040 in 92 but I think you need a 33mhz 486 to get on a par with a 25mhz 040 for raw integer code execution in assembler.

[dreadnought]

Ok, but what about Doom?

[CCCP alert]

Originally Posted by dreadnought

Ok, but what about Doom?

What about it?

[grelbfarlk]

Does Doom?

[Bruce Abbott]

Originally Posted by dreadnought

Ok, but what about Doom?

Doom was released in December 1993 and came on four 1.44MB floppy disks. A 68040 accelerator was not included in the box.

[ShK]

It was Wolfenstein 3D, not Doom

[ Show youtube player ]

[CCCP alert]

This is not a thread about games, nobody was buying a 486 for gaming at the start of 1990/end of 1989. There is already a thread for Doom, Wolf and DOS vs Amiga gaming respectively so please keep those comments there.

The 486 is a pivotal CPU for Intel in the scientific, industrial (CAD/CAM and engineering) and commercial/governmental office markets to rival UNIX Workstations etc which obviously would have been first to market with an 040 machine I am guessing. Was curious to know how early an Amiga+AMIX or UNIX could have been possible. Probably take me seconds to find out the first 040 Mac but the 040 Falcon was canned, Acorn didn't do CISC so the only other real option for a serious 040 based 'budget' Workstation option was Amiga. The 4000 uses an EC040 with no MMU right? This means no UNIX and I presume AMIX either.

Was just trying to get a feel for how far Motorola and all their clients who would use an 040 were behind the first 486 server Apricot VX FT in Sept 89 and whenever the first 486 workstation was released seeing as Motorola were a bit late to this battle and therefore companies like Sun etc had to make do with 030 vs 486 for at least 6 months or more until the 040 was done.

Lots of knowledge about Amiga accelerators among members here and the 2000 and 3000 could run a decent UNIX OS with the right spec/upgrades but 'when' is the key. Of course you would need an Amiga SCSI controller with the option of RAID arrays attached for serious workstation environments. Like I said this is not about little Johnny getting a 486 PC at Xmas to play Doom vs Gloom deluxe on a £400 2mb A1200 or even A4000/030 for £1000. Take that talk to other threads meant for that 'problem'

[patrik]

Full 040 in the A4000. Internet says some FPU-less variants were sold, but that must be a later thing a not very common.

[Bruce Abbott]

Originally Posted by CCCP alert

The 486 is a pivotal CPU for Intel in the scientific, industrial (CAD/CAM and engineering) and commercial/governmental office markets to rival UNIX Workstations etc which obviously would have been first to market with an 040 machine I am guessing. Was curious to know how early an Amiga+AMIX or UNIX could have been possible.

UNIX was possible since 1988 when Commodore released the A2500 with 68020 and 68851 MMU. They even produced a model called the A2500UX which came with Amiga UNIX.

The 4000 uses an EC040 with no MMU right? This means no UNIX and I presume AMIX either.

No. Originally it was a full 040. The later 'cost reduced' A4000 had an LC040 (which has the MMU but no FPU). AFAIK the EC040 was never used in any Amiga model or accelerator card.

Was just trying to get a feel for how far Motorola and all their clients who would use an 040 were behind the first 486 server Apricot VX FT in Sept 89 and whenever the first 486 workstation was released seeing as Motorola were a bit late to this battle and therefore companies like Sun etc had to make do with 030 vs 486 for at least 6 months or more until the 040 was done.

A 50 MHz 030 has about the same processing power as a 25 MHz 486, so they didn't have to 'make do'. One example of a machine using it was the HP/Apollo 400s, introduced in 1990.

Lots of knowledge about Amiga accelerators among members here and the 2000 and 3000 could run a decent UNIX OS with the right spec/upgrades but 'when' is the key.

Key to what?

[TCD]

Originally Posted by Bruce Abbott

Key to what?

Maximum whatifism

[Seiya]

the first was Fusion Forty per A2000 adverted in 1990:


and again in 1991:


released in 1991:


in 1992 68040 is released for A500:
advert Jun 1992:


https://amiga.resource.cx/search.pl?...&base=dec&pid=

[d4rk3lf]

Originally Posted by TCD

http://amiga.resource.cx/manual/Progressive040_2000.pdf
From 1991 for the A2000 (not sure if this is the first though).

Lol, imagine a guy that had that in 1991, and in 1993 he meet few friends, a proud Aga owners.
Friend 1: I have A1200 with 020, it's very fast.
Friend 2: I have A4000 with 030, it's speed is unmatched.
Our guy: I have 040, 2 years already.

[Karlos]

I remember that 040 board for the A500. It was insane in 1992.

[CCCP alert]

Thanks for the info about the LC/EC040 in A4000.

I had a 9mb A2000 with 48mb SCSI wordsync Supra HDD card in 1990, used to grab 250+ 16 greyscale low res video caps via a Rombo digitizer and use then as animbrushes. Later added a VLAB Y/C card and a 24 bit video upgrade in the Denise socket which supported standard full PAL/NTSC broadcast quality at standard 15.5khz video output so worked nice enough.

As for 50mhz 030, this was top end max speed, the 33mhz 486 wasn't released that much later than the 25mhz 040, the 33mhz 486 not long after the 25mhz. There are better benchmarks than MIPS like Specmark for integer and FPU performance but the DX33 is 27 MIPs (25mhz=20 MIPs) and a 50mhz 030 is not that high (about 11-12 MIPS). UNIX for 680x0 is more mature and well implemented than i386/i486 flavours so MIPS isn't the whole picture. Perhaps the 68882 is better than the 1989 486 FPU.

Specint/SpecFPU and other benchmarks are better (especially those run on a UNX Kernal) but you get the idea.

The key = to understanding the time delay between various 25mhz 486 vs 25mhz 040 potential workstation purchases/builds becoming available. Nothing to do with "what if" I want to know WHEN these machines came out in this battle for this the most sophisticated market sector and where Amiga fits in. It's the Formula 1 engine/chassis Williams vs Mclaren battle of the computing world.

Jay Miner had a great passion for graphics workstations, HAM is just the RGB version of HSV 24 bit video compression used by flight simulators he had seen, the area fill/line draw etc extensions to blitter also from his exposure to top end graphic workstations in 83/84 too IIRC.

[idrougge]

Originally Posted by CCCP alert

The 486 is a pivotal CPU for Intel in the scientific, industrial (CAD/CAM and engineering) and commercial/governmental office markets to rival UNIX Workstations etc which obviously would have been first to market with an 040 machine I am guessing. Was curious to know how early an Amiga+AMIX or UNIX could have been possible. Probably take me seconds to find out the first 040 Mac but the 040 Falcon was canned, Acorn didn't do CISC so the only other real option for a serious 040 based 'budget' Workstation option was Amiga. The 4000 uses an EC040 with no MMU right? This means no UNIX and I presume AMIX either.

Unfortunately, AMIX doesn’t run on 68040. It was already abandoned when the A4000 came out.

[Bruce Abbott]

Originally Posted by CCCP alert

There are better benchmarks than MIPS ... but the DX33 is 27 MIPs (25mhz=20 MIPs) and a 50mhz 030 is not that high (about 11-12 MIPS).

According to this page a 50MHz 030 does 20 Dhrystones, 25MHz 80486 does 24, and 25MHz 68040 does 40.

AIBB benchmarks indicate that a 25MHz 040 (Blizzard PPC, Falcon 040/25, Apollo 1240) does 52% more Dhrystones than a 50MHz 030 (Blizzard 1230-IV). Floating point showed a larger difference, with the Blizzard PPC's 25MHz 040 rendering the beach ball 2.17 times faster than the Blizzard 1230-IV's 50Mhz 030+8882.

So a 25MHz 040 was definitely worth getting over a 50MHz 030, but a 25MHz 486 might not have been.

The key = to understanding the time delay between various 25mhz 486 vs 25mhz 040 potential workstation purchases/builds becoming available. Nothing to do with "what if" I want to know WHEN these machines came out in this battle

Knowing when machines were released doesn't tell you much about why.

The i486 was announced by Intel on April 10, 1989, with samples expected in the third quarter and production quantities in the fourth quarter. A manufacturer who wanted to 'get the drop on' others could buy those samples and put them into machines before the end of the year. They would be paying a premium and the chips might have bugs, but might be worth it to be first on the market.

According to this article in the January 1990 issue of Computerworld magazine, Motorola was already shipping samples of the 68040 with production quantities expected by the 3rd quarter. However the prices quoted were $795 for the 040 vs $950 for the i486. If this is accurate then the 040 was better value even though only in sample status. MIPS were quoted at 20 vs 15 for the i486, and MFLOPS at 3.5 vs 1.0, which would make the 040 even more attractive. However in terms of time to market the i486 beat it by 6 months.

The next question is what would the OEMs do? The architectures of the 040 and 486 are very different. If an OEM was using 68k in current models it would pay to wait for the 040 rather than switch to Intel if they knew it was coming. Manufacturers who worked closely with Motorola would probably get advance notice, and so would plan for its release.

This is enough to explain the 6 months delay you found, and sets a limit on how early the Amiga could have gotten an 040. It also fits in with Motorola supplying Commodore with a 'special blessed' MC68040 for the A3000 launch in April 1990. According to Brian Bagnall in "Commodore the Final Years", the chip Motorola sent them was the latest revision, nicknamed 'Queen Bee'. Before that they must have had sample chips to test the prototype 040 boards with (by this time the board was up to revision 4).

As to what relevance this delay had to do with the Amiga, I'd say not much. Few if any customers were so desperate for increased performance that they would pay 'anything' to get an 040 before large scale production started. Only fans consumed by PC envy would get upset about the PC world getting 486's a bit sooner (in low quantities and great expense). By July 1990 GVP was offering its Impact A3001 card for the A2000 with a 50MHz 030. With twice the performance of an A3000, most customers would be pretty happy. Their first 040 board wouldn't be released until 1992.

Jay Miner had a great passion for graphics workstations, HAM is just the RGB version of HSV 24 bit video compression used by flight simulators he had seen, the area fill/line draw etc extensions to blitter also from his exposure to top end graphic workstations in 83/84 too IIRC.

yes. The Amiga was originally going to be a games console, but Miner had other ideas. The partners who wanted to keep it low-end were pushed out of the company. After Commodore took it over and produced the A1000, Miner wanted to take it even higher with 'Ranger', which would have an ultra-high resolution of up to 1024x1024 for workstation applications.

Luckily for us and Commodore, management recognized that a low-end machine was needed. Miner didn't want that and tried to put a spanner in the works by suggesting that fat Angus wouldn't work, but his own designs weren't ready when crunch time came and Commodore chose the A500 and A2000 instead. Miner's Los Gatos team was then let go, but other engineers also dreamed of having an Amiga workstation - probably as result of using workstations themselves for product development. That's the main reason we got the A2500 with A2620 and A2630 cards, and the A3000. It's also why the A2410 graphics card and A2024 monitor were developed.

While the engineers worked on building the machines they wanted, they neglected updates to the OCS chipset that would have addressed the perception that the Amiga was slipping behind. This did not help Commodore's prospects at all. Jay Miner deserves credit for creating the Amiga, but in the end he (and others) did it a disservice by trying to make it into something most of us didn't want.

[Thomas Richter]

Originally Posted by Bruce Abbott

Jay Miner deserves credit for creating the Amiga, but in the end he (and others) did it a disservice by trying to make it into something most of us didn't want.

Maybe not "most of you", but "most of some other group of customers" that would have had the money to sustain the development of the platform, other than "most of you".