Was anyone else disappointed with the A1200?

[oscar_ates]

1. Lew Eggebrecht wasn't in the hot seat. Bill Sydnes has the hot seat. Who demanded for IDE? Hint: Commodore Germany.

2. You're forgetting that the year 1992 was a large financial blow due to double debacles from A600's release and A500 cancellation. This is Bill Sydnes factor!

3. Fact: Bill Sydnes is the person who was fired by Ali, NOT Lew Eggebrecht.

4. AA3000+ and A1000+ AA projects were frozen for "more than 6 months" to focus on the ECS adventures i.e. A1000Jr and A300(A600). This is the Bill Sydnes factor! AA3000+ revision 1 reached surface-mounted chip design before being frozen. "A1000Jr" refers to Bill Sydnes.

"More than 6 months" wasn't used to complete AA machines. AA3000+ is the prototype AA.

"A1000Jr" did NOT include Gayle or AA-Gayle and Budgie since they were NOT completed in 1991. AA-Gayle is dependent on A300's Gayle R&D.

What was the reasoning of Mehdi Ali & Bill S. that A600 would sell like hot cakes?

[TCD]

Originally Posted by oscar_ates

What was the reasoning of Mehdi Ali & Bill S. that A600 would sell like hot cakes?

Reasoning? One of the better jokes I heard this week

[hammer]

Originally Posted by oscar_ates

What was the reasoning of Mehdi Ali & Bill S. that A600 would sell like hot cakes?

Renaming A300 into A600 was to cover Bill Sydnes' mistakes. Due to failed promises made by Bill Sydnes, Mehdi Ali fired Bill Sydnes.

As for David Pleasance and Psygnosis upgraded CD32 for minimal cost increase, Mehdi Ali wasn't aware of competitive pressures on CD32's math power deficit, despite spending for the CD32 FMV module's $50 CL-450 SoC with a vision grandeur being a digital VCD player that departs from the core gaming business.

Argonaut Games (UK) pushed for mass-produced "cheap RISC" (SuperFX, 16-bit custom RISC CPU) and Nintendo supported the request.

Sony's Ken Kutaragi designed SNES's 16-bit SHVC-SOUND chip's S-DSP and was influenced by Nintendo's SNES visual DSPs and Mode 7.

SNES's SHVC-SOUND solution has Sony SPC-700 microprocessor (based on MOS/CSG 65xx 8bit CPU, ), S-DSP, stereo 16-bit 32000 Hz digital audio DAC (NEC uPD6376) and 64 kb of internal RAM.

[hammer]

Originally Posted by Bruce Abbott

I knew you would do that. Your nitpicking is BS.

On a technical basis, your argument is bullshit.

Originally Posted by Bruce Abbott

Generally when people say 'Windows XP' they mean 'Windows XP', not 'Windows XP 64-Bit Edition for Itanium'.

Your argument is superficial labeling bullshit.

For AMD64 (X86-64), there's Windows XP x64 edition. AMD attached thier existing X86 sales channel and economics of scale with K8 AMD64. AMD K8 was the "gaming king" CPU before Core 2. The gaming PC argument is a very important factor.

Originally Posted by Bruce Abbott

I didn't mention Itanium because it was a non-entity in the personal computer market,

Your argument is not made on a technical basis. Windows XP x64 Edition and Windows XP 64bit Edition for Itanium systems targeted similar markets.

IA-64 is Intel's 64-bit desktop computing path and replacement for IA-32 a.k.a i386.

IA-64 includes a very slow IA-32 mode. Intel's Pentium Pro has gimped 16-bit X86 instructions in the past.



Without the second source AMD's X86-64, Intel's IA-64, and IBM PowerPC 970 would be fighting for 64-bit desktop supremacy.

Windows XP 64-Bit Edition for Itanium Systems is for personal desktop PCs.

Like IBM MCA's attempt to assert control over the PC clone market, it's an Intel tactic to remove x86 cloner AMD.

Originally Posted by Bruce Abbott

and because it was specifically designed to be Intel's successor to x86 - not a 'foreign' architecture like PA-RISC or ARM.

WRONG.

1. IA-64 is a foreign instruction set (Explicitly Parallel Instruction Computing, EPIC, a VLIW variant) and it's nothing like IA-32.

2. IA-64 supports both little and big endian modes. Itanium's big endian mode targets HP's PA-RISC legacy. PA-RISC's big endian mode targets HP's 68K HPUX legacy.

Itanium's instruction set is based on HP's research on VLIW

https://en.wikipedia.org/wiki/Itanium

In 1989, HP started to research an architecture that would exceed the expected limits of the reduced instruction set computer (RISC) architectures caused by the great increase in complexity needed for executing multiple instructions per cycle due to the need for dynamic dependency checking and precise exception handling.[c] HP hired Bob Rau of Cydrome and Josh Fisher of Multiflow, the pioneers of very long instruction word (VLIW) computing. One VLIW instruction word can contain several independent instructions, which can be executed in parallel without having to evaluate them for independence. A compiler must attempt to find valid combinations of instructions that can be executed at the same time, effectively performing the instruction scheduling that conventional superscalar processors must do in hardware at runtime.

HP researchers modified the classic VLIW into a new type of architecture, later named Explicitly Parallel Instruction Computing (EPIC), which differs by: having template bits which show which instructions are independent inside and between the bundles of three instructions, which enables the explicitly parallel execution of multiple bundles and increasing the processors' issue width without the need to recompile; by predication of instructions to reduce the need for branches; and by full interlocking to eliminate the delay slots. In EPIC the assignment of execution units to instructions and the timing of their issuing can be decided by hardware, unlike in the classic VLIW. HP intended to use these features in PA-WideWord, the planned successor to their PA-RISC ISA
....

Intel's engineers were impressed when Jerry Huck and Rajiv Gupta presented the PA-WideWord architecture they had designed to replace PA-RISC. "When we saw WideWord, we saw a lot of things we had only been looking at doing, already in their full glory", said Intel's John Crawford, who in 1994 became the chief architect of Merced, and who had earlier argued against extending the x86 with P7.


HP's Gupta recalled: "I looked Albert Yu [Intel's general manager for microprocessors] in the eyes and showed him we could run circles around PowerPC, that we could kill PowerPC, that we could kill the x86."

Itanium is effectively PA-WideWord, the planned successor to their PA-RISC ISA.

Itanium preserves HP EPIC's three instructions per bundle. Like other semiconductor companies, Intel drank the VLIW cool-aid.

Commodore's PA-RISC hitched their ride with HP's PA-WideWord which is ultimately, Intel's Itanium!

The other major VLIW-based CPU designs are Transmeta Crusoe/Efficeon (faking IA-32/X86-64, no native MMU hence the MMU is emulated) and NVIDIA's Project Denver/Carmel (faking AArch64).

NVIDIA had VLIW-based GeForce FX.
ATI/AMD had VLIW-based Terascale and Northern Island series before Graphics Core Next (GCN) architecture.

3. IA-64 has a slow in-order processing P5-like block for IA-32.

Itanium with native big-endian AROS and userland 68K emulator would have run Amiga's 68K OS-friendly apps.

You're wrong.

[TuRRIcaNEd]

Originally Posted by MagicSN

Yes, definitely. To the point where I would never have considered to buy it (68020 ? Seriously ? No Fast RAM ?

The thing is, even at the time, I can see your point if we're talking purely "on-paper" specs. But with the benefit of 20/20 hindsight, things were somewhat more complicated than that

Though I today also have an A1200. With a PiStorm Cm4 included. Guess that also counts as a Bigbox Amiga ^^

A work colleague (when I was 16 years old in 1994 and working for a small computer sales chain in the UK) built his own homebrew "A1200T" tower system - several months before the third-party conversion kits to do something similar came out. The truth was, that "low-end" mainboard ("Channel Z" IIRC) was rather more capable than it first appeared.

Commodore's errors ... were too much "low end system", always focusing on the lowend - A1200, A600, A500+ - when what would have been needed was more on the Highend.

With all due respect, I beg to differ. What ended up happening to the A500 and other subsequent "home" platform variants in the US market was, as you say :

...caused them to get irrelevant when other gaming platforms overtook them.

I presume you're talking about the Sega Mega Drive/Genesis and the Super Nintendo there - we'll come back to that.

Focusing on Lowend caused them to get irrelevant...

Now if we're going to be entirely truthful regarding the history, from the late '70s until the mid-late '80s CBM's raison d'etre (as expressed by Jack Tramiel) was along the lines of "Computers for the masses, not the classes". In other words, making systems that were affordable for people on a modest domestic budget was CBM's core mission.

Even in terms of the US market (which became the source of CBM's undoing in the late '80s to early '90s), their biggest successes up until 1985 were the VIC-20 and C64 - the latter being able to take advantage when the bottom fell out of the home console market catastrophically circa 1982. Strictly speaking, from 1982 to 1986 (when the NES was released en masse in the US), the C64 held the almost-unique position of being (arguably) the best affordable home gaming platform as well as being an effective home computing platform if the user so desired.

It's generally accepted that the 1985 OG Amiga (later A1000) was a technical marvel years ahead of its time in several ways - but the business shenanigans around the bidding tug-of-war between Tramiel's Atari Corp. and CBM, along with CBM's "power vacuum" in terms of management and business strategy meant that the initial product was streets ahead of the competition in most technical regards, but the price point was too expensive for a "home market" machine (i.e. the sector where CBM traditionally performed the strongest) and the hardware was not really expandable enough for the business market.

I'll always be of the opinion that Tom Rattigan's '86-'87 strategy of separating the platform into "home" (A500) and "business/professional" (B2000) product lines was the most logical and effective way to go. It's true that both Nintendo and Sega had committed to a big push into the North American market by then, however in 1987 all they had were their 8-bit offerings (namely the NES and Master System). As far as their 16-bit console offerings went, the Sega Genesis/Mega Drive was not released into that market until August 1989 and the Super Nintendo did not properly appear until September 1991.

Why do I mention that? Because if Irv Gould had not reflexively shitcanned Rattigan in 1987 and Rattigan had been able to supervise a proper US promotional campaign for the A500 and B2000, it would be reasonable to wager that the A500 could have been positioned as a 16-bit successor to the C64 approximately two-to-four years before the Japanese console titans had their 16-bit products ready for the US.

Rattigan's strategy appears to have covered a lot of potential scenarios, but it began with getting a relatively affordable "home" version of the Amiga to market as early as was practical - well in advance of the 16-bit Japanese consoles as far as the US/Canada market was concerned. The USP of the "classic" Amiga platform was that it was a powerful creative tool out of the box; between DPaint and Noisetracker it was possible for an Amiga user to create smooth and impressive audio-visual work without needing to know how to code at all or having to buy any expanding hardware - which was damned-near unique at the time!

I think it's worth pointing out that when it came to the introduction of the two new Amiga machines that came out of Rattigan's plan in 1987, it was far more a matter of "home" vs. "business" or "professional" use rather than "high-end" vs. "low-end" - because fundamentally, the hardware and chipset specs "out-of-the-box" were more-or-less the same, but what the B2000 offered was future expandability. In 1987, the A500 potentially offered a gaming experience well in advance of what Sega and Nintendo could offer for at least another two years - and on top of that, because it was a computer rather than a console it could provide creative capabilities for the user that no console could ever match. In other words, what the A500 and B2000 models represented was in combination a contiguous platform for both "the masses" *and* "the classes" if marketed properly.

Obviously, this is wild-ass speculation, but I'd argue it's reasonable to consider it possible that had Irv Gould not been so paranoid and left Tom Rattigan in charge the strategy may well have paid off, and by the time Sega and Nintendo released their 16-bit offerings in America, the A500 could have had two years to build a user base which could then have provided a baseline for the platform as a whole to move upmarket and at least challenge Apple, if not the PC compatible sector.

And to save the gaming part they would have needed to do what Apollo did today... (maybe just up to 640x480 and 16 Bit color, at that time of course no 3D Options yet)

Even the all-conquering PSX was limited to a native vertical resolution of 200/240 pixels because it had to be capable of being hooked up to a common-or-garden TV set. Just like the 1985 Amiga OCS, the only way it could achieve 640*400/480 was via an interlaced (i.e. "flickering") mode.

In any case - when it comes down to it; as far as I'm concerned the question as to whether the A1200 was "disappoint[ing]" depends on the context. By which I mean - had CBM Intl (as in the US parent firm) been competently managed post-1987 and it had still taken until 1992 to release a machine with a comparable spec to the A1200, yes, that would have been very disappointing.

However in real terms, what we're talking about is a company dealing with a toady CEO whose original business roadmap was predicated on killing the Amiga platform and re-tooling as a commodity Wintel PC supplier, only to belatedly discover that there was no way that proposed model could compete with the likes of Dell and Gateway in terms of scale. Having effectively gutted the hardware R&D budget over at least three years (and rendered the promising AAA architecture obsolete in the process), CBM's engineers were tasked not only with developing a (mostly) backwards-compatible - but significantly improved - chipset on a breakneck schedule, but also designing the "home" version of the platform in such a way as to take advantage of emerging standards such as IDE and PCMCIA while still keeping the cost of the base unit in 1992 equivalent to that of the A500 in 1987.

In that sense, the A1200 ought to be considered nothing less than a minor miracle.

[hammer]

Originally Posted by TuRRIcaNEd

With all due respect, I beg to differ. What ended up happening to the A500 and other subsequent "home" platform variants in the US market was, as you say :

For 1987, the A500 introduction aimed for the $699 USD price range which is $863.20 USD in 1992.

For 1992, the A1200 aimed for $599 USD which is about $485.05 USD price in 1987. A1200's "guts" are about $399 retail CD32 level.

There's a healthy profit margin with A1200.

The person who priced A1200's budget has aimed for Atari ST's price range i.e. "PCJr mentality". Jeff Porter's "8 bit planes with 16 million colors" combined with Bill Sydnes's "A600Jr" = A1200. Bill Sydnes cuts the 32-bit compute power cost to the bone.

A500 has a hardware premium over the Atari ST.

3DO's $699 price target in 1993 is closer to A500's premium price.
LG released 3DO at $399 to price match Sega Saturn's $399.

A minor hardware cost increase for A1200 would have $20 DSP3210-50, 1 MB Fast RAM (80 ns access FP DRAM about $25), and 68EC020-25 ($19 without discounts). A better configured "Falcon" for A1200.

Reference
https://www.minneapolisfed.org/about...ion-calculator.

For 1987 to 1990, A500 has a sweet spot with enough premium hardware for a low enough price.

As for A1200, when there's CPU power, AGA is a reasonably good low-cost entry-level SVGA-like chipset and faster than IBM's original VGA. $20 DSP3210 @ 50 Mhz would give AGA some FP32 math power beyond 486DX-66 or 68040-40. This is why other 32-bit 3D game consoles have fast MUL instructions as their cornerstone math power.

Remember, early PC 3D accelerator Rendition Verite v1000 has a 25 Mhz MIPS-like CPU with a fast 1 cycle 32x32 MUL instruction, fast 1 clock approximate reciprocal instruction, hence 2 clock approximate integer divide, usual set of RISC instructions, "bilinear load" instruction and tiny cache for 4 pixels. No hardware for Z buffers.

[TommyTomorrow]

This thread should be called "were you disappointed with the A1200 in the 90's?" and most people compare the A1200 to what is available and what we know about it now.

[redblade]

Originally Posted by hammer

PiStorm32 RPi 4B has a very low-end price.


It wouldn't make any difference since the Amiga platform wasn't in the corporate office knowledge-based systems e.g. combined GUI word processing with custom workflow business rules and accounting software fusion.

Amiga platform has a weak "business" relational database software.

Oracle 7.1 was released on Windows NT in 1994.
Oracle v6 was released on Novell Netware 386 in 1988. Before the arrival of Windows NT Server, Novell claimed 90% of the market for PC-based servers.

Amiga would have been okay for SQL based databases? I know Oracle released one for NeXT and other 68k based unix systems.

There seemed to be quite a battle in the /X BBS door scene for the fastest zippy search, surely some of those could have been used in SQL. (Yes I know SQL is a lot more than text search of table fields).

They could have used A1200's in Video stores in the early 90s as a DB system or any store.

[malko]

Here is an answer I got from ChatGPT (that has been turned on end of 2021) about performance comparison of Win95, OS2 and AmigaOS in the 90's :

It is difficult to definitively say whether one operating system was less performant than another, as performance can be influenced by a wide range of factors including hardware specifications, software optimizations, and user experience. However, it is generally agreed upon that Windows 95 was less stable and had more compatibility issues than OS/2 and Amiga OS. OS/2 was known for its stability and efficiency, while Amiga OS was praised for its advanced multimedia capabilities and multitasking abilities. Ultimately, the performance of an operating system can vary depending on individual needs and preferences.

[AestheticDebris]

Originally Posted by malko

Here is an answer I got from ChatGPT (that has been turned on end of 2021) about performance comparison of Win95, OS2 and AmigaOS in the 90's :

And that's about as much proof as anyone needs that ChatGPT will quite happily spout utter nonsense. Compatibility with what exactly? OS/2 could only run a subset of Windows 3 applications whereas Windows 95 ran pretty much all of them (and AmigaOS didn't run any).

And, as much as people berate Win95 it had both preemptive multitasking and virtual memory support, putting it slightly ahead of AmigaOS technically. PC hardware in general was still behind the Amiga though, which lead to stuttering when you loaded files from floppy disk etc (although having a hard drive made this a lot less of an issue).

Both of them were substantially nicer to use than OS/2 though which, in my experience at least, was utter frustration.

[malko]

Originally Posted by AestheticDebris

[...] Compatibility with what exactly? OS/2 could only run a subset of Windows 3 applications whereas Windows 95 ran pretty much all of them (and AmigaOS didn't run any). [...]

The question was about performance on each OS. Meaning :
- Win95 apps on Win95 OS;
- OS/2 apps on OS/2 OS;
- Amiga apps on AmigaOS.
So there is no point in saying that Amiga cannot run Win 3 applications or that OS/2 can only run a subset of them.

[TEG]

Originally Posted by redblade

Amiga would have been okay for SQL based databases? I know Oracle released one for NeXT and other 68k based unix systems.

Lack of memory management disqualify it for me. And as @AestheticDebris say, lack of virtual memory is another problem I think. Database often need a lot of resources.

[hammer]

Originally Posted by redblade

Amiga would have been okay for SQL based databases? I know Oracle released one for NeXT and other 68k based unix systems.

There seemed to be quite a battle in the /X BBS door scene for the fastest zippy search, surely some of those could have been used in SQL. (Yes I know SQL is a lot more than text search of table fields).

They could have used A1200's in Video stores in the early 90s as a DB system or any store.

There was Oracle Server Release 7 for A/UX (Apple's Unix with System 7) e.g. https://68kmla.org/bb/index.php?thre...-for-dl.17363/

MMU wasn't a guaranteed deployment for 32-bit 68K CPUs when compared to IA-32 (i386) CPUs.

In 1992, Am386-40 was priced like 68030-25 and 386DX-25.

[TEG]

Originally Posted by hammer

In 1992, Am386-40 was priced like 68030-25 and 386DX-25.

I see their was a 40Mhz of the 68030. Do you have an idea of the price at the time?

P.S.
For ignorant like me, Am386 was AMD clone of the Intel 386.

[Bruce Abbott]

Originally Posted by AestheticDebris

And that's about as much proof as anyone needs that ChatGPT will quite happily spout utter nonsense.

And even if it isn't nonsense you can't trust it, so it's useless.

And, as much as people berate Win95 it had both preemptive multitasking and virtual memory support, putting it slightly ahead of AmigaOS technically.

And it needed it. On a 4MB machine you were hitting the swap file just getting to the desktop.

PC hardware in general was still behind the Amiga though, which lead to stuttering when you loaded files from floppy disk etc (although having a hard drive made this a lot less of an issue).

Floppies were terrible - the whole system would freeze up if you put a bad disk in the drive.

User response was often sluggish too. It got better in XP, then worse in later Windows versions as all pretense of efficiency went out the window.

To speed up operation Windows 95 etc. used several techniques. One was cashing icons so they could be displayed faster. Windows 98 had a bug where when the icon cache file reached a certain size it would get corrupted and all your icons would change to one (random) icon image. To fix it you deleted the icon cache, then you got to see how slow it was when not cached. Made the Amiga look fast!

I recently acquired a PC motherboard with 386SX-16 CPU and 16MB RAM. Windows 95 runs fine on it, but takes about 5 minutes to boot. Every time you click on something it has to think about it for a bit. Strangely my relatively modern laptop with 4GB does the same thing in Windows 10, but worse.

In comparison my A1200 is a breeze. Even the stock A500 is much faster!

Both of them were substantially nicer to use than OS/2 though which, in my experience at least, was utter frustration.

I tried out the OS/2 Warp pre-release and it was quite nice. Problem was it needed 8MB RAM when most PCs only had 4MB. Of course Microsoft made sure that Windows 95 would work (badly) in 4MB, and that was the end of OS/2 Warp!

[malko]

Originally Posted by TEG

[...] For ignorant like me, Am386 was AMD clone of the Intel 386.

And "Cx" was used for Cirix CPU (those CPU were used on some Compaq and Packard Bell models of computers). Equivalent to Intel's CPU for office (for a lower price) but a bit less performant for gaming.

[CCCP alert]

Can't remember where I heard it recently but the A300 went off the rails was because Commodore couldn't cost reduce the machine enough.

One major part of cost reducing a computer is doing massive production runs, due to Irving and Medhi's ridiculous bonus checks this would never happen. Jack Tramiel obviously didn't piss the company resources away and hence the C64 sales dominated the sales of rivals so much it's like comparing VCS to Vectrex/Coleco sales.

You want to survive the console threat you need a CEO who doesn't piss away CBM profits and make it impossible to replicate power/price of something like a console/C64 via massive up front bulk raw material purchasing/productions runs. 75% of the A1000 to A500 cost reduction is down to less RAM (no 192kb WORM daughterboard for Kickstart) and the massive drop of DRAMs in that 24 month period an the massive middleman profit margins of A1000. Gould's Commodore were total fail in minimising production costs.

[Bruce Abbott]

Originally Posted by hammer

MMU wasn't a guaranteed deployment for 32-bit 68K CPUs when compared to IA-32 (i386) CPUs.

Amiga OS didn't use an MMU because the 68000 didn't have one. This was fortunate because it meant Commodore couldn't screw up the OS by giving it virtual memory.

In 1992, Am386-40 was priced like 68030-25 and 386DX-25.

Remembering of course that the 386 had no internal cache. To get the best out of that 40MHz CPU you needed expensive fast static RAM and a cache controller. Most 386SX motherboards had no cache, so performance at higher clock speeds was not nearly as good as you might expect.

[Thorham]

Originally Posted by Bruce Abbott

Amiga OS didn't use an MMU because the 68000 didn't have one. This was fortunate because it meant Commodore couldn't screw up the OS by giving it virtual memory.

And because of that we're now stuck with legacy crap that prevents us from having proper memory protection

[Thomas Richter]

Originally Posted by Bruce Abbott

Amiga OS didn't use an MMU because the 68000 didn't have one. This was unfortunate because it meant Commodore screwed it up by not giving it virtual memory.

I fixed it up for you. Anyhow, Apple *did* give the Mac virtual memory, despite using a 68K. So it was clearly possible, but it requires more programming discipline CBM had.