Was anyone else disappointed with the A1200?

[Bruce Abbott]

Originally Posted by Thorham

I thought they only did CPUs

Not even that. Arm don't make hardware.

[Bruce Abbott]

Originally Posted by chb

If you honestly think that a 68000@7 MHz is only 1.4 times faster than a 6510@1 MHz, I've got a bridge to sell.

I just quoted the official MIPS. What that equates to in practice depends on the application.

For example, here's the same game on the C64, PC, ST and Amiga. Doesn't look like the 68000 is making it run any faster or smoother.

[ Show youtube player ]

[Thomas Richter]

Originally Posted by Bruce Abbott

Not quite sure what you are trying to say here - that the A1000 was effectively no better than the C64?

Memory-wise , the original A1000 with its 256K was pretty much on par with the C64, yes. There was fourfold as much RAM, but you also needed fourfold as much RAM for all the basic stuff. Wider instructions, larger screen resolution, larger pointers. My best guess is that this is how the memory size was actually picked in first place. It turned out that the 256K were actually even a bit on the low end, which is why it became 512K shortly after.

Originally Posted by Bruce Abbott

Sure, we all would have liked that (so long as it was affordable). But the market needed a lot more than just better hardware - it needed PC compatibility.

But that's a little bit a difference what Amiga users wanted and what the general computer market wanted. At the time of the demise of the Amiga, PC compatibility was mandatory, but at the time the Amiga was introduced, the world was still a bit different. What the market wanted was a system you bought for serious applications, and could use for gaming. There was no real niche for the Amiga, though there could have been one, as there was also a niche for the Macintosh, even though it was not PC compatible. This is what the argument is about: CBM management followed its legacy and attempted to sell it as home computer, though *that* niche was dead. They failed to develop a new niche - but that requires a bit more than just excellent (at that point in time) hardware.

[hammer]

Originally Posted by Bruce Abbott

Realistically, we had to switch to PCs just to be compatible. I had an A3000 with 60MHZ 060, 32MB of RAM and RTG, but I couldn't run Netscape or Microsoft Internet Explorer so websites I wanted to use were not available.

A3000 with 060 @ 60 Mhz, enough RAM, and RTG should be able to run Shapeshifter's MacOS 8.1, Netscape Communicator 4.0.8, and Internet Explorer 4.0.

Originally Posted by Bruce Abbott

I also couldn't read word documents or pdf files,

Run MS Office 4.2 for MacOS 68K.

[hammer]

Originally Posted by AestheticDebris

But probably not for much longer:

https://en.m.wikipedia.org/wiki/X86-64#X86S

It's another Intel-driven UEFI Class 3 push. There's a "second source" AMD.

Unlike Intel, NVIDIA's ADA and AMD's NAVI 3X still have VBIOS and VGA legacy.

For Intel IGP and GPUs, Intel has removed legacy VBIOS support since 2020.

[hammer]

Originally Posted by AestheticDebris

The PC absolutely died. The brand lived on. First the 16-bit DOS was replaced by a 16-bit Windows kernel. Then that was replaced by a "386 Enhanced" hypervisor. Then that was replaced by a fully 32-bit Windows 9x. Which was subsequently replaced by a totally redesigned, 32-bit, VMS clone called Windows NT. Which was then replaced by a 64-bit NT derivative. And modern versions of that bear little resemblance to their original....

If Mac is Triggers Broom, the PC is The Ship of Theseus.

IA-32 Windows XP to 10 can run Win16 Windows applications. X64 Windows XP to 10 has jettisoned Win16 support.

[ Show youtube player ]
How to easily run 16-bit apps on 64-bit modern Windows?

Open-source WINE VDM enables Win16 apps to run on Windows X64.

[hammer]

Originally Posted by Thorham

All platforms have non-consistent game quality. Games always range from great to bad and everything in between.

The PC had VGA's Mode 13h chunky pixels and Mode X year 1991 published baseline.

[Thorham]

Originally Posted by Bruce Abbott

Not even that. Arm don't make hardware.

Yeah, they license designs or something, but I didn't realize they had a home computer back in the day.

Originally Posted by Bruce Abbott

I just quoted the official MIPS.

You're comparing 6510 MIPS to 68000 MIPS. You know that's absurd, right?

Originally Posted by Thomas Richter

Memory-wise , the original A1000 with its 256K was pretty much on par with the C64, yes.

You can't be serious

Originally Posted by hammer

The PC had VGA's Mode 13h chunky pixels and Mode X year 1991 published baseline.

I'm not talking about graphics fidelity but game play.

[chb]

Originally Posted by Bruce Abbott

I just quoted the official MIPS. What that equates to in practice depends on the application.

Where are those values published that itmakes them "official"?
MIPS ist a remarkably ill-defined quantity. It's pretty clear that those for the 6502 are peak MIPS - the fastest instructions on that CPU take 2 cycles, so they give the theoretical maximum of 0.5 MIPS@1 MHz. Using the same definition, a 68000@7MHz reaches 1.8 MIPS, not 0.7, not taking into account that it can do more in a single instruction than the 6502. Those 0.7 MIPS given for the 68k on the other hand probably use a definition based on a realistic mix of instructions, which at 10 cycles/instruction sounds about right.

The 68k can be massively faster for a number of applications; take e.g. a 16x16->32 multiplication. About 500 cycles on a 6502 using a typical mulu routine, 60 cycles on average on the 68k. So an Amiga 1000 performs that task about 60x faster that the C64. That's an extreme example, but e.g. soft floating point is also extremely slow on the 6502.

Comparing random games running on vastly different hardware architectures isn't very helpful. By that definition a 68k can be considered equivalent to a 486...

[Karlos]

Originally Posted by Bruce Abbott

Not even that. Arm don't make hardware.

The ARM evaluation system was built for the BBC Micro as a second processor using it's side expansion slot. Designed and built by basically the same team, operating as the same company, that built the BBC Micro itself. This was before the Archimedes, this was the inception of ARM itself. Like most other alternative platforms, they saw the writing on the wall but they knew they had something special (high performance, low power) so they decided to carry on with the processor and licensing model and became ARM holdings.

ARM originally stood for Acorn RISC Machines.

[Bruce Abbott]

Originally Posted by Karlos

ARM originally stood for Acorn RISC Machines.

Yes, I know the history of ARM. But a CPU isn't a computer line. The Archimedes was discontinued in 1995. After that Acorn produced the Risc PC, which initially cost £1249 with 2MB RAM, 210MB hard drive and 14" monitor. Acorn stopped making Risc PCs in 1998 and that was it for them, another 'home' computer bites the dust.

Interesting to note that the Archimedes had a shorter lifespan than the Amiga (8 years vs 11 years), during which time it received virtually no improvements apart from a slightly faster CPU. Yet strangely, fans don't complain bitterly about it.

[Bruce Abbott]

Originally Posted by chb

It's pretty clear that those for the 6502 are peak MIPS - the fastest instructions on that CPU take 2 cycles, so they give the theoretical maximum of 0.5 MIPS@1 MHz. Using the same definition, a 68000@7MHz reaches 1.8 MIPS

OK lets use those numbers. 1.8/0.5 = 3.6. Now compare 68020 vs 68000. We know from sysinfo and AIBB that a 14MHz 68020 is 4 times faster than a 7MHz 68000, so the A1200 has a similar CPU speed increase over the A1000 as it does over the C64.

Comparing random games running on vastly different hardware architectures isn't very helpful.

Proof of the pudding and all that. Gamers only care about what the machine did in practice. In practice the most popular game of the day played better on the C64 than on the Amiga. If there was a reason to be disappointed with your Amiga, that was it.

[Bruce Abbott]

Originally Posted by hammer

A3000 with 060 @ 60 Mhz, enough RAM, and RTG should be able to run Shapeshifter's MacOS 8.1, Netscape Communicator 4.0.8, and Internet Explorer 4.0.

Good point, except that doing this legally was very difficult, and I didn't have access to the required software and firmware.

At one time I had a real Mac Color Classic, and the experience put me off Macs for life. If I had to emulate a Mac then I might as well have just used a PC - a lot less hassle and guaranteed compatibility with anything that was thrown at me.

[aeberbach]

Originally Posted by Bruce Abbott

In practice the most popular game of the day played better on the C64 than on the Amiga. If there was a reason to be disappointed with your Amiga, that was it.

You make me laugh, not enough to engage with this nonsense-as-a-thread, but I do laugh at you sometimes. Back it up Bruce, name the game!

[minator]

Originally Posted by Bruce Abbott

Yes, I know the history of ARM. But a CPU isn't a computer line. The Archimedes was discontinued in 1995. After that Acorn produced the Risc PC, which initially cost £1249 with 2MB RAM, 210MB hard drive and 14" monitor. Acorn stopped making Risc PCs in 1998 and that was it for them, another 'home' computer bites the dust.

Interesting to note that the Archimedes had a shorter lifespan than the Amiga (8 years vs 11 years), during which time it received virtually no improvements apart from a slightly faster CPU. Yet strangely, fans don't complain bitterly about it.

It was more than slightly faster, it was several times faster. Which was important given that unlike the Amiga, the CPU did everything.

The history after Acorn stopped making Risc PCs is a bit like the Amiga, one company continued the line for another 10 years but there was a second company who also had rights to the OS and sold a different version...

However unlike the Amiga, it all eventually ended up together and the OS was eventually open sourced in 2018.

It's been ported to all sorts of boards, including various RaspberryPi variants. However, it wont run natively on Arm CPUs that only support 64 bit.

[AestheticDebris]

Originally Posted by Bruce Abbott

OK lets use those numbers. 1.8/0.5 = 3.6. Now compare 68020 vs 68000. We know from sysinfo and AIBB that a 14MHz 68020 is 4 times faster than a 7MHz 68000, so the A1200 has a similar CPU speed increase over the A1000 as it does over the C64.

Reducing things to numbers like this is clearly nonsensical. But if you're going to do that you need to factor in that there was 3 years between the C64 and A1000, but 7 years between the A1000 and A1200 - so it'd need to be over double the speed increase to be "equivalent"

Originally Posted by aeberbach

You make me laugh, not enough to engage with this nonsense-as-a-thread, but I do laugh at you sometimes. Back it up Bruce, name the game!

Loads of them. Let's be honest, a lot of early ST and Amiga games weren't much more than the same game as their 8-bit counterparts with slightly tarted up graphics. And a lot of them got rather more attention on looking pretty over being well implemented.

I'd take the excellent Spectrum version of Chase HQ over the lacklustre Amiga port any day (heck I think it's a better game than even the arcade).

[dreadnought]

Originally Posted by Bruce Abbott

Proof of the pudding and all that. Gamers only care about what the machine did in practice. In practice the most popular game of the day played better on the C64 than on the Amiga. If there was a reason to be disappointed with your Amiga, that was it.

Flappy Bird played better on even a cheap smartphone than beefed up PC or console. Ergo: cheap smartphones are better gaming devices than beefed up PCs or consoles.


This is how crazy a thread can get when you flog the same dead horse for 350 pages. So, yeah, a few outlier games (mostly quick'n dirty ports) played better on the ZX or C64, but that still doesn't make water dry and sky green.

[TCD]

Originally Posted by AestheticDebris

Reducing things to numbers like this is clearly nonsensical.

You must be new to this thread...