real amiga vs winuae

[Karlos]

The 68000 was incomplete in some areas, particularly with regard to virtualization and the separation of privileged and unprivileged instructions, so Motorola made changes to the design.

The 68010 is really what the 68000 should have been.

[Karlos]

Performance aside, emulated 68K has another advantage. You can support every user mode opcode that's ever been implemented on the 680x0/6888x. That allows you to maximise the throughput of all 68K code. If you emulated just the subset supported by a given CPU model, then your emulated CPU would have to jump through the same (emulated) exception traps to handle them. With a fully virtual CPU, unimplemented instruction exceptions need never happen for any legal opcode. You can just support them.

[Karlos]

My point is valid for 680x0 processors not just 68020 which I used as an example because happen to have it in front of me to play around with.

No, your point is not valid. Read and comprehend: Any code that uses any instruction not implemented in silicon will fail fatally on any 680x0 that does not have the requisite support code installed

If your 68000 code touches the SR outside of supervisor mode, it will crash on 68010, 68020, 68030, 68040 and 68060 despite being legal on 68000.

If your 68881/68882 code touches any transcendal function or uses certain rounding instructions, it will crash on 68040 and 68060, unless trapped and handled in software (provided by 040/060 libraries).

If your 68020/30/40 compatible code uses any 64-bit integer division or 32x32->64 bit integer multiplication it will crash on 68060, unless handled in software (provided by 060 library).

If all you have at your disposal is a 68020 then you obviously haven't felt the pain others have when their 68040.library or 68060.library has been missing or had problems. Count yourself lucky.

[Karlos]

Your ignorance was just proven-- you think a library is required and it's not.  You have yet to address many other points made and I won't waste my time linking to them as you are proven time and time again to blurt out blind statements like above.

What kind of misinformed drivel is this? Of course the_leander is perfectly correct to say you need the library if you are using a 68040/68060. After several dozen recoverable alerts during booting I just about managed to get into workbench, just, before a fatal error took it down again.

Don't believe me? See attached. Now, if you don't mind, I'll re-enable my 68040.library so that I can actually get the thing booted.

[Karlos]

Notice the fatal error in the middle shot? That's right, an unhandled line F exception, which if you knew anything at all about 680x0 you'd recognise as an unimplemented instruction exception.

Directly compatible, my erse. You want to use an 040 or 060? You need the library, unless rebooting your machine is your favourite hobby.

[Karlos]

@the_leander

LOL, bloody hell, where do you find these?

[Karlos]

Ah, how I love the sound of skull against swedish cluebat...

[Karlos]

Your experiment is a failure because you aren't testing for compatibility.

As far as being compatible, here another link for you proving I'm not alone in this world running code on 680x0 originally written for 68000:

http://www.digchip.com/datasheets/parts/datasheet/522/68040.php

OMG, what planet are you on? I tested compatibility using my Operating System. That is the single most important piece of software I have for the system and the one I use more than any other software.

You don't need to test much other software because without the OS it was designed to run on you probably aren't going to it working in the first place.

Not being able to boot the OS that will run on the 68020 even, simply because the 68040.library is temporarily disabled more than adequatly demonstrates that the 68K series does not do complete hardware backwards compatibility. The 68K series requires software support for full backwards compatibility. This is a simple fact and nothing you can say changes this.