Apollo Maggie 3D Chip - Preview Demo

[Promilus]

@grond - what do you mean 68k instruction does more work? It doesn't. Moving registers, all math calculations - those instruction works the same. It makes difference in memory accesses due to segmented memory (different peripheral, external memory etc.) vs linear memory so it overall might take more x86 instructions to get the job done but that's not a great difference. Yes, AC68080 has impressive memory controller (why shouldn't it - it has support for memory several generations ahead of what 68k and Pentium had with EDO DRAM). And while it helps to reach results closer to the peak performance that peak performance isn't way ahead of what Pentium (MMX) could do.

[Ffin]

Originally Posted by meynaf

For retro gaming we don't need more than what we already have.

YOU don't need more than you already have, but you don't speak for me. I welcome more!

[Ffin]

Originally Posted by Bruce Abbott

I am more concerned for those of us who can't justify the hassle and expense or don't want to further compromise the 'retro' status of our machines. Fancy 3D is all very well if that's all you want (but then why use an Amiga when a modern PC does it so much better without even trying?). But many of us spent a lot of money and some effort to make a 'high-end' Amiga with an 060 or Vampire, only to now find that new games produced for the Amiga need more. It shouldn't be necessary to keep paying and paying for ever more 'upgrades' - that's not the Amiga way!

Well this clearly just isn't for you, is it? Stick with what you have, but don't expect the rest of us to hang back with you.

[meynaf]

Originally Posted by Promilus

what do you mean 68k instruction does more work? It doesn't. Moving registers, all math calculations - those instruction works the same.

68k instruction set is better than x86 instruction set.
It has more registers so it spends less efforts in swapping them around and can do more calculations directly. It has better addressing modes and much less constraints overall. As a result, even though individual x86 instructions are of same size and often even smaller, 68k programs are denser than x86 (for big programs x86 version can be 1.5 times bigger than 68k - compare them yourself if you don't believe me).

Originally Posted by Ffin

YOU don't need more than you already have, but you don't speak for me. I welcome more!

In retro gaming there is the word 'retro', perhaps something you've missed.

[grond]

Originally Posted by Promilus

@grond - what do you mean 68k instruction does more work? It doesn't. Moving registers, all math calculations - those instruction works the same.

I mean exactly what it says. E.g. the 68k ISA has more powerful address modes than x86. For this reason the x86 needs to sometimes use two instructions where the 68k needs just one. The one 68k instruction does the same work of two x86 instructions and thus does more work than either x86 instruction.

[Ffin]

Originally Posted by meynaf

In retro gaming there is the word 'retro', perhaps something you've missed.

Your retro isn't my retro, friend. We all had different experiences. You are the one who is missing something.

[Promilus]

For this reason the x86 needs to sometimes use two instructions where the 68k needs just one. The one 68k instruction does the same work of two x86 instructions and thus does more work than either x86 instruction.

While I understand "sometimes" I don't understand that "more powerful ISA" How dense compiled output code is - that's something irrelevant nowadays as both storage media and RAM itself is quite abundant. And when it comes to actual calculation performance it seems 68k vs x86 ain't making great difference. And in aspect of actual math performance even RISC-core based K5 outperformed Pentium (and by extension 060) in integer math. So while ASM instructions 68k is more user friendly quality of compiled high level language code and it's performance is what matters to me.

[grond]

Originally Posted by Promilus

While I understand "sometimes" I don't understand that "more powerful ISA" How dense compiled output code is - that's something irrelevant nowadays as both storage media and RAM itself is quite abundant.

It's not irrelevant. A instruction cache of size X will hold more instructions if the code is denser increasing the likelihood of a cache hit. The same is true for bandwidth. If you can load X MB/s of instructions from RAM, you will load more instructions to be processed if the code is denser.

[Promilus]

@grond - Pentium already had twice the cache of 060 and pentium mmx had 4x and also chipset cache (L2) which could be few hundreds KB. So... your point? And how much bigger is typical x86 code over 68k anyway? 68k only real advantage came with asm developers. And developers have little or no control over what comes to cache and when. By the time Pentium came out most PC apps were being compiled from high level language and with 95 using winapi, system libraries and abstraction layers. And there are system libraries and API for Amiga as well. ASM development is encouraged mostly to non-expanded amiga or timing critical routines.

[grond]

Originally Posted by Promilus

@grond - Pentium already had twice the cache of 060 and pentium mmx had 4x and also chipset cache (L2) which could be few hundreds KB. So... your point?

You may be talking about the 060 and the Pentium MMX. I am not. I was talking about the wrong comparison of the 080 to a Pentium 75.

[Promilus]

Well I didn't go into "AC68080 is only P75 level" theme. I did however put Pentium-class and AC68080 together and I don't think I made mistake by that (arguments already presented). And x86 vs 68k asm number of instructions generated in typical usage makes small difference in size (when using rigid 4 bytes wide A32 of ARM obviously does, that's why Thumb-2 was conceived for microcontrollers which have little RAM and not that much FLASH ROM, also with RISC it really takes more instructions to make the same job as most/all RISC are unable to grab data out of the memory directly. It needs to be loaded up to registers first. Doesn't mean it takes a lot longer to execute - only that it takes more - very simple - instructions to do so).

[meynaf]

Originally Posted by Ffin

Your retro isn't my retro, friend. We all had different experiences. You are the one who is missing something.

Then what am i missing exactly ?
Or perhaps you're just arguing on semantics ?

[Ffin]

Originally Posted by meynaf

Then what am i missing exactly ?
Or perhaps you're just arguing on semantics ?

You wrote "we don't need", when you should have written "I don't need"! You speak for yourself and noone else. You certainly never speak for me! If there ever were to be a single voice for the group, I sincerely doubt that it would be yours!

[meynaf]

Originally Posted by Ffin

You wrote "we don't need", when you should have written "I don't need"! You speak for yourself and noone else. You certainly never speak for me! If there ever were to be a single voice for the group, I sincerely doubt that it would be yours!

You didn't answer the question of what i am supposed to be missing.

[Ffin]

Originally Posted by meynaf

You didn't answer the question of what i am supposed to be missing.

You are missing the point that your opinion is not a universal fact! You are also clearly missing the ability to admit being wrong! And WOW, DID THAT TAKE SOME EXPLAINING! Big ego or tiny IQ? Both?

[grond]

Originally Posted by Promilus

Well I didn't go into "AC68080 is only P75 level" theme. I did however put Pentium-class and AC68080 together and I don't think I made mistake by that (arguments already presented). And x86 vs 68k asm number of instructions generated in typical usage makes small difference in size (when using rigid 4 bytes wide A32 of ARM obviously does, that's why Thumb-2 was conceived for microcontrollers which have little RAM and not that much FLASH ROM, also with RISC it really takes more instructions to make the same job as most/all RISC are unable to grab data out of the memory directly. It needs to be loaded up to registers first. Doesn't mean it takes a lot longer to execute - only that it takes more - very simple - instructions to do so).

Well, then you won an argument against a point I didn't make. My point was that the 080 is better than a Pentium 75 (even if sometimes only slightly) in every isolated aspect that matters (memory interface, IPC, code density, cache size, clock rate) and thus also as a whole. I did not claim it to be faster than e.g. a 233 MHz Pentium MMx.

[meynaf]

Originally Posted by Ffin

You are missing the point that your opinion is not a universal fact! You are also clearly missing the ability to admit being wrong! And WOW, DID THAT TAKE SOME EXPLAINING! Big ego or tiny IQ? Both?

Speaking of yourself here, i see.
For your information, "we" does not necessarily include "you".
Furthermore, to admit being wrong i must first see arguments - and you've shown none so far.
So please stop that nonsense right now, otherwise your trolling will be signaled to moderation.
During all that time it would have been a lot better to just say what's included in your vision of retro gaming that we currently don't have...

[tygre]

Originally Posted by meynaf

Speaking of yourself here, i see.
For your information, "we" does not necessarily include "you".

Oh, Meynaf, who's "just arguing on semantics" now?

Just drop the matter and use "I" rather than the (royal? ) "we"...

Peace!

[meynaf]

Originally Posted by tygre

Oh, Meynaf, who's "just arguing on semantics" now?

Sorry, i had to return the gift. Not the smartest move i admit, but little choice.

Originally Posted by tygre

Just drop the matter and use "I" rather than the (royal? ) "we"...

Peace!

Of course my english might be betraying me, but it seems to me that "I" implies only one person. I supposed there were several, actually the majority because many people here keep their machines "as is" and are happy with them. Or are you telling me nobody is satisfied with their current config ?

I supposed that "retro" means things from the past, i.e. retro gaming was about what we had back then and still enjoy today.
Sorry, but that chip is obviously not for anything retro, it is for "modernizing" the platform.

[coldacid]

Not to totally OT the thread, but yeah it'd be nice if English had inclusive and exclusive "we" (i.e. shared first and second person plural, exclusive first person plural).