@Kronos

Quote:

While that sounds fantastic it is also yet another 1 man project aiming (too) big.

It's work, sure, but sans the 32bit abiv0 layer most of it already exists. The end results is a bi-product of other existing things.

@ppcamiga1
Quote:

there is no native version of lv 3.5 for arm so what you wrote is total bs

It's not often I find myself questioning the right of free speech but when I do, it's usually having just read something this moronic.

Are you really as utterly stupid as you sound, or does it take effort?

All of the systems in the benchmark were running 68K lightwave because, shock horror, the Amiga version of lightwave is for 68K.

From it we see the relative performance of your beloved PPC for this task versus Pi. And guess what? You lose. When you factor in the cost/performance, you lose so hard it's unwatchable, even for me.

Given that the maximum performance of the 68K code is also ultimately a function of the maximum performance of the generated native code being executed (but also the efficiency of that translation), two possibilities emerge (the truth being somewhere between them):

1. The inexpensive Pi is comparable to the vastly more expensive PPC.
2. The inexpensive Pi is not as fast as the vastly more expensive PPC, but had a vastly better JIT.

Neither of these are good adverts for the vastly more expensive PPC.

_________________
Doing stupid things for fun...

@ppcamiga1

Quote:

ppcamiga1 wrote: ... It is still fun to use to use it.

Can you please let us killjoys know, what is so "fun" about using a computer for entertainment and productivity which happens to use a PPC cpu?

I used to use lots of entertainment and productivity software for PowerPC Macs (G3/G4/G5) running MacOS 9.x and then MacOS X, from 1999 until about 2010. Any "fun" I had on the Mac during the PowerPC time was the same "fun" I had on intel Macs, and now on ARM Macs. Same "fun", irrespective of CPU.

Then on some of my old PowerPC Macs I used applications and games on MorphOS. It was OK, but not especially "fun".

On my Power Mac G5 Quad 2.5GHz for a little while I ran Linux PPC. Zero fun.

So I ask you again: What special "fun" is it to use a computer that runs on PPC?

_________________
All the way, with 68k

@Karlos

switch because it is cheap.
it is shit compared to win/lnx/osx but switch because it is cheap.
want people switch to commodity hardware give people os as good as win/lnx/osx.

@agami

it is fun because it is not pc.

here little better result that my
probably other version
64 fps on pistorm with rpi3 on a500

https://www.youtube.com/watch?v=gJwdColUxUs

@ppcamiga1

Quote:

ppcamiga1 wrote: @agami it is fun because it is not pc.

Idiotic! Intel outside has become your religion!

_________________
"Art challenges technology. Technology inspires the art."
John Lasseter, Co-Founder of Pixar Animation Studios

From the same author who brought us this little gem:
Quote:

switch because it is cheap. it is shit compared to win/lnx/osx but switch because it is cheap. want people switch to commodity hardware give people os as good as win/lnx/osx.

We present the world its sequel:
Quote:

it is fun because it is not pc.

_________________
Indigo 3D Lounge, my second home.
The Illusion of Choice | Am*ga

@agami

You have a great point there.

Fun is the key and it should stay like this when working with computers.

Having a state of the art hardware is one way to attract new users, but having fun is the reason of sticking with it.

Amiga of the past won that, and some nostalgic users want to have that again.

Programming Assembly on a x86 architecture is NOT fun, but is well fun on 68x computers.

I thought that programming in PPC/CELL is also fun (never got the chance trying), but hearing from many developers PPC is a pain in the butt, and that was one of the reasons that PS3 didn't go well when was out.

Some new protection methods need to be applied to the new 68x hardware in order to be safe when connected to the network though ... I'm not that tech guy to go deeper on the requirements.

But, low level programming should be easy and open for new developments.
I'm tired of "open can" programming..

Last edited by liquidbit on 04-Jan-2024 at 11:04 AM.

@BigD

Quote:

BigD wrote: @ppcamiga1 Quote: ppcamiga1 wrote: @agami it is fun because it is not pc. Idiotic! Intel outside has become your religion!

He can't even present a consistent argument there, he often says PC gaming is fun.

_________________
Doing stupid things for fun...

@liquidbit

Quote:

I thought that programming in PPC/CELL is also fun (never got the chance trying), but hearing from many developers PPC is a pain in the butt, and that was one of the reasons that PS3 didn't go well when was out.

Having written quite a bit of assembly (as well as higher level) on a number of platforms, I can say that there is fun to be had on any architecture if you enjoy low level programming.

However, PPC, at least as far as the examples used in Amigas go, is not as much fun as it could be. All the hardware I had was usually bottlenecked by bus/memory so you might conceive of an awesome way to implement an algorithm in assembly and discover it was literally no faster when tested than the most naive version you could write in C with an optimiser enabled. You don't even need to use assembler to use any of the vectorisation, the compiler provides intrinsics for those, and can emit functionally equivalent scalar only code for targets that won't have it.

Contra example. I worked on a game engine (never finished) that was meant to be an axonometrically projected block based world for a syndicate type game (blocks were reusable bits of prefab geometry rather than Minecraft type solid blocks). I'd devised a pretty neat 8 bit fixed point format for the coordinates of geometry within a block that made data nice and compact, and wrote code that would convert an array of such into a array of screen transformed floating point data ready to he used in Warp3D indexed array drawing (complete with out of bounds culling though that was an edge case as the list of blocks were already culled to the view) with basic scissoring of what was left unculled so it was very fast. I pretty much got that routine down to copy speed on the hardware I had at the time. It was rather satisfying.

_________________
Doing stupid things for fun...

@Karlos

So in short:

Assembler is "fun" and even somewhat useful for ancient primitive CPUs that could be understood by a human with the 68000 being the top option.

For everything better/newer it is a painful exercise in futility,

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

It depends on your personality traits to an extent, but essentially, yes. Even 68060 blurs the line a bit, C code can perform well enough on it that unless you are very proficient, the gains you may make on any general code can be marginal. Obviously, there are exceptions everywhere. It's possible to write extremely well tuned x64 assembler even, but the effort/reward ratio is so skewed it's almost never worth it.

This is why I decided that MC64K should focus only on how easy it was to program, since it's whole raisin d'etre is to run code written in a heavily 68000 assembler influenced way for no other purpose than to have fun in an assembler like environment. Optimisation rules are really obvious: generally, least number of instructions wins (because it's an interpreter) and there are faster register to register only operations for selected logic and arithmetic.

Last edited by Karlos on 04-Jan-2024 at 12:46 PM.

_________________
Doing stupid things for fun...

@Karlos

68060 != 68000

Everything with even the tiniest cache means that you can't count cycles for given opcodes to determine which algorithm is the best.

Also once you go beyond single digit MHz and 1MB of RAM you will have a hard time coding anything worthy of the HW within reasonable time compared to much fewer lines of C code.

No with CPUs in the 10-100MHz range you might get to a point where hand optimizing some tight loop will make sense, but that really best done by coding it in higher language and looking at the generated code compared to straight up writing it in assembler.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

You overlook the fact that some people actively enjoy the challenge of assembler. My 68K skills tend to top out at 040 because I never had an 060 and I never really became proficient in PPC - only ever rewriting individual functions and such for the reasons previously stated: it almost always never made a meaningful difference. My G4 was bottlenecked by memory and so was my 603e.

It was still fun to try, even when so often fruitless.

Last edited by Karlos on 04-Jan-2024 at 12:52 PM.

_________________
Doing stupid things for fun...

@Karlos

I say it is a "a painful exercise in futility".

You say it can be "a joyful exercise in futility".

Mkay.....

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

Masochism, you mean?

_________________
Doing stupid things for fun...

@Karlos

The theme for everything "Amiga" in the past 20+ years I'd say.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

Nice to be agreeing on something for a change, eh?

_________________
Doing stupid things for fun...

@Karlos

Boring....

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada