@NutsAboutAmiga

Quote:

- ZeldaROTH_m68k extra binary for non AMMX Amigas

@IanP

Quote:

Full pipelining, multiple execution units and operation fusing means you don't necessarily need a GHz clock for a Gigaflop. Gunnar clearly states the new FPGA will be faster so isn't limited to ~100MHz, so for example 4 operations per clock would only need a 250MHz clock to reach a Gigaflop. If Gunnar says it's possible I'm inclined to believe him on this. I'm looking for to see what they have planned for Vampire 3 or Vampire 4.

So..you're "inclined to believe him" because it would only take "4 operations per clock" and "a 250 mhz clock" fpga.

Ever work with fpgas?
I have a few devices that use them.

Those two requirements are actually pretty lofty.

How about just delivering a compatible, functional fpu?

@iggy

https://www.altera.com/products/fpga/features/dsp/arria10-dsp-block.html

@IanP

REALLY nice.
Arria10 components are pretty high end though.
Cool to see Intel making such big improvements.

I gotta check those out more thoroughly.

Thanks for posting notice of it, something like that (IF you could convince people to pay for it, and I personally am rather stupid when it come to how much I'll shelll out, would be VERY cool).

Now...does anyone want to build an improved X86/X64 board?

@iggy

Quote:

REALLY nice. Arria10 components are pretty high end though. Cool to see Intel making such big improvements.

Pricing seems to vary greatly depending on the version, but does anyone have a rough breakdown of the difference between the 3, 5 and 10 series FPGAs?

And can they be used fairly interchangeably or does each require specific board logic?

It would be nice to have modular processor and simply buy the model that you can afford.

@iggy

Quote:

Now...does anyone want to build an improved X86/X64 board?

Yikes... what are you up to?

@ne_one

Quote:

Yikes...

Ah...what the heck?
Why not?

I'm sitting here looking at my A2000, and reminiscing of the days when this kicked ass.
It could run 68K software with a nice range of color in hi-res, with stereo sound, and you could plug in a card and run PC software.

Now, some of us are looking at running derivatives of AmigaOS directly on PC hardware (wasn't that Gateway's plan originally), some of us are still holding on to the intermediate step of the PPC, and others want to push the 68K further. Then there is ARM (which we haven't explored nearly as much as I'd like).

Those Arria10 Socs have nice ARM processors built into them, btw.

Last edited by iggy on 08-May-2017 at 05:53 AM.
Last edited by iggy on 08-May-2017 at 05:46 AM.
Last edited by iggy on 08-May-2017 at 02:18 AM.

@iggy

tu run 68k software, today, there is only one solution and it is not hardware


In these days there some new 68k productions that cut any possibilty to use on real Amiga even 68060 + RTG.
Vampire will be a good hardware solution, but slowly could divide amiga community again because developoer should be cover alls amiga models and amiga like: OS3 (68k classic and 68k AMMX), OS4, MOS, AROS,
In the past we saw that support so many platforms is equal to no optymizations and now some of these announced 68k games are are not fully playable on real 68k classic.

_________________

@Seiya

The NG options are not fully compatible to 3.X already and of course different to the other NG OSs too so split is there already. Vampire has a big advantage, it will standardize 68k as the main hardware platform.

@OlafS25

Quote:

it will standardize 68k as the main hardware platform.

Nonesense like AMMX will split the platform even further. OS compliant 68k applications run faster on NG OSs anyway.

@pavlor

most "classic user" are not interested in the so called "NG" otherwise they would use it already

@Seiya

OS3 is already split by things like RTG, AHI and processor generation (68000, 020, 030 etc), clock speed (7MHz 68000 to 80MHz 060), FPU/no FPU, chipset (OCS, ECS, AGA) chip ram (512KB, 1MB, 2MB), slow mem (less of an issue with patched software), fast ram (none to 256MB), Drives (Hard disk, CD Rom) and probably a few other things I haven't thought of. Vampire (AMMX) is just one more but will cure lots of others by bringing up for example a standard A500 to a "high end" Amiga with AGA compatibility, RTG etc. It may piss off some 040/060 accelerator owners that can't run AMMX software but it allows others to run a much broader range of software. A platform that stands still is a dead platform and if there's one thing we can be sure of it's not easy to kill the Amiga.

@OlafS25

Quote:

most "classic user" are not interested in the so called "NG" otherwise they would use it already

Well, Vampire today quite resembles NG (RTG, AHI) with limited CPU power. If new applications use artificial Vampire only features (AMMX), then even 68k Amiga platform suffers another split. People buy Vampire, because it gives them the very NG features (big resolutions, higher CPU speed, more RAM) without sacrificing classic roots. I even see people dreaming on forums to use Vampire for NG needs (eg. webbrowsing or movie playback) which is amusing considering real raw power of Vampire (or rather lack of it). It is clear Amiga users want NG... in classic disguise.

@Seiya

Quote:

tu run 68k software, today, there is only one solution and it is not hardware

Perchance, what would that be?
As all software requires hardware to run.

Currently everyone is pointing to X64.
But that still requires instruction interpretation and translation.

Ultimately, if we follow your logic, the real answer would be to not run 68K software at all, run native software.

As you're probably aware, I'm already an NG advocate, but I still use legacy hardware to run legacy applications in some cases (primarily software that directly hits hardware).

And legacy hardware is aging, so for me a replacement is welcomed.

I can support more than one solution to this equation.
For some apps, especially updated/ported apps, I go with my NG hardware, others that run better on a genuine Amiga I use legacy hardware (or an FPGA system) for.

Oh, and sometimes (God forbid), I use Windows.
Because sometimes you just have to compromise and make a pact with Satan himself to get something done.

BTW - On hardware vs. software, for NG, we need to make some decisions regarding the former.
I told myself that the X5000 would be my last PPC investment, but then I joined the group working on the T2080 laptop, BUT PPCs are DYING.

We need to take a good hard look at X64 (or ARM).
Hmm...hardware.

Last edited by iggy on 08-May-2017 at 01:05 PM.

@pavlor

they want to use it as far as possible and with more ressources people will want to use it for more advanced uses. They want to continue the road that was no longer available when Commodore went bankrupt, so to say "what would happened if...". I see nothing wrong with it. What is realistic is another question, let them have the fun. NG has not the same touch or emotions, it is (except processor) using the same hardware basically as modern systems but combining high prices, mostly oldfashioned hardware lagging years behind and bad or no driver support. Hardly a winner combination too

@OlafS25

Raw power of NG is much greater than that of Vampire. Not only CPU speed, but also 2D (compositing) and 3D GFX (even with slooow Warp3D). You may play Quake 3 class game in HD resolution even on A1-500, but this would be next to impossible for Vampire. The difference is so great even emulated 68k applications run faster on NG than on Vampire.

@iggy

Seiya means WinUAE.

@pavlor

if that would be important everyone would use AROS on X86

@OlafS25

Quote:

if that would be important everyone would use AROS on X86

Too limited compatibility.

@pavlor

Quote:

Seiya means WinUAE.

Yeah, I rather figured that, but I've had Amiga Forever for years, and while WinUAE (especially when tweaked) is pretty damned good, it still isn't perfect.

@OlafS25

Quote:

Hardly a winner combination too

Its all a matter of focus, legacy or legacy+, and its all just a hobbyist pursuit anyway.

@pavlor

Quote:

Quote: "if that would be important everyone would use AROS on X86" Too limited compatibility.

Funny, that's what I'm advocating in the X64 version of MorphOS, accepting compatibility limitations in order to enhance the OS' capabilities.

Last edited by iggy on 08-May-2017 at 01:15 PM.