@wawa

I wish the Vampire team all the best. I could very well become their customer.
It is very good for the future of Amiga. Nice to see that even Hyperion support them in latest OS 3.x release. To me it is this and AROS for Pi that is exiting for the future.

@megol

Quote:

megol wrote: @kolla Quote: kolla wrote: @megol Ok, right, then there's even less reason to talk about this as "improving C++ compilers". Please inform yourself before having an opinion as this is just ridiculous.

In fact, as far as I can remember GCC can directly compile C++, i think that's the point to have a full port with Apollo native instruction compiling capabilities

_________________
God created Paula so that Allister Brimble and Dave Whittaker could do music

Check my Amiga gameplays (ITA)!

@ErikBauer

Actually its multiple layers.

And "make" controls the compile process its responsible for figures what's what need to be recompiled, and it solves that needs to be compiled and in what order,

Pre processor generates program code, G++ generates assembler, AS generates machine code, and then lnker connects everything into one blob of executable file.

As pointed out before G++ has no interpreter.

C# has as runtime, its more like Java (actually C# is based on Microsoft java, that Microsoft had to rename after being sued by Sun Microsystems), C# has intelligent garbage collector. Actually it's not a C language.

On subjects of interpreter, I only say that computer has execute things in sequence, some that were hard to do, if there are like mathematical rules of operations where like * and / has higher priority then + or -, or where stuff inside "(" ")" has calculated before rest can be calculated and so on, the interpreter is able solve this things, so the idea of converting a interpreter into compiler is not a bad idea.

Last edited by NutsAboutAmiga on 26-Aug-2019 at 09:45 PM.
Last edited by NutsAboutAmiga on 26-Aug-2019 at 09:43 PM.
Last edited by NutsAboutAmiga on 26-Aug-2019 at 09:35 PM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

@NutsAboutAmiga
Thank you a lot for the kind clarification, it is always good to (re)learn something

_________________
God created Paula so that Allister Brimble and Dave Whittaker could do music

Check my Amiga gameplays (ITA)!

@ErikBauer

Just put my name & address down for a pre-order with Vesalia, to help them gauge how much interest there is for the Vampire 1200 V2+. I normally do support developers like yourself, who have projects that I'm interested in. Looking forward to seeing the finished product.

Ed

_________________
Amiga 1200 - ACA 1233 68030 128MB Ram
8GB CF With tons of Classics

AmigaOne X5000

Raspberry PI 400 - PiMiga 1.5
"That which doesn't kill you, only makes you stronger" - Someone important, but I forgot who

erased by me

Last edited by Trekiej on 27-Aug-2019 at 06:44 AM.
Last edited by Trekiej on 27-Aug-2019 at 06:21 AM.

_________________
John 3:16

@ErikBauer
Yes GCC have a C++ front-end which will produce the correct code using a proper 68k back-end (which is shared for all supported languages).

@kolla

On Coldfire old amiga binaries would work without recompile?

If not you have a problem because many programs never will be recompiled because developers and/or companies are not existing anymore or have left the market or there is a legal chaos and nobody knows who owns the software.

Regarding obscured 68080... even if the new commands are not documented and because of that not used, as long as the old software and compilers work with it you have a good platform with at least 125 MB RAM, very fast 68060 and RTG. If I would program amiga software I would anyway not directly target vampire but program it in a way that the software runs on as much hardware as possible. That happened back in the time too, software (expecially games) were developed for A500, not for the already existing better platforms. If you target RTG you can already do a lot. The specialities of Vampire are only needed for certain cases. Software could f.e. check ram and processor and make graphics better if it runs on a vampire. That was the case with ambermoon that detected ram and processor if it run on A1200 (unfortunately AGA was not used).

Regarding Commodore you are correct but I assume that was affected by the decision of motorola to drop the 68k line, if new processors after 68060 would have been announced amiga would have stayed 68k. But we will never know...

Last edited by OlafS25 on 27-Aug-2019 at 03:18 PM.

Pre-sales must be slower than expected (and summer is over)...

Quote:

The APOLLO 68080 is like an improved 68040 or improved 68060. This means behave 100% like 68040 MMU would make sense.

http://www.apollo-core.com/knowledge.php?b=4¬e=22867&z=R3yw9I

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@OlafS25

Quote:

On Coldfire old amiga binaries would work without recompile?

Amiga didn't even come into the picture here, it was a dead platform already and Motorola could not have cared less. Other 68k platforms, the ones that were actually alive and able to recompile, could and did.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@OlafS25

Quote:

On Coldfire old amiga binaries would work without recompile?

Not, not to my knowledge and this was the dealbreaker. Why we never saw any ColdFire 68K accelerators. It's said to be source compatible but not binary compatible. Or at least not fully. I've read it misses out on some bit instructions as well.

https://en.wikipedia.org/wiki/NXP_ColdFire

@kolla

Yes but why do you mention coldfire here then?

For most of the amiga software (at least most of the former commercial software) there are no sources and no chance anyone will ever recompile it. So even if a coldfire based hardware is faster it makes no sense

@Hypex

Gunnar said that too. As far as I can remember they looked at coldfire when natami was started and dropped the idea because of that.

I assume Kolla knew that too so I do not understand him

@OlafS25

If you read what I replied to, you see why I mentioned ColdFire.
It's not like Motorola made 68k exclusively for Amiga, 68k development went to ColdFire, but Amiga was not in a position follow, and instead went a much more sane route (at the time) of PowerPC.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@Hypex

It would have been fine if Coldfire simply didn't implement all 68k instructions, the real problem are instructions that doesn't cause an illegal instruction trap and have significant differences with the original 68k. The similarities makes a JIT a bit easier/cleaner but it still requires a type of emulation.

@megol

Well, noone is talking about "improving FORTRAN compilers", which would be just as correct. It was about one thing, improving gcc, primarily for C code as that is most relevant.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

Quote:

This means behave 100% like 68040 MMU would make sense.

Vampire do not have to have full 68k MMU.
Tool for catching null pointer access will be enough.

Last edited by ppcamiga1 on 07-Sep-2019 at 04:06 PM.

@ppcamiga1

Too late, it already has an MMU, it's just not compatible with any previous 68k MMU and also not available from any operating system running on the card.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

it's not "too late" if it's not documented no one will use it, if no one use it, then there is no consequences if its changed. Let vampire enginers mess around figure it out they most likely get around to fix it. I do not see point of going rampage on the forums.

Last edited by NutsAboutAmiga on 08-Sep-2019 at 04:59 PM.
Last edited by NutsAboutAmiga on 08-Sep-2019 at 04:58 PM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

@kolla

I spent at beginning 90s a lot of money on a new A4000 with graphic card but without MMU... if I had known that I spent lots of money on useless garbage. Nobody explained it to me