The OS4 kernel (known as ExecSG) is a completely rewritten version of the old OS 3.1 kernel that was all in 68K assembler.

The new kernel (which was written by the OS4 core developers, Thomas and Hans-Joerg Frieden) is written mainly in C, but with a few critical parts written in PPC assembler. It is backwards-compatible with the old 68K design and extended considerably with virtual memory and other design features. It also contains support for (ie, runs on) every type of PPC CPU that the several hardware platforms contain.

_________________
cheers
tony

Hyperion Support Forum: http://forum.hyperion-entertainment.biz/index.php

Gee, thanks. You people are so helpful...

_________________
Common sense - So rare it's almost like a super power

Quote:

I don't think it is the actual AmigaOS. The reason why...I have read it is based on a third party kernel called ExecSG. Who knows what that is but it's probably Linux with an Amiga API wrapper. Code riding on top may appear to be AmigaOS but I want the real deal!

naturally AmigaOS 4.x IS AmigaOS, it has nothing to do with Linux...
It is well known that AmigaOS 4.x is the direct evolution of AmigaOS 3.x.

Last edited by andres on 03-Sep-2011 at 12:37 AM.

_________________
A1200/020+68882 - 6 MB RAM - AmigaOS 3.0

Home Recording Audio

@X1000_Booster

Video is sometimes better than plain words:

AmigaOS Twenty-Five


And of course...

Welcome!

@pavlor

Nice reminder !

AOS4 is the real thing. Native, actual AMIGA OS on PowerPC (CBM was planning to use HP PA RISC, but CBM died).



AOS4 can also run most AOS3.1 productivity applications "natively". I mean that AOS4 uses a petunia "tool" to turn the 68k binary to native PPC form "on the fly".

My SAM is the fastest true AmigaOS system I have had so far. I hope also old AGA (and AAA, Hombre and Caipirinha) features will one day be available via "thin emulation", by using Gallium3D & OpenGL etc....

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

@KimmoK

Quote:

and AAA, Hombre and Caipirinha

Multi-core support on the X1000 would be far more useful.

@pavlor

Quote:

Multi-core support on the X1000 would be far more useful.

I bet +2-3 years after first release of x1000 we will have initial multi-core suppot, which of course, will be buggy at first release, and will be fixed in the next 2 years. So in summ, i think if you will swith off from amiga stuff for the next 5 years, and come back after, then, pretty possible multicore support will be done. But its just "maybe" of course :)

Personally for me, i start to feel that resource tracking and protection of the necessary OS parts are very-very need it, to avoid all those crashes, intuition lockups and freezes. much-much more than any multicore support, which will not solve for us all our current problems.

Last edited by kas1e on 03-Sep-2011 at 08:11 AM.

_________________
Join us to improve dopus5!
zerohero's mirror of os4/os3 crosscompiler suites

@kas1e

Quote:

much-much more than any multicore support, which will not solve for us all our current problems.

Of course! However, it would need even more time than multicore-support.

@pavlor

:)) 10 years ?:)

_________________
Join us to improve dopus5!
zerohero's mirror of os4/os3 crosscompiler suites

excellent news!

_________________
AmigaNG, YouTube, LeaveReality Studio

@kas1e

Sorry to disappoint you but the issue of adding SMP support was discussed and refined on the OS 4 developer list since MAI sent Hyperion a dual Freescale MPC 7410 card which could have worked in the AmigaOne (proof of concept).

Since there were no actual AmigaOS machine available with multiple cores or CPU's, SMP support was ofcourse a waste of time to implement in itself.

The way in which it will be implemented however is already clearly defined and was subject to peer review by other developers.

Obstacles to an efficient implementation were removed (e.g. the use of Forbid) and replaced in many OS components over the years (e.g. DOS).The foundation for SMP support was put in place, a clear picture exists what needs to be done to accomplish it and how.

I'm willing to take a bet that it won't take 2 years ;)

(Disclaimer: unless something better comes along that is even more desirable than SMP).

_________________

@Hyperionmp

Quote:

Obstacles to an efficient implementation were removed (e.g. the use of Forbid) and replaced in many OS components over the years (e.g. DOS).The foundation for SMP support was put in place, a clear picture exists what needs to be done to accomplish it and how. I'm willing to take a bet that it won't take 2 years ;)

Sounds good. If it will take less that 2 years it will be better for everyone, but we all live in real world, so problems can happens because of millions reassons, and so, its all can take a longer way as usuall.

The better is to have no OS crashes anymore and no intuition lockups and freezes :)

Btw, while you write posts and read answers, maybe you can bring some light if you have in the pipeline support of any ppc-notebook (any kind of mac-laptop of course, just because they already exists), so, every aos4 programmer can works on apps even if he not at home.

For example i am for the last 3 weeks are not at home, and will be in next 2 weeks too. For that time, if there will be some sort of virtual machine for aos4, or support of any ppc-laptom where os4 works, i can do some programming for aos4. And as i know, many aos4 developers just have no time sometime, just because they not have their aos4 right by hands when they have time.

And why you say "Sorry to disappoint you" ? I saying just truth and how all of this happens in last 10 years. I will be only glad if os4 will develops faster, but all the facts point that is not (console handler are good example, which take 5 years, so its very logical assume, that SMP will take 5 or 10 years as well).

Last edited by kas1e on 03-Sep-2011 at 12:04 PM.

_________________
Join us to improve dopus5!
zerohero's mirror of os4/os3 crosscompiler suites

@kas1e

"nd why you say "Sorry to disappoint you" ? I saying just truth and how all of this happens in last 10 years. I will be only glad if os4 will develops faster, but all the facts point that is not (console handler are good example, which take 5 years, so its very logical assume, that SMP will take 5 or 10 years as well)."

Big difference between a spare time developer like Tony W and full time developers ...

_________________

@Hyperionmp

nice to hear, but from what i perceive about all os components take quite long to complete. good to see usb support improve at last and lets see what comes next, but looking at it as an outsider it pretty much as kas1e sees it.

@Hyperionmp

Quote:

a clear picture exists what needs to be done to accomplish it and how.

I guess one part of that clear picture is deprecating the 68k support?

Chapter 1 of the PowerPC Virtual Environment Architecture, Book 2 is quite clear on memory ordering being rather weak in multi-processor PowerPC systems unless explicitly enforced by memory barriers --- and even with explicit memory barriers, mis-aligned memory accesses are still non-atomic, and there are no guarantees whatsoever regarding coherency between processors.

While OS4's support for 68k code already has an inherent race condition (68k read-modify-write instructions were atomic on all 68k Amigas, but are not in OS4's task-based emulation), trying to run legacy 68k code (with about 20% of all memory accesses being mis-aligned) in an SMP environment would take ignoring-race-conditions-and-hope-for-the-best to a whole new level....

(Life isn't all that much more comfortable in PowerPC land, either. Something simple like a producer/consumer setup with shared memory FIFO becomes *really* tricky when implemented on something with as few memory order guarantees as the PowerPC. Add a number of guarantees that OS4 currently makes, but which no longer hold in an SMP environment, and "two years" starts to sound like "March 2002"....)

IMHO, SMP is not mandatory for future AmigaOS. I rather see some form of ASMP than a buggy SMP (or some compromises with the rest of the system). The bottom line is that one day we must be able to use those other cores, somehow. It does not have to be the same way that is used in the mainstream and it does not have to be the best way in every aspect.

Perhaps some PPC virtualisation features can be used.

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

@KimmoK

Quote:

I rather see some form of ASMP than a buggy SMP (or some compromises with the rest of the system).

Guess why true ASMP haven't been used since late 60's? Because it's horrible cluster***k.

@Rose

I meant asymmetric multiprocessing like some kind of new WarpUP or second CPU as a cluster node (good "enough" for 3D rendering and video conversions) or just some special way in juggling around tasks.

Last edited by KimmoK on 03-Sep-2011 at 08:18 PM.

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

@zerohero

Quote:

Gee, thanks. You people are so helpful...

Hmm.. I would have thought most people understood your sarcasm the first time..
Hope they get it with this comment.

_________________
/Harry

[SOLD] µA1-C - 750GX 800MHz - 512MB - Antec Aria case

Avatar by HNL_DK!

slightly OT but amigakit now sell 2GB RAM for.... X1000
http://amigakit.leamancomputing.com/catalog/product_info.php?products_id=1061

_________________
AmigaOne X1000