@EntilZha

Quote:

AFAIK, there's no possibility to find out wheter code is 68k or PPC in MOS. This is a no-brainer in OS4, so where is MOS more mature ?

You can know whether the task is in 68k mode at any given time, of course, but since 68k code and ppc code are executed transparently, I'm not quite sure why you would need more than that, or need that at all?

Last edited by falemagn on 07-Dec-2004 at 10:35 AM.

_________________
“It is well enough that people of the nation do not understand our banking and monetary
system, for if they did, I believe there would be a revolution before tomorrow morning.”
~~ Henry Ford

Remembering posts from Titan computers in the past on various websites and forums I would guess that financial reasons have nothing to do with it.

As for Papyrus Office it was known as Crapyrus Office by those that used to use OS/2 around where I used to work. Not used it myself, not really bothered to be honest.

As for emulation I seriously doubt there is any reason why a MorphOS emulator could not be developed for OS4, the question really comes down to "why would anyone want to?".

@fryguy
Quote:

What transfering speeds can you get with Roadshow on OS4?

Here the fastest speed I can achieve is around 10MB/s in and 8MB/s out.

Optimal speed depends on many factors though:
Transfers between A1 and Amithlon or QNX are much faster and steady than between the A1 and XP.
Both machines have 3Com chips and are connected directly with a cross-over cable.
Using a router/switch could slow the traffic a bit.
Reading writing from/to ram is probably faster.

For the internet it doesn't really matter anyway, as most people get less than 5Mbps (500KB/s) from their ISP, e.g. my cable provider caps it at 3Mbps down and 256Kbps up.

Alex

_________________
AlexC's free OS4 software collection

AmigaOne XE/X1000/X5000/UAE-PPC OS4 laptop/X-10 Home Automation

@falemagn

Hooks, for example... or about any callback like interrupts. In OS4, you just stuff a pointer to your interrupt routine into the is_Code field of an interrupt, regardless of whether it's 68k or PPC, and it's either emulated, or executed natively, depending on the code type.

_________________
Thomas, the kernel guy

"I don't have a frigging clue. I'm norwegian" -- Ole-Egil

All opinions expressed are my own and do not necessarily represent those of Hyperion Entertainment

@DaveP

Quote:

Remembering posts from Titan computers in the past on various websites and forums I would guess that financial reasons have nothing to do with it.

A few days ago I have been in contact with Michael Garlich regarding these issues.

However according to him there is nothing against an OS4 port of Papyrus, but they want money guarantees and developer machines.

@EntilZha

Quote:

Hooks, for example... or about any callback like interrupts. In OS4, you just stuff a pointer to your interrupt routine into the is_Code field of an interrupt, regardless of whether it's 68k or PPC, and it's either emulated, or executed natively, depending on the code type.

We've already had this conversation, remember? MOS can do that just as well, if not better, at the expense of a slight source code incompatibility (Old code usually uses ASM anyway in those parts, so it's not big deal). and I'd say that jumping to a function which checks for a particular opcode (mos) is faster than jumping to a function which has the burden to check whether the code is 68k or not (frame unwinding and PC address masking?).

Last edited by falemagn on 07-Dec-2004 at 10:48 AM.

_________________
“It is well enough that people of the nation do not understand our banking and monetary
system, for if they did, I believe there would be a revolution before tomorrow morning.”
~~ Henry Ford

@falemagn

Quote:

We've already had this conversation, remember?

Not really (remembering, that is, I'm not saying we didn't).

Quote:

at the expense of a slight source code incompatibility

Not in OS4.

Quote:

I'd say that jumping to a function which checks for a particular opcode (mos) is faster than jumping to a function which has the burden to check whether the code is 68k or not (frame unwinding and PC address masking?).

Why that ? Maksing the address is really not a problem.

Besides, in OS4, this also applies to 68k code: 68k code can call PPC hooks or even PPC function pointers without the need for special treatment.

Last edited by EntilZha on 07-Dec-2004 at 10:58 AM.

_________________
Thomas, the kernel guy

"I don't have a frigging clue. I'm norwegian" -- Ole-Egil

All opinions expressed are my own and do not necessarily represent those of Hyperion Entertainment

@DaveP

Quote:

Probably around the same time MOS1.5 comes out.

That would be, never.

The posts about OS4Emu proving MorphOS is somehow better are pure bull####. You could emulate MOS or OS4 on a C64, if you somehow attached enough RAM to it to hold all the PPC registers and boot the OS.

You can emulate anything on top of anything else with enough RAM and glue code.

I happen to know for a fact that a similar wrapper for MOS is possible for OS4.

_________________
Playstation Network ID: xeron6

@xeron

Quote:

The posts about OS4Emu proving MorphOS is somehow better are pure bull####.

No, the whole issue which OS is better or more advanced is pure ####. And this thread falls nicely into that category.

@falemagn

Quote:

falemagn wrote: We've already had this conversation, remember? MOS can do that just as well, if not better, at the expense of a slight source code incompatibility (Old code usually uses ASM anyway in those parts, so it's not big deal). and I'd say that jumping to a function which checks for a particular opcode (mos) is faster than jumping to a function which has the burden to check whether the code is 68k or not (frame unwinding and PC address masking?).

GIven the fact that you don't know how IsNative() is implemented I don't think you are in a position to tell weather it is fast or slow. But I find it rather amusing that you have to restort to things like 'which has the burden' (well, burden sounds hard yes) and 'frame unwinding ' (just because things are complicated or hard to understand doesn't mean they are slow). So could you just refrain from making comments about things you don't know anything about and stop trying to convince people with ghost tactics?

Now, given the fact that I know that IsNative is quite fast and I assume that you know the fact that emulating the emul gate instruction (this is what you claim, right) I think that we can conclude that there is no real speed difference between the 2 solutions.

As a side note, the number of if(pc & 1) hacks in the morphos version of ixemul are many. The os4 version is simply alot cleaner. And if 10-50 cycles are wasted due
to that (I am not saying they are) I'd gladly sacrifice that for a cleaner implementation.

regards,
Stefan

ps.
For you who wonder, the complicated 'frame unwinding' is a simple 'addi r1,r1,x' instruction

@Samwel

Quote:

The specs promise a lot of things, but all are not in the current 1.4.2 version. Alot was to be fixed in the next release, which maybe never will be released..

From my perspective the Abox has already grown far beyond what was *originally* planned by its developers. At the early stages of MorphOS development Ralph only had a Quark microkernel and some time later a complete PPC re-implementation of the orginal AmigaOS 3.1 exe kernel hosted on top. AFAIK at this point this kernel functionality was solely meant to lure AmigaOS users as potential customers as it allowed people to run Workbench 3.x. But a native environment would at some point replace Workbench and this re-implemented kernel.

As most people know the MorphOS team/bPlan was unable to get a deal for bundling AmigaOS3.x with their solution for whatever reasons. So this left Ralph with a rather useless re-implementation of the OS3.1 exec kernel (and other libraries) in PPC.

Only due to AROS' already done re-implementation of AmigaDOS and the hard work done especially by people like David Gerber (Ambient) and Stefan Stuntz (MUI) the ABox environment eventually turned into something useful.

IMO based on my communications with Bill Buck he was obsessed by destroying Amiga Inc and anything related. Wayne said something similar on AO a week or so ago. Also within an email threat directed towards Jens Schönfeld BBRV claimed to have "steamrolled the competition". I think this obsession was one reason why the ABox has been pushed so far by Genesi, other reasons of course include mainstream visibility and recognition as AmigaOS had many applications, supporters and developers.

Now with the ABox market fairly satured and inability to claim that AmigaOS4 doesn't really exist (as meanwhile over a thousand people have a pre-release), this market isn't much of interest to Genesi's management anymore.

Last edited by MikeB on 07-Dec-2004 at 01:30 PM.

@MikeB

Quote:

to claim that AmigaOS4 doesn't really exist (as meanwhile over a thousand people have a pre-release)

Their post is dated 28.4.2002. Was there "public" OS4 or not back then, I won't comment.

@hooligan

AmigaOS4 did exist back then (02-28-2004), but it wasn't available yet to general users. There's a big difference. Now it's useless to make such statements.

Last edited by MikeB on 07-Dec-2004 at 02:58 PM.

@Yabba

> And if 10-50 cycles are wasted due
> to that (I am not saying they are) I'd gladly sacrifice that for a cleaner implementation.

They aren't. The isNative() function is really fast, and offers you the possibility to call e.g. hooks from the PPC side without having to check for special opcodes. The aim behind the design in OS 4 has always been to favorize PPC-code over legacy code, i.e. having the possibility to build PPC libraries and applications without any legacy ballast. Emulation gates *are* legacy ballast.
The fact that the OS 4 68k emulator is slower than the MOS one results from the fact that MOS has a JIT. This will change soo when the integrated version of Petunia becomes public, be it with a subsequent update or the final version.

_________________
"Who do you think is this guy - god ?"
"No, god knows mercy !"

@MikeB

I think it's a little pedantric to start questioning his sentence.. he clearly meant there is no OS4 for the masses. But of course anyone can interpret the sentence as he/she wishes to.

Last edited by hooligan on 07-Dec-2004 at 01:07 PM.

@hooligan

Quote:

I think it's a little pedantric to start questioning his sentence.. he clearly meant there is no OS4 for the masses. But of course anyone can interpret the sentence as he/her wishes to.

You could compare the situation if Amiga Inc management would be claiming: "There is NO QBox. "

Not a good idea, if you ask me. IMO without providing proof & context such a statement can only be regarded as FUD.

Last edited by MikeB on 07-Dec-2004 at 01:21 PM.
Last edited by MikeB on 07-Dec-2004 at 01:08 PM.

@MikeB



Quote:

You could compare the situation to Amiga Inc management claiming: "There is NO QBox. "

In which case I would agree with AInc for the first time in world's history.

Last edited by hooligan on 07-Dec-2004 at 01:10 PM.

@hooligan

I would disapprove of making such statements by AI management, because:

1) I believe Amiga Inc management should concentrate on their own products,
2) The QBox may actually exist as has often been touted by Genesi when being confronted with limitations and issues surrounding the ABox environment.

There's one difference though, the QBox han't been seen by anyone outside of its developer team yet, but AmigaOS4 was already being demonstrated at many Amiga fairs, product demonstration events, shows, parties and user meetings at that point of time.

Last edited by MikeB on 07-Dec-2004 at 01:22 PM.

@MikeB & Hooligan

Let's agree that it is dangerous in general to try to figure out what BBRV means or doesn't.

( Example " PegasosG4 exists " In Aachen 2002. )

And in the end. It means nothing. Any Bill & CEO combination sucks, always.

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