@umisef

Library reference counts are incremented and decremented in Forbid() state. But your point still remains valid. For example Intuition supports manipulating window flags directly but underlines you must use atomic instructions. Applications also depend on atomic instructions in their internal threading.

However, when you are not emulating OS but running our own OS natively you can take this into account and have a hidden dupe field for internal system flags etc. etc.

Last edited by itix on 04-Sep-2011 at 07:46 PM.

_________________
Amiga Developer
Amiga 500, Efika, Mac Mini and PowerBook

@umisef

Quote:

A 68k emulator should emulate a 68k. That includes making the same guarantees. The only way to do that with a task-based 68k->PPC emulator is to call Forbid()/Permit() (or Disable()/Enable()) for every single emulated 68k instruction that executes more than one memory access.

Calling Forbid() would not be enough because atomic instructions can't be interrupted by, eh, interrupts. It would be mandatory to call Disable() for every instruction. But those cases are really rare and only very few modern applications rely on atomic instructions (mostly those written in machine language or using machine language macros) and (as I wrote above) you can also take it into account in the OS level.

FORBID macro advertised by Commodore system developers is quite nasty, though

Last edited by itix on 04-Sep-2011 at 08:08 PM.
Last edited by itix on 04-Sep-2011 at 08:05 PM.
Last edited by itix on 04-Sep-2011 at 08:04 PM.

_________________
Amiga Developer
Amiga 500, Efika, Mac Mini and PowerBook

@umisef

Quote:

Take the old classic "ADDQ.B #1,TDNestCnt(A6)", which was compiled inline into plenty of 68k code, because that's one of many such macros in include_i/exec/ables.i[1].

While you make a good case for what would trap the emulator and possibly the reason some 68K games and software freeze upon loading it must be said that code doing things like Forbid() ASM MACROS directly is definately not supported in OS4.

Only code using the actual Forbid() function would be supported and even then going out on a limb as software employing such techniques as disabling multitasking would be warned against. OS4 with its 68K emulation integrates to a high degree 68K software; perhaps too high, as a rogue Forbid() will halt the entire system. Unless it waits. And 68K interupts are actually executed in an actual real PPC interupt. I thinks they've made that too close to the system there.

@Rob

Quote:

OS 3.5 and 3.9 mainly consisted of 3.1 with third party components.

Yes they did. And not everyone agreed with the choice. I accepted it as it was when I bought into these releases. Well it was exciting at the time.

Bit I did reject that VincED and put KingCON back in its palace.

Quote:

I think the 3.9 feature set was exceeded quite some time ago in the 4.0 days.

Counting features OS4 may have more but the OS3.9 feature set was never matched. Soon maybe but here are some examples:

Internet: You could use a GUI to go off/online and check status. OS4 relies on CLI commands. Don't know if Genesis supported internet sharing but it's not easy to do this on OS4.

Prefs Palette has a nice colorwheel. Now missing.

Installer. Does not support OS3.5+ extenstions. Doesn't even acknowledge them or ignore them forcing one to use buggy old installer for OS3.5+ applications.

Shell. Scroll bar and other features added. Gone in OS4. Basic OS3.1 shell in its place.

That's a short list off the top of my head.

I hope there have some betatester test the blender benchmark and see the real speed of the pa6.

http://www.eofw.org/bench/

there are several other MOS or OS4 Systems to compare.

If somebody tell its slow because driver for X1000 are not optimize or so, look at the sourcecode of blender when execute this benchmark code, if you not believe that Raytracing execute 99.9 % own code, so OS or driver speed doesnt matter in performance with this benchmark

Last edited by bernd_afa on 21-Sep-2011 at 05:00 PM.

@bernd_afa

Quote:

bernd_afa wrote: I hope there have some betatester test the blender benchmark and see the real speed of the pa6.

NDA

Unless/until there is a proper end user release all one can do is speculate good or bad things.

_________________
Power Mac G4 "Quicksilver 2002"
800Mhz, 1.5GB RAM, Radeon 8500
MorphOS 2.7 (Registered)
$225 total spent!

Is there anyway of telling how many Beta Testers received their boards?

I would think there would be a lot of YouTUBE videos up by now.

@Gmeanie

Quote:

Gmeanie wrote: Is there anyway of telling how many Beta Testers received their boards? I would think there would be a lot of YouTUBE videos up by now.

I think you missed something, I'll remind you of it just for your convenience (and avoid you searching just one post above):

N  N  DDD     AA
N  N  D  D   A  A
NN N  D   D  A  A
N NN  D   D  AAAA
N  N  D  D   A  A
N  N  DDD    A  A

_________________
AOS 4.1 : I dream it, Hyperion did it !
Now dreaming AOS 4.2...
Thank you to all devs involved for this great job !

@abalaban

i presonally know about 2 betatesters that received boards, and both confirmed that publishing photos of their boards is not restricted by nda. both have put a lot of snapshots online. i trust there is more borads delivered, just not everybody is doing that much fuss.
f.i. here:
http://www.a1k.org/forum/showthread.php?t=28870

@wawa

Pictures of the boards and even videos of the manufacturing process have been already published, so I don't see any problem with that. But people want to see now benchmarks and videos of AOS 4 and software running on it, which is what I think would be under NDA.

_________________
PowerBook 5.2 MorphOS 3.15
PowerBook 5.8 MorphOS 3.15
Amiga 1200 BPPC/BVision AOS4.1 FE

@emeck

yes apparently nothing what concerns os4.1 on x1k may be spilled. the user above reports the speed of the system as "insane", of course not backing it up by anything, and im not sure if he ever owned any ng os4 system. thoght his comments are very entusiastic (as they ever were) so there is no major letdown, it seems.

@bernd_afa

Quote:

I hope there have some betatester test the blender benchmark and see the real speed of the pa6. http://www.eofw.org/bench/ there are several other MOS or OS4 Systems to compare. If somebody tell its slow because driver for X1000 are not optimize or so, look at the sourcecode of blender when execute this benchmark code, if you not believe that Raytracing execute 99.9 % own code, so OS or driver speed doesnt matter in performance with this benchmark

bernd this is all fine and right but i dont think anybody will provide results openly at this time.

Since the OS is in beta, I don't see why there should be any need for benchmarks, since the system isn't fully optimised and therefore would probably be running slower than it otherwise would. And all that would achieve is to trigger another round of those very helpful posts, saying how it should be faster for that price, etc. etc.

I can post a screenshot of 4.1 running on my A1, it'll pretty much look the same as 4.1 running on an X1000 - I don't see the point in that. It's like comparing screenshots of Windows 7 running on a Sony Viao to Windows 7 running on an Alienware machine. Nonetheless, I'm sure the betatesters will post pictures in due course.

_________________
RobTheNerd.com | InstallerGen | SMBMounter | Atoms-X

@wawa

The user quoted by wawa, is long term AmigaOS user. He used to have SAM (Flex I think). He also owns a fast A1200 running OS4.1upd 3. However, I would consider him as very enthusiastic.... and if says it is insane fast, he means it is insane fast. He knows fast Systems very well from e.g. Amiganoid, where we had an AmigaOne G4 XE 1,266 Ghz OS 4.1.

Fair enough they are not allowed to post benchmarks, remember the endless discussion going on about the slow bootprocess, as Videos had been published by AmigaKit...

_________________
Geomarketing at
www.geobiz.de
www.gebietsplanung.net
www.geomarketing-consultant.de

@gerograph

I know what is really fast.At the amiganoid i was playing with a AONE from MIKKA with 1,266GHZ !!!!

_________________
Amiga600/Vampire2/PrismaMegaMix

@tommysammy

that's what I said. BTW I tried iBatch on Mikas Amiga = fast... did you try it on X1000 yet ?

_________________
Geomarketing at
www.geobiz.de
www.gebietsplanung.net
www.geomarketing-consultant.de

@gerograph

No not at this monent.Give me please some time
I have too much work with all here
The answer will come via PM

_________________
Amiga600/Vampire2/PrismaMegaMix

@thread

The AmigaOne X1000 Experience - Steven Solie

#6

_________________
This posting, in its entirety, represents solely the perspective of the author.
*Secrecy has served us so well*

@Hyperionmp

Quote:

Hyperionmp wrote: 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 ;)

Bet lost - and now?

_________________
My programs: via.bckrs.de
MorphOS user since V0.4 (2001)

@Zylesea

Implementation date and release date do not have to be the same.

http://blog.hyperion-entertainment.biz/?p=863

Last edited by Spectre660 on 13-Sep-2013 at 07:22 PM.

_________________
Sam460ex : Radeon Rx550 Single slot Video Card : SIL3112 SATA card