As we know the bounty for an OS4 release it has been reached, the A1 machine it's done to Marcin too !

http://amigabounty.net/sputnik/sputnik.html

Maybe a new version of Sputnik isn't too far, the beta 2 was released in March 2007

http://www.ppa.pl/khtml/index_en.php

So the question is, any news

Last edited by samo79 on 12-Jul-2007 at 07:27 PM.

_________________
BACK FOR THE FUTURE

http://www.betatesting.it/backforthefuture

Sam440ep Flex 800 Mhz 1 GB Ram + AmigaOS 4.1 Update 6
AmigaOne XE G3 800 Mhz - 640 MB Ram - Radeon 9200 SE + AmigaOS 4.1 Update 6

@samo79

I am in contact with Marcin now and then but don't have much to tell except that Marcin doesn't have an A1 yet and that a new beta (MOS) is on it's way.
I guess Marcin should get the A1 when the new beta is out, hopefully within some weeks.

_________________

He only needs OS4 for his Pegasos. Would be much easier.

@samo79

He's currently rewriting stuff. First, the memory allocations were quite slow, as some pages could do up to a million memory allocations. So he's redoing it in a way that will make it much faster. I believe this part is complete now. The other bit hes doing is rewriting the render engine as the old engine was tricky to implement some features, and this part was the source of most instability anyway. He was hoping to finish up on this version end of July, but dont hold him to this.

Yes I interrogate him weekly/monthly. :P

Would be more interesting if sputnik were released for 3.x. At least more amigans would get his hands on.

_________________
Amiga 1200 desktop. Apollo 030/50 Mhz 8mb ram + ClassicWB + Wb 3.1
Amiga 500 + ACA500plus 8mb + 30gb CF
Raspberry Pi 3b+ and Amibian 1.4
Mac Mini G4 1GB Ram with the butterfly!!

@jahc

Quote:

jahc wrote: @samo79 He's currently rewriting stuff. First, the memory allocations were quite slow, as some pages could do up to a million memory allocations. So he's redoing it in a way that will make it much faster.

Memory allocation time seems to be a much bigger problem on MorphOS than OS3/OS4. The author of FryingPan said that Morphos was much slower than Amiga OS4 (and 3.x) with memory allocations. From his website (I hope he doesn't mind me quoting it):
Quote:

i am aware that fryingpan takes some time to burn discs on morphos. i want you to consider certain patch - not for fryingpan, but for morphos itself. it turns out memory allocations are real pain in the butt on morphos. i've done some benchmarks and it seems that i can perfrorm ten million (10'000'000) allocations on OS3 and OS4 within few seconds (~5.6s and 1.8s respectively), while ten times less (one million, 1'000'000 allocs) take over 30 seconds on morphos. i've tried to reduce the allocation time by reducing security mechanisms (semaphores). to make allocations safe i tried to associate tasks with pools (so one memory pool per task), but the code turned out to execute 3x longer (due to extremely slow FindTask(0) command, that is unthinkable). So I would like you to give 'APooler' a try. You can find the utility here: Morgoth's Download Page And.. no, I won't reduce the allocations. Memory allocation is a system time critical operation. FP operates on memory alot (virtually every string now uses at least one memory allocation, many string operation create a temporary buffer, and the whole is object oriented to give me some control over behavior and memory control). I haven't done that for fun or to make fp a pain in the ass. i won't revert to one huge structure. Talk to MOS team to fix this flaw.

Still, reducing the number of allocs with something that's faster is probably going to help on all platforms; unless the new solution is still slower than memory allocs on OS4.

Hans

_________________
Join the Kea Campus - upgrade your skills; support my work; enjoy the Amiga corner.
https://keasigmadelta.com/ - see more of my work

@Hans

What I'm suggesting with the above, is that if excessive memory allocs are the problem then an OS4 port from the current code would probably be much faster than the MorphOS version, i.e., this rewrite wouldn't be necessary for the OS4 version.

Of course, it's Marcik's decision which platform he develops for.

Hans

_________________
Join the Kea Campus - upgrade your skills; support my work; enjoy the Amiga corner.
https://keasigmadelta.com/ - see more of my work

@Hans

Interesting…

Anyway think they have done hell of job on ExecSG, its where fast and secure.

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

@jahc

Thanks for the news

P.S. (OT)
How is procede WookieChat for AROS PPC ?

_________________
BACK FOR THE FUTURE

http://www.betatesting.it/backforthefuture

Sam440ep Flex 800 Mhz 1 GB Ram + AmigaOS 4.1 Update 6
AmigaOne XE G3 800 Mhz - 640 MB Ram - Radeon 9200 SE + AmigaOS 4.1 Update 6

@samo79

I just had a look, and it appears that AmiDevCpp doesnt support PPC-AROS yet, or at least not without some manual installation of other files somehow. Sorry about that!

@jahc

So the only way is to send the sourcecode to an AROS PPC developers ? :-/

_________________
BACK FOR THE FUTURE

http://www.betatesting.it/backforthefuture

Sam440ep Flex 800 Mhz 1 GB Ram + AmigaOS 4.1 Update 6
AmigaOne XE G3 800 Mhz - 640 MB Ram - Radeon 9200 SE + AmigaOS 4.1 Update 6

@jahc

Tell him to try out the AllocTaskPool() command (or whatever it's called) on MOS. IIRC every task has its own memory pool associated with it ready to be used. Should probably be a lot faster than the normal AllocMem() function.

_________________
- Don't get fooled by my avatar, I'm not like that (anymore, mostly... maybe only sometimes)
> Amiga Classic and OS4 developer for OnyxSoft.

@Deniil715
It's AllocTaskPooled(), but that won't help. First, it's a normal pool, with the only advantage, that you don't have to create/free a pool for your own needs yourself. Second, he used pools already, AFAIK. Also, people claiming such a rewrite wouldn't be necessary on OS4 are stupid and clueless. No matter how good, or how bad a memory subsystem in your OS, doing lots (and i mean LOTS) of memory allocations on the fly WILL cause a slowdown, and properly written software just doesn't do that. So if there is a way to avoid that, it's smart to do it. And that also applies to non-Amiga systems.

_________________
MorphOS, classic Amiga, demoscene, and stuff
"When a bridge is not enough, build a Viaduct!"
"Strip the Amiga community of speculation and we can fit every forum on a 720k floppy" (by resle)

@Hans
Quote:

The author of FryingPan said that Morphos was much slower than Amiga OS4 (and 3.x) with memory allocations.

I'd be happy to know what kind of benchmark he used that he got thoose results. I haven't found a way to reproduce his results, even that MOS was slower than the other two systems (i compared myself to his results, but i'll do my own benchmarks with my own code on OS3 and OS4 soonish), but not by that magnitude he talks about. Well, to be honest, that's not true, because i found ONE way to reproduce it, by creating/destroying a whole memory pool on every alloc, but no sane software should really do that... I was also unable to reproduce the FindTask(0); slowness he talks about. So, example code reproducing the results are welcome. Until that, i consider it as spreading FUD.

_________________
MorphOS, classic Amiga, demoscene, and stuff
"When a bridge is not enough, build a Viaduct!"
"Strip the Amiga community of speculation and we can fit every forum on a 720k floppy" (by resle)

@Chain-Q

Why not write to the author and ask for the test code he used? Perhaps if you had a look at it you could spot if he is doing something wrong. Also keep in mind that he probably uses MOS 1.4.x rather than the latest MOS beta (1.5/2.0?) that you use.

@COBRA

Because IMHO it's much more easier to shout "FUD" as soon as someone hurts the "beloved and naturally perfect baby".

@Chip
Quote:

Because IMHO it's much more easier to shout "FUD" as soon as someone hurts the "beloved and naturally perfect baby".

That behaviour is sooo familiar to me... I remember i saw it already, from some OS4 troll... maybe the one called... Chip, IIRC.

_________________
MorphOS, classic Amiga, demoscene, and stuff
"When a bridge is not enough, build a Viaduct!"
"Strip the Amiga community of speculation and we can fit every forum on a 720k floppy" (by resle)

@Chain-Q

Big own goal, personal attack & . And you call me troll.. sigh...

@Chain-Q

Chip is always being accused to being a troll, but he always speaks his opinion and posts facts to support his ideas, he can be wrong as everyone but he is not a troll. Usually are MOS guys who seem to bring us the light of the truth and often revel themeselfs as clueless trolls. So write this not only to Chip, but to a lot of MOS users that in these years have posted FUD on forums. I agree with Cobra, it is the simplest thing to do, ask Frying Pan developer, maybe is wrong, just ask.

Last edited by ikir on 17-Jul-2007 at 11:05 AM.

_________________
ikir

@ikir
Quote:

I agree with Cobra, it is the simplest thing to do, ask Frying Pan developer, maybe is wrong, just ask.

The point is, he posted numbers without any proof, and people just took them as facts. Then i say, i can't reproduce his results, and suddenly, i become the evil. Oh well... I will probably ask anyway.

_________________
MorphOS, classic Amiga, demoscene, and stuff
"When a bridge is not enough, build a Viaduct!"
"Strip the Amiga community of speculation and we can fit every forum on a 720k floppy" (by resle)