@cHaOs667

Quote:

The decision for the PowerPC is right. Why? Because i think that the Desktop is already dead

Quote:

and at the embedded market there rules the PowerPC in all its varations.

Not quite, PowerPC is also represented at the high end in supercomputers and servers, the fastest processor you can currently buy is called POWER5+ which uses the PowerPC instruction set. IBM are putting the next gen into mainframes, that'll secure PPC's future at the high end for a long, long time.

Quote:

How many Users does have Zeta today? Sometimes it looks like that there are less Users than AmigaOS has Users...

Yellowtab has (primarily in Germany) sold over 100 000 copies of Zeta. More than Be ever sold and several times the number of Amiga users these days.

The decision to go PPC was taken 10 (or more) years ago, a lot has happened in that time. Today if you were to male a decision it'd almost certainly be x86. PPC had an distinct advantage then, today it doesn't, Cell could change all that however.

@hammer
Quote:

What was the number Intel?s multi-core development again?

Of the 4 chips I mentioned 3 have brand new core designs, Freescale also have one in development, Intel have 1 new core I know of (Merom) and Itanium Monticeto, AMD have a new core design due in 2007.

Quote:

Only with SPU i.e. it's VMX is shown to be less effective than P4?s SSE and G5?s VMX e.g. refer to cloth simulation test. PS; SPU?s cloth simulation doesn?t run same code as Pentium IV due to SPU optimization.

The cloth simulation was by their own admission an early experiment yet it still came out 5X faster, they didn't do much (if any) SPU optimisation.

Now quote the figures for the terrain renderer...

Quote:

Anyway, it?s not technologically hard for AMD to produce an eight core K7 (minus L2 cache, 22 million for each K7 core) from 250 million transistor budget (with 1MB L2 shared cache). Unlike Cell, this eight K7 core will retain OOO (out-of-order) processing. This is already possible with 90nm process.

Putting a large number of cores on a single chip is no simple operation, a lot of the work on the Cell went into the I/O and memory infrastructure, it'll be more complex with x86 due to L1 cache coherency. 1MB L2 cache would be far to small and even then it'll have a high(ish) latency due to contention. It'll also need an absolute beast of a memory controller which again will have high(ish) latency.

There's also the question of what sorts of apps could use a large number of OOO cores, Sun's Niagara uses very simple in-order cores but for servers it's an absolute monster, it even beats an 8 way POWER5 on one benchmark.

AMD are planning multi-core chips, 4 way in 2007 and 8 way sometime after. Both they and Intel are talking about including simplified or special purpose "cores" sometime after, sounds very Cell like. Cell is starting with a major advantage for high compute stuff, x86 is going to have a very hard time catching it.

_________________
Whyzzat?

@Hammer

US gov at the federal level is not yet interested in macs for it agents. (only the CIA issues Macs to some of it's civilain workers) But as a whole the US gov is Wintel.

US state agencys are also Wintel dominated.

Look logical- think about it Mr. Jobs has a responsibility to his share holders and right now they are happy because of the Ipod and they want to hear about mor innovative research and development on additional Ipod products and services (Video pod and the intergrated phone Ipod/Video pod)

The share holders are not interested in the Macs and Os X. Remember the mid 90's when Apple was on the edged becuse its Macs had become expensive and lackluster and all its major developers like Adobe where releasing the same products on the Mac to the PC.

It took Jobs to remove the squar box of the PC and introduce the hip fashion IMac and all the Apple diehards where back. Why? It was Fashionable to have a Mac again because the design (aqua) was hip even the PC makers try to copy it.

Now here we are again with Apple PC not doing to well but The revolutionary Ipod breaking sale records and helping sagging mac sales so Apple coporation can post positive figures to wall street.

Lets face it and be logical about it If you were Mr, jobs and had a killer product and app. would you pull the Irving Ghoug or Medi Ali? No, your are Steve Jobs you understand about capitalizing on someone elses dumb mistake (Sony for never taking the WalKman digital or developing a successor like the Ipod). Being Steve Jobs you are going to run with this product and cut out everything else that would drag you down.

Do not be suprise if you see Apple selling off it Pc division- 'Hint Apple has position it's self well by being able to run on intel chips so that a Major Chineness pc delveloper may aquire it's Pc division -while Apple concentrates on its mobile devices. Remember IBM never Put PPC in it's consumer PC even though it created them.

Quote:

Hitback wrote: Do not be suprise if you see Apple selling off it Pc division- 'Hint Apple has position it's self well by being able to run on intel chips so that a Major Chineness pc delveloper may aquire it's Pc division -while Apple concentrates on its mobile devices.

It was already the Chinese who made the PPC machines for Apple.

@BlueTroll

Missing the Point here guys. Wall Street and investors are not looking at long term investments but short term gains. Apple has made it self atrractive ( the old gal put on some nice sexy hi heels - a nice above the knee skirt- a nice soft but contouring to the body blouse- her nails/toes are french manicured- her hair all doneand flowing long) all to attract a suitor for its allling PC division.

Another chineness competior to Leveno is sniffing around and With Apples new found embracing love for the IA it looks good- because it is the standard and Apple knew this if it wanted to sell its holdings to anyone. think Business here guys not OS or love of a platform. Jobs is looking out for the Money he must produces for his major investors.

Right now those investors are demanding further RD be done on the Ipod Perticually it swiss army knife IPOD which will include Music/Video/PDA/blackbery type servic/Cell phone capability. And Apple is Listening. Look at their recently filed earning reports and see wher all their profit comes from. Hint it aint from sells of IMacs- Mac-minis or Powerbooks. No NO NO NO it is all IPOD base and wall Street loves it. Now figure out what Apple will RD and what will take a back seat?

Last edited by Hitback on 11-Jan-2006 at 06:17 PM.

@Manu

Quote:

Manu wrote: I believe Amiga will remain hobbyist computer wathever platform there is so....bring it to x86 and save me 800 bucks.

A man after my own heart.

_________________
"If you want to tell people the truth, make them laugh, otherwise they'll kill you." - Oscar Wilde

@Hitback

I think "well marketed" better describes the Ipod than "revolutionary".

@Atheist

Quote:

Atheist wrote: Even APPLE wasn't enough of a consumer for them to continue with faster and higher CPU generations past 68060 75 MHz.

AFAIK Apple didn't even bother staying with 68k untill the 68060, they jumped to PPC after the 68040.

_________________
"If you want to tell people the truth, make them laugh, otherwise they'll kill you." - Oscar Wilde

@Mr_Capehill

Quote:

Mr_Capehill wrote: @Bobsonsirjonny There is AROS if you want AmigaOS on x86. I think we can't have cheap x86 hardware for AmigaOS because it's so hard to support everything ("hey, why AmigaOS doesn't run on *my* pc system?!1!?") and if you pick custom design, you are back to square 1, with expensive custom hardware.

AmigaOS4 doesn't support "all those PPC configurations" either, it only runs on AmigaOnes. If the AmigaOne were x86 based, it would have been based on an easily attainable board, just like Eyetech tried to do with the AmigaOne based on a Teron. It wouldn't have to be custom, it would just have to be "chosen." Instead of Eyetech looking at the Teron, and using it, they'd look at an Asus KTwhatever, or an MSI1234. Pick a board, just like Eyetech did, and call that board an AmigaOne, support it with OS4, and if the board becomes obsolete, pick another board, and support the new one. Even in the worst case scenario, like the board that was chosen becomes unavailable IMMEDIATELY after launch, no problem, pick another manufacturers equivalent as the new "revision" AmigaOne.

_________________
"If you want to tell people the truth, make them laugh, otherwise they'll kill you." - Oscar Wilde

@Rob

Well you could say that But lets look at History -

Ibm ( the dumest corporation I have ever seen) develops the PC but says will only sell a few hundred Mainframes is where its at so the let Ms retain the righ to dos an get 1 percentage of every sale of the pc plus they don't pursue the market.

Xerox develops the GUI and then says ah no one will ever use this So they let Jobs see it and what happens?

Jobs creates the Macintosh from what he saw at Xerox He wants to follow a stratagy of place one in every school. Bill see this an he creates Ms win and goes after the Business market Rest is history

Sony made a Big mistake by not bring its walkman up to date it instead decide to look at the game market and miss this whole digital music device craze. If sony would have developed the walkman to a Walkpod they would own the digital revolution now - Walkpod-WEG TV LCD-Digital Cameras-Laptop and DesktopPC- Playstation 1,2 &3. Apple got there first and the rest is history.

@firbodi

Quote:

firbodi wrote: @Dave Haynie There is an idiom in my native language that says: "You're sitting outside the wrestling arena and say knock him out!" Firbodi

Yea, but he built the arena, and invented the game and it's rules.

_________________
"If you want to tell people the truth, make them laugh, otherwise they'll kill you." - Oscar Wilde

@Hitback

Huge corporations have a tendency to have risk-averse people in charge. Even if some of their engineering talent makes something that's cutting-edge, the company as a whole will typically not follow through because of those high-level decision makers that don't want to rock the boat. For example, consider Apple, widely regarded as the most innovative computer company. Look at the periods in time when Jobs was in charge, vs when he wasn't. The only thing keeping Apple cutting edge is Jobs, as soon as he's out of the picture (the guy will have to retire some time) I predict some moron will run the company into the ground within 5 years.

Hi @itix

ISTR, back in November of 2002, that there was Petition to Amiga Inc. to make OS4 available for x86 PC.:

from treblesix message ->
My claim is and has been that AmigaOS, or a clone, should have been ported to a PC. I just bought a 2.6GHz P4/Celeron machine, including 80GB hard drive, DVD/CD-R/CD-RW drive, 256MB of DDR-DRAM, etc. for $199. Ok, it was a good sale, this is one of my son's Christmas presents. This runs many times faster than any "neo-Amiga" class PPC machine. And many other people have...

This was essentially the argument. I opposed it in 2002 because the A1 wasn't delivered. After two years with the A1-XE/g4, i agree with the petition now; i kind of unsterstand the whole hardware factor now.

A1

_________________

I just had an interesting realization.

Even if we had gone x86 using a single board to mitigate the driver problem, we still would be in a bind wrt hardware today. Why? Work on the current version of OS4 started in 2002 IIRC. Assuming that the engineering problems that had to be solved for the AmigaOne version would have been about the same for any given x86 motherboard (except for CPU and NB, all components in the AmigaOne are pretty standard), the time taken to get where we are (basically, OS4 final) would have been about the same; namely, OS4-x86 would have been "done" round about now. Question: what board/chipset that was widely available in 2002 is as available now (other than eBay/the bargain bin)? Remember that Hyperion only has so many people working for them, so by the time they have negotiated with Intel, SiS or whoever for chipset docs and rewritten the HAL for the new version of any chip, another version is out and the "current" one is obsolete. Imagine what would happen if they need to switch to a different chipset altogether because the manufacturer decides that the relationship with Amiga is not worth their while any more - you need start from scratch. The only people who can cope with that kind of development cycle are Microsoft and Linux, simply because they have so many people to throw at the problem (Apple does not have that problem, because they are Intel partners)

So, with hindsight it looks like those advocating custom boards we have control over (AmigaOne, PV, Amy'05) were actually correct in their thinking. Moreover, those advocating using embedded chips (eg those new all-in-one chips from Freescale) are even more correct, because embedded devices tend to have a much longer shelf life than desktop chips (as an extreme example, you can still buy 6502s and z80s in bulk, with the companies responsible for these chips having no indication of slowing down production)

@CodeSmith

I could not agree more with that statement. Apple was not a cutting edge company when t Jobs and his partner( I always forgrt the name of the other guy he did the US festivals back in the 80s) ran it it just happens that they stumbled on Xerox Parc( now there was innovation !!! but as you said the higher up could not see what the talent they assembled had really did) any way when he left Jobs brought in Scully ( I think that was his name) from some soda company (Pepsi) for a while Apple languished in pain as Job could not get his bearing and there was really no innovation in the PC world Just amber or green Pc dispalys. eventualy Jobs was force out by share holders and moved on to create Next coumputers. The year left Apple did more business than ever before but that could only last so long as PC and MACs became boring again Just a system to do spreadsheets and word processing

Enter Jay Miner ( Atari wonderboy) Jay actually if you looked at it created the grapich card market adding GPU to a little machine like the Amiga was revolutionary so revolutionary that Apple was afraid and Ms was not even thinking pass EGA/CGA

And you know the story the higher ups at C could not see what they had and wanted to concentrate on selling C128 and PC clones.

Not capalizing on this gave rise to house like ATI and NVida soon everybody had graphic cards and Amiga fell behind.

Now As I stated before Apple with steve Jobs did not revolutionize the ditgial music industry the just capitlized on Sonys abandaoing of it. The Sony higher ups just did not see the potential here

Thus all we are now waiting for is some engineer to develop som sort of 3d interacting system for computing some where and the higher ups to say this will never sell> then have them invite bill gates over and the rest will be history.

@Rogue

I would very much like to see an explanation on how Dave Haynie is missing the point.

Based on Dave Haynie's posts on the Team Amiga mailing list, where Dave have posted more or less regularly since Team Amiga was formed, I'd say he's rarely missing the point.

Vidar

Last edited by VidarL on 11-Jan-2006 at 08:03 PM.

@Hitback

Other companies had mp3 players on the market years before Apple. It's just that Apple marketed theirs better.

@Rob

Yes but who had it as stylish as Apple? No one again this is what I am saying Apple did not invent the digital player-Just like they got the GUI idea fron Xerox parc and improved on it they did the same with the music player. Sony was the leader and should have continued the evloution of the walkman to digital.

Look at all the other MP3 players an see how blan they look. But look at the IPOD it resembles the stylish Zippo lighters they look great!!!

and everyone wants to belong to a stylish hip crowd and right now the IPOD is hip.

@Hitback

its too big for my liking, doesnt fit in that stupid pocket in your jeans, where nothing ever fits, apart from those elongated MP3 players.

_________________
Dark Lord Design Wicked Solutions For Damned Problems

@minator

Quote:

The cloth simulation was by their own admission an early experiment yet it still came out 5X faster, they didn't do much (if any) SPU optimisation.

Actually they have applied SPU optimisation and to quote

Quote:

Porting to the Cell Architecture ... In order to port the prototype cloth simulator to the Cell processor, the code and data structures had to be restructured. First, changes had to be made to allow for parallelization of the algorithm. Even for more conventional multi-core targets, this is a challenging process. With the Cell architecture, there is the additional requirement to manage memory transfers between the main processor (the PPE) and the SPEs. Second, where feasible, scalar code was re-written to take advantage of the vector support in both the PPE and the SIMD units.

- Page 1, Alias Cloth Technology Demonstration for the Cell Processor

Quote:

Putting a large number of cores on a single chip is no simple operation, a lot of the work on the Cell went into the I/O and memory infrastructure, it'll be more complex with x86 due to L1 cache coherency. .

K8 can already handle an eight core processing through HTT protocols.
K8's Direct Connect link is just an internal version of HTT link.

The model for the current dual core Athlon 64 X2 is from a prototype dual Socket 754. The AMD64 ISA itself has already defined a 256 core per package.
Refer to latest version AMD SimNow for cache coherency for scaling beyond 8 processors.

Quote:

1MB L2 cache would be far to small and even then it'll have a high(ish) latency due to contention. .

Depends on the allocation of the L2 cache. OOO processing with smaller cache would be less effective. Such a processor makes it more reliant on software optimization.

Note that certain low end K8 Sempron (Palermo 90nm SOI) has 128 KB L2 cache. Eight core Palermo would total to 1MB L2 cache.

In game terms, K8 Sempr0n is still quite competitive against Pentium IV EE(advantages of having medium depth pipeline and integrated MCH).
AMD has a patent that shares K8's front end for two execution cores.

Quote:

It'll also need an absolute beast of a memory controller which again will have high(ish) latency.

Refer to AMD’s licensing of Rambus’s technology.

Quote:

There's also the question of what sorts of apps could use a large number of OOO cores, Sun's Niagara uses very simple in-order cores but for servers it's an absolute monster, it even beats an 8 way POWER5 on one benchmark.

Note that POWER5 has longish pipeline while Niagara is a 6 stage processor.
"The pipeline on Niagara is quite shallow, only 6 stages deep". 6 stage processor is less dependant on OOO HW and may have faster instruction retirement. Niagara's core can't be compared to PPE nor SPU i.e. both PPE and SPU has a longer pipeline depth compared to K7, K8, Dothan, Yonah, 7447A and ‘etc’.

AMD K8 core has 12 stage for integer pipeline with early exits techniques.

Last edited by Hammer on 11-Jan-2006 at 10:41 PM.
Last edited by Hammer on 11-Jan-2006 at 10:38 PM.
Last edited by Hammer on 11-Jan-2006 at 10:37 PM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@Hitback

Quote:

US gov at the federal level is not yet interested in macs for it agents. (only the CIA issues Macs to some of it's civilain workers) But as a whole the US gov is Wintel.

I was not referring to the federal US government.

Quote:

Do not be suprise if you see Apple selling off it Pc division-

Apple's HW manufacturing is already being done by ODMs. I don't see this as a major surprise.

Quote:

'Hint Apple has position it's self well by being able to run on intel chips so that a Major Chineness pc delveloper may aquire it's Pc division -while Apple concentrates on its mobile devices. Remember IBM never Put PPC in it's consumer PC even though it created them.

Hint Apple is still dependant on PC HW revenue and Steve Job’s has stated that shifting to Intel would not change their business model i.e. no OEM MacOS X licenses (at this time). Remember, IBM’s PC division doesn’t include its OS and application software group.

Michael Dell already stated they are willing to OEM license MacOS X on their Dell boxes. Note that Dell’s willingness to ship MacOS X in it's PCs is in stark contrast to Dell’s willingness to ship AMD processors in its PCs. MacOS X (X86) OEM license would present Microsoft a real world competition in the desktop OS market. It will directly attack Microsoft’s fundamental revenue streams.

Last edited by Hammer on 11-Jan-2006 at 11:19 PM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)