@Spirantho

by same logic, neither the 65xx family is dead... or 68k/CF.

But... there is an very obvious trend in the industry signaling a slow, but steady move away from PPC over to competing products... And Freescale doesn't seem capable of reversing the trend - if anything, by introducing ARM cores in their telcom products they seem like slowly giving up and admiting defeat -I don't know how many years it will take, but to me it seems PPC is on the verge of becoming irrelevant...Wii U might be it's swan song.

IBM works only on custom products and their POWER line is highly specialized so PPC might have a future there but that future will not hold consumer products.

_________________

@WolfToTheMoon

Freescale's counter attack is not active yet.
It will take one year before we see what happens.

PPC had monopoly in telecom. That's why it's development practically stopped for years. Now there is also 6...32 core MIPS on the market. So PPC is not alone any more, they needed to start to compete.

ARM is not as powerfull as PPC. Neither is MIPS. So PPC has change to continue to keep good marketshare in telecommunication etc. ARM is smaller core than PPC, that's why it has some use in the lowest power use setups.

" It's been losing market share to x86 and ARM. Hence the Freescale ARM announcement."

You've got it wrong there. Freescale woke up because of MIPS. MIPS is the biggest competitor for PPC. ARM is not yet in telecommunication, it's only going there, to the lowest end and as a small security core beside PPC core.
Intel became PPC competitor in high performance embedded computing (that's why SIMD is back in PPC).

Because markets are changing, also freescale makes changes. They can not any more rely on Telecommunication + Military, because PPC is not the only one there any more.

I think it is possible that we see a PPC also for handheld devices, because PPC needs to counter attack to be on the safe side.

Last edited by KimmoK on 10-Sep-2012 at 06:43 AM.

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

@KimmoK

Applied Micro will present their 64bit ARMv8 CPU later this year or early next. Up to 128 cores, up to 3 GHz... Somewhat later nVidia will introduce their 64bit ARM chip.
And there are soon A15 Cortex SoCs on the market. ARM has the power to challenge PPC even on the high end, if not right now, it will be very soon.

And Freescale is just too damn slow... they're almost always late with their products and they rarely meet all their design goals.


I don't think they stand a chance in hell on the handheld market. Look at how much money Intel is throwing at it and they'll still pretty small there. Freescale is just too slow and probably not rich enough to compete with various ARM makers and Intel. Intel can allow itself to burn money for years to establish a foothold in the market... I very much doubt Freescale is in the same position.

Last edited by WolfToTheMoon on 10-Sep-2012 at 07:10 AM.

_________________

@WolfToTheMoon

I don't understand why producers doesn't follow the same schema with PowerPC. Many PPC core are licenceable. Maybe those IP costs more?

https://www.power.org/kshowcase/view/catalogs_by_category?categories=614ae4fcf15b15d6849dd6611f69855bc4a916da

http://www.linleygroup.com/newsletters/newsletter_detail.php?num=4048&year=2010&tag=3

This is a customizable processor. It's seems interesting. It's a Risc processor.

http://www.tensilica.com/products/xtensa-customizable#features

Last edited by AmigaBlitter on 10-Sep-2012 at 08:46 AM.
Last edited by AmigaBlitter on 10-Sep-2012 at 08:44 AM.

_________________
retired

@AmigaBlitter

Yes, there must be some "issues" on PPC licencing. Perhaps Irin Gould / Medhi Ali is there somehow.
At this rate the best RISC core will fail.

As long as only mobile devices used ARM chips, it's roadmap did not seem good for desktop-wannabe-devices. But in a year, a lot can change. So far ARM is not a sane option for our niche.

"What PPC CPU could be a good candidate as the next Amiga chip?"
Several options exist.
APM82181 - with this you can do 200eur computer that outperforms SAM440
APM86290 - with this you could do some 400eur computer that outperforms SAM460
T2080 - with this you could do the most powerfull Amigalike computer so far (30kMIPS + SIMD) Also this kind of computer should cost less than 1000eur.
T4240 - This could give us 130kMIPS + SIMD. But CPU cost would drive the price far north from 1000EUR.

We need SMP/SMP now.

Last edited by KimmoK on 10-Sep-2012 at 10:08 AM.
Last edited by KimmoK on 10-Sep-2012 at 09:21 AM.
Last edited by KimmoK on 10-Sep-2012 at 09:12 AM.
Last edited by KimmoK on 10-Sep-2012 at 09:11 AM.

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

@WolfToTheMoon

Quote:

WolfToTheMoon wrote: @Spirantho by same logic, neither the 65xx family is dead... or 68k/CF.

If people are still using them actively like they are PPC, and still developing them actively like they are PPC, then they're not dead. If, however, they're not actively used and developed then no, my statement does not apply to them.

Please don't deliberately take my logic to extremes to try and discredit it - you know what I meant and so do other people here; there's a world of difference between the users and developers of embedded PPC and old CF and 65xx.

@KimmoK

Quote:

I think it is possible that we see a PPC also for handheld devices, because PPC needs to counter attack to be on the safe side.

I don't think so, that would cut into Freescale's i.mx sales and Freescale is spending boatload of capital on i.mx for the consumer market. Which sounds more reasonable for Freescale to do, spend the money on adapting new generation of ARM (like the A15) technology or create new (modern) PPC cores that cut the cord from old core technology? Only thing I've been seeing from Freescale on PPC is adding more old school cores with a few added tricks to stay competitive in the router market.

Let me remind you of the Linux benchmarks that were posted around Spring, ARM (Galaxy II and iPhone4) kicked PPC's ass. Freescale has no reason to do much with the PPC series other then to keep it limping along for router sales.


Nice quad core i.mx6 $250 tablet: http://youtu.be/uGRCFFLcZ3Q

_________________
Dammy

@damocles

Yes, perhaps it's not Freescales task to do, but power.org or IBM should consider it.

"Let me remind you of the Linux benchmarks that were posted around Spring, ARM (Galaxy II and iPhone4) kicked PPC's ass"

I must admit that I do not remember them.
Anyone got link to those benches?
(A9 MIPS rating per core per Mhz does not hint that it outperforms PPC)

Other than that, I believe mx6 was the first ARM to have PCIex1 on board.
Time will tell if (or some else) it gets more. (would be nice to get that info)
Untill there is more PCIe available I doubt ARM will be used much on dekstop.

Last edited by KimmoK on 10-Sep-2012 at 12:31 PM.
Last edited by KimmoK on 10-Sep-2012 at 12:16 PM.

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

@KimmoK


Sun Spider chart has the ARM devices included.

_________________
Dammy

@damocles

Ok, it's then benchmark of JIT vs nonJIT + not the same SW
Dualcore A9 can not otherwise be 4x faster than G4.

A15 seems to be the first one to deliver more than previous PPC generation (not as much as the high end of PPC next generation).

If our OSs do architecture switch, it should be to chips that are usable everywhere. And the decision should not be made before there are real chips to use.
(most likely PDAs will take ARM ahead of PPC in RAW CPU performance, but would their built in GFX be enough for 3D if the PCIe offerings do not match... we will be wiser next year, perhaps)

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

@KimmoK

Quote:

Ok, it's then benchmark of JIT vs nonJIT + not the same SW Dualcore A9 can not otherwise be 4x faster than G4.

It doesn't matter to the consumer, they know what is fast and what isn't fast because it's happening right in front of their eyes. If the app/game is running like a slug on a faster hardware, why bother in the first place? Consumers sure as hell won't and you can take that to the bank.

_________________
Dammy

@damocles

That is one way to look at it.

But if you install AOS on those ARM devices, would OWB suddenly be faster? No. it would be slower than on PPC.

(A9 at 1.2 Ghz would be slower than G4 1.1Ghz etc. Btw, no matter on what HW AOS would run, it would look bad (to some people) because it's browser does not have JIT for javascript etc...)

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

@KimmoK

Quote:

But if you install AOS on those ARM devices, would OWB suddenly be faster? No. it would be slower than on PPC.

WebKit ARM JIT has been around for awhile.

Quote:

(A9 at 1.2 Ghz would be slower than G4 1.1Ghz etc. Btw, no matter on what HW AOS would run, it would look bad (to some people) because it's browser does not have JIT for javascript etc...)

Like that tablet I shared the video of, it's multi-core which has more then enough horse power. Like Terminills pointed out a few days ago.

If you can't compete against a $250 consumer device because it's faster plus it has richer capabilities, it's time to step back and examine the big picture on the reason to continue development.

Last edited by damocles on 10-Sep-2012 at 08:16 PM.

_________________
Dammy

Because PPC is going to die.?
ARM has a lot of time there since 1987, little known by users do not seem to have much success until recently mobile devices have not yet died, continued to develop. So why will stop to exist PPC?
I think they are just fads and what now appears to be dead is alive tomorrow. I think if Amiga new generations pass cpu cpu surely becoming obsolete (because of bad luck Amiga)
Best stick with big PPC after all was the natural successor to 68k and Commodore had chosen.

As for the x86, has been saying for the first Pentiums is the future, x86 will be here forever, it seems that will never go, I do not want it next Amiga.

_________________

@natamix

Quote:

Because PPC is going to die.?

There will always been a need for router CPUs and PPC fills that niche for now.

Quote:

ARM has a lot of time there since 1987, little known by users do not seem to have much success until recently mobile devices have not yet died, continued to develop.

That would be because of ARM was the single product, AFAIK, for the ARM group to push forward as it was do or die for them. I don't see IBM being all that interested in PPC (vs money for POWER CPU) development and Freescale is selling what it see it can make a profit on and that appears to be more and more focused on ARM SoCs for large scale production.

Quote:

Best stick with big PPC after all was the natural successor to 68k and Commodore had chosen.

Commodore never chose PPC and the natural successor to the 68K is Coldfire.

Quote:

As for the x86, has been saying for the first Pentiums is the future, x86 will be here forever, it seems that will never go, I do not want it next Amiga.

It already is for the Amiga. AmigaOne, OTOH, I agree, will most likely remain with PPC. My question was not about hardware development, I can see Freescale continue on with whatever improvements (more G2 cores or whatever) to keep the upper end routing capabilities their products need to be competitive. It's the big picture for the Community.

_________________
Dammy

@damocles

BTW, Freescale profits plunged 76% down in the last quarter of this year...

_________________

@KimmoK

Quote:

A15 seems to be the first one to deliver more than previous PPC generation (not as much as the high end of PPC next generation).

You'll be able to buy them fairly soon. Should give ARM G5 type power.

Next year the Applied Micro 64 bit ARMs will appear at 2.5GHz complete with PCIe, SATA etc. I expect they'll be even faster.

Quote:

If our OSs do architecture switch, it should be to chips that are usable everywhere. And the decision should not be made before there are real chips to use.

Plenty of ARMs out there - they are made in *vast* quantities.
You can get 10" ARM laptops off ebay for £83 (new).

Quote:

(most likely PDAs will take ARM ahead of PPC in RAW CPU performance, but would their built in GFX be enough for 3D if the PCIe offerings do not match... we will be wiser next year, perhaps)

NOVA 3

That runs on a phone from last year, next generation mobile GPUs will also be out soon.



BTW If the laptop is too expensive look up "cloud stick". An ARM computer the size of a USB stick for £30.

_________________
Whyzzat?

@damocles
Well, the coldfire is another story, at that time would still development. PPC was the most direct competitor to x86 and manufactured by Motorola (besides others) so I say what successor.
Commodore chose Risc processor and surely had finished picking the PPC

Last edited by natamix on 10-Sep-2012 at 11:53 PM.

_________________

@natamix

Quote:

Well, the coldfire is another story, at that time would still development. PPC was the most direct competitor to x86 and manufactured by Motorola (besides others) so I say what successor. Commodore chose Risc processor and surely had finished picking the PPC

Actually Commodore chose a variant of HP's PA-RISC that was used in printers for Hombre.

Funnily enough, HP's printers went on to use Coldfire and became the main user of that processor.

_________________
Whyzzat?

@minator

PA-Risc 7200 if i remember correctly.

_________________