IBM's PowerPC 65-nm Plans?

[MacRumors]

AppleInsider reports that IBM is working on mobile/embedded 64-bit PowerPC processors under the PowerPC 300 series.

This information echos information from a previous MacRumors report.

In light of the ongoing difficulties in incorporating desktop-class processors (such as the PowerPC 970) into the PowerBook, Apple and IBM have laid the foundation for an architectural revamp of the PowerPC to produce an ultra low power 64-bit mobile processor.

The November report also aimed at 2005 for the debut of a low-power mobile PowerPC 300 series.

[spencecb]

Well, I guess that is a start...but here's hoping that they will figure it out sooner, because 2005 is more than a year away still, and that is an awful long time to wait to get the Pro processor into the the Pro notebook. Maybe they will throw out some cool things in the meantime, like dual processor PowerBooks...that would be sweet.

[AmigoMac]

3.5 GHz G5 PB.... maybe... but I'll wait a couple of months and will get the next version of the G4... unless the rumors lead me to wait a bit more...


cool, great for Apple

[tazznb]

I hope it comes to fuition much sooner, but I'm not complaining.

[Dont Hurt Me]

Originally posted by AmigoMac
3.5 GHz G5 PB.... maybe... but I'll wait a couple of months and will get the next version of the G4... unless the rumors lead me to wait a bit more...


cool, great for Apple

dont be so sure of another anything from motorola, after all we had the 1.42 a year ago and today the fastest g4 being sold is a 1.33? at this rate we will be back to 500 mhz in no time Just more rumors of moto going bye bye!

[QCassidy352]

well, it's nice to know they have a plan, but 18 months doesn't sound so great. Does this imply that the .90 nm 970s will *not* be suitable for laptops? i.e. that it will be 18 months until we can see a powerbook G5?

If that's the case, I hope apple switches the powerbooks to G3 + altivec at 1.5 - 2 Ghz ASAP.

[Steven1621]

let the PB rumors begin!!!

[beelesbob]

Embedded suggests much smaller than is sensible for laptops, especially high end laptops - it suggests that IBM are wanting to sell some ludicrously powerful chips for embedded systems, but under the line for powerbooks... I would expect Powerbook G5s to use some generation of the 970/980.

Bob

[the_mole1314]

And P4 will hit 4ghz by late 2004. Man, hurry up IBM!

[JW Pepper]

Newton upgrade cards anyone!

[Lanbrown]

They are probably going after one of the markets Broadcom is in. Cisco uses embedded processors from Motorola in some of their routers. In their high-end gear they use MIPS processors and they have started to use the MIPS from Broadcom in some of their recent gear. Cisco could want a more powerful processor for some of their line and an embedded 970/980 could fit the bill.

[ITR 81]

This doesn't mean the PPC970 90Nm is not going into the PB it only means this is a likely path Apple will follow after the PPC970. I believe mostly to reduce weight and power consumption. Remember smaller and the more the effienct a processor is the smaller the battery can be and the slimer the chassis can be resulting in a thinner and lighter PB.

I'm willing to bet the PPC350 will run 3+Ghz.

[bousozoku]

Originally posted by QCassidy352
well, it's nice to know they have a plan, but 18 months doesn't sound so great. Does this imply that the .90 nm 970s will *not* be suitable for laptops? i.e. that it will be 18 months until we can see a powerbook G5?

If that's the case, I hope apple switches the powerbooks to G3 + altivec at 1.5 - 2 Ghz ASAP.

.90 nm 970s would be amazing, but the current 90 nm units are too hot for convective cooling and the PowerBook case too slim for effective fans along with convective cooling. The iBook case might be better.

IBM's tests with PPC970 blades will be helpful in determining cooling alternatives. Blade/processor density is high and so is heat.

65 nm processors are a good idea but making certain they're stable is the problem. Breakthroughs are easier in the lab than they are in production.

[MacsRgr8]

Originally posted by beelesbob
Embedded suggests much smaller than is sensible for laptops, especially high end laptops - it suggests that IBM are wanting to sell some ludicrously powerful chips for embedded systems, but under the line for powerbooks... I would expect Powerbook G5s to use some generation of the 970/980.

Bob

I agree.
I expect the 90 nm 970 to be able to make it in a PowerBook somewhere 2nd half of 2004.
As someone already pointed out, the G4 seems to be slowing down, i.e. back to 1.33 GHz. I don't really expect any speedbumps of the G4 anymore.
On the other hand, the Mojave IBM chip could also become next year's PowerBook chip....

[Longey Nowze]

Originally posted by beelesbob
Embedded suggests much smaller than is sensible for laptops, especially high end laptops - it suggests that IBM are wanting to sell some ludicrously powerful chips for embedded systems, but under the line for powerbooks... I would expect Powerbook G5s to use some generation of the 970/980.

Bob

IIRC the G4 is also used as an embedded CPU, I think it's used it routers and stuff like that isn't that why most PPC CPUs come in daughter cards?
and didn't IBM just prototyped a dishwasher full of embedded CPUs? wasn't that very powerful? imagine how fast washing dishes will be! :P

Thanks
MaT

[Longey Nowze]

Originally posted by Dont Hurt Me
dont be so sure of another anything from motorola, after all we had the 1.42 a year ago and today the fastest g4 being sold is a 1.33? at this rate we will be back to 500 mhz in no time Just more rumors of moto going bye bye!

wasn't the 1.42GHz G4 an overclocked version? moto never had anything in their sites about the 1.42GHz G4, the highest they had was a 1.33GHz or something...

correct me if I am wrong
MaT

[Gymnut]

Moto who??

[pilotgi]

I wouldn't roll my eyes at the possibility of a 750VX in the PowerBook. It will have a 400 Mhz bus which is nearly two and a half times faster than the current bus with clock speeds near 2 Ghz.

This sounds like a good alternative to Moto chips to me, since I don't think the G4 will see any more improvements.

[dongmin]

Originally posted by Longey Nowze
wasn't the 1.42GHz G4 an overclocked version? moto never had anything in their sites about the 1.42GHz G4, the highest they had was a 1.33GHz or something...

correct me if I am wrong
MaT

The 1.42 G4 7455s used in Power Macs are different than the 1.33 ghz 7457s used now. It IS a step forward, even if it's a small step. The 7457s are supposed to be faster than 7455s at the same clock speed. The 7455 was probably stretched to the limit at 1.42 ghz (hence the windtunnel cooling) but the 7457 should have some room to grow, albeit not THAT much more room. And of course, the 7457 draw less power than the 7455s which in itself is huge.

I wouldn't be surprised if Apple did another iteration, possibly two with the 7457s. I believe the current ones don't have any L3 cache; it could be added to eek out more performance.

[altglbrs]

Originally posted by the_mole1314
And P4 will hit 4ghz by late 2004. Man, hurry up IBM!

Even if Apple/IBM only reach 3Ghz by late 2004 (which has been promised), we'll be at 75% of the top PC MHz rating. When was the last time we were that close? Sure, its a full GHz back, but percentage differences are much more meaningful in my opinion.

[Rincewind42]

Originally posted by Longey Nowze
wasn't the 1.42GHz G4 an overclocked version? moto never had anything in their sites about the 1.42GHz G4, the highest they had was a 1.33GHz or something...

They were simply the fastest chips that Motorola could make (and thus also the lowest volume). Basically since Motorola had such poor yields, Apple bought every G4 over 1Ghz, so Motorola couldn't produce any for anyone else, thus making the fastest G4 Motorola would sell 1Ghz.

Presumably Motorola could get higher than 1.33Ghz yields on the 7457, but since Apple doesn't need them for their high-end towers anymore, they probably don't care to push the limits yet.

[k2k koos]

Even if Apple/IBM only reach 3Ghz by late 2004 (which has been promised), we'll be at 75% of the top PC MHz rating. When was the last time we were that close? Sure, its a full GHz back, but percentage differences are much more meaningful in my opinion.

And numbers aren't the only thing remember?
We are comparing a P4 to a G5, quite a different processor, so the gains are even better in the real world, not dramatic, but still very good news.

The real fun starts when the OS and applications are all 100% 64bit native, which isn't the case yet, so the G5 still can't show it's true potential.

[howard]

what about that whole ibm g3 with altivec processor? that could boast the powerbooks over the next year and then when they pick up a g5 the ibooks will get the g3 altivec from ibm

[legion]

Originally posted by Longey Nowze
IIRC the G4 is also used as an embedded CPU, I think it's used it routers and stuff like that isn't that why most PPC CPUs come in daughter cards?
and didn't IBM just prototyped a dishwasher full of embedded CPUs? wasn't that very powerful? imagine how fast washing dishes will be! :P

Thanks
MaT

Actually, the processors used in the "dishwasher" IBM prototype supercomputer (which will be the foundation of the Blue Gene series of super computers including the one (Gene/L) that will wipe away any current SC (including the EarthSimulator)) uses PowerPC300 series chips (info can be found on the IBM Blue Gene webpages)

The concept (for the Blue Gene series) is hundreds of chips running at lower frequencies, packed tightly into a box verses fewer chips at a higher frequency with less density (current Power series of SC.) This is where the PPC300 is coming into play.

[tychay]

Originally posted by legion
The concept (for the Blue Gene series) is hundreds of chips running at lower frequencies, packed tightly into a box verses fewer chips at a higher frequency with less density (current Power series of SC.) This is where the PPC300 is coming into play.

I'm surprised Arn hasn't done a story on BlueGene/L as it has just hit the supercomputer list at #73. I question whether or not this chip is related to the PPC300 mentioned in the article. Where did you get that data?

Another PPC thing that has missed the Mac radar is the Playstation 3. If you do a net search on the "cell processor" which has patents/technology from Sony and IBM (to be produced at Fishkill), you should be suitably impressed.

Both BlueGene/L and "cell" use a network-centric approach to processing. BlueGene uses a dedicated network coprocessor; cell divides processing units into cells that are distributed in blocks across multiple peripherals (think of your TVR or DVD player giving CPU cycles to your PlayStation or PlayStations available in a tiered manner which can play the same games but have amazing frame rates or better rendering). This is great stuff and I hope some of it trickles down (or up) to the Mac world.