Intel's Plans for Low-Power Chips

[dglsmhn]

Cnet says Yonah will be released January '06

...which means there is a good chance we will see an intel PowerBook at MacWorld SF...

http://news.com.com/2061-10791_3-5842839.html?part=rss&tag=5842839&subj=news

 

[shawnce]

...which means there is a good chance we will see an intel PowerBook at MacWorld SF...

http://news.com.com/2061-10791_3-5842839.html?part=rss&tag=5842839&subj=news

...or an Mac mini, eMac, even possibly iMac.

 

[VicMacs]

...or an Mac mini, eMac, even possibly iMac.

here dies the powerbook G5

 

[Ghibli]

here dies the powerbook G5

Hmmmm, I tought it was dead a LOOOOOOOng time ago...

 

[~Shard~]

here dies the powerbook G5

It was alive to begin with?

 

[shamino]

Not necessarily. New buyers may be attracted to Intel Macs if they're capable of running Windows (and Linux) in addition to OS X and competitively priced compared with the alternatives.

I've written about this several times before.

Dual-booting should be possible. Assuming that Windows itself runs on the box (it does on the developer models), a boot loader shouldn't be a difficult thing to develop. I wouldn't be at all susprised to see a port of GRUB for the platform.

As for whether this ability will affect the market any, I have strong doubts. I have used dual-booting systems many times in the past. I've older Macs that can boot two different Mac OS's (an SE that can boot 6.0.8 or 7.5.5 and a Quadra that can boot 7.6.1 or 8.1), and several different dual-boot PC's (a 486 that boots DOS, Win95 or Linux, a dual-PPro that boots DOS, Linux or Win2K).

It's been my experience that it is just too much of a pain to work this way. Every time you need to switch OS, you must quit all your apps, do a full shutdown, and then reboot. After a short while, you get tired of all this and you pick one OS to work with and ignore the others.

Back in the days of DOS and Mac OS 6, it was different. You tended to reboot quite a lot anyway, so switching OS's was only slightly more annoying than switching apps. But today, when people typically leave systems running for weeks at a time without rebooting, it's no longer something you want to do a lot.

IMO, if Apple started marketing Macs as dual-boot systems, the market will split into two groups. Those who run Mac OS all the time, and those that run Windows all the time. Those in the first category won't care about dual booting. Those in the second category will ask themselves why they bothered buying a Mac in the first place when they could've spent the same (or less) on a PC designed for Windows.

IMO, the best way to give customers the best of both worlds is to give/sell them Virtual PC or some equivalent. When the different apps can all run at once, you will use them all.

 

[sjk]

I've written about this several times before.

I never doubted someone hadn't and even apologized for any redundancy... I can only read so many posts in a day.

As for whether this ability will affect the market any, I have strong doubts. [...]

It's been my experience that it is just too much of a pain to work this way. Every time you need to switch OS, you must quit all your apps, do a full shutdown, and then reboot. After a short while, you get tired of all this and you pick one OS to work with and ignore the others.

Same conclusion here with my brief dual-booting experience.

What could make it more tolerable is a system suspend/resume capability, like hibernation on Windows or sys-suspend on Solaris. There are other useful reasons for having that feature, like saving/restoring your environment when moving your Mac mini or iMac to another room. Once I ran sys-suspend on a Solaris system, moved it to my new house, and booted it back to where it had been... it even reconnected to the internet audio stream I'd left playing.

IMO, if Apple started marketing Macs as dual-boot systems, the market will split into two groups. Those who run Mac OS all the time, and those that run Windows all the time. Those in the first category won't care about dual booting. Those in the second category will ask themselves why they bothered buying a Mac in the first place when they could've spent the same (or less) on a PC designed for Windows.

Maybe.

I'd think certain business and schools that want to run both OS X and Windows might find advantage to using the same systems for both. Apple could sweeten the deal for those buyers.

IMO, the best way to give customers the best of both worlds is to give/sell them Virtual PC or some equivalent. When the different apps can all run at once, you will use them all.

That would be ideal for some of us even if it's messy from a usability perspective (which nagromme pointed out in some other thread).

 

[Ghibli]

IMO, if Apple started marketing Macs as dual-boot systems, the market will split into two groups. Those who run Mac OS all the time, and those that run Windows all the time. Those in the first category won't care about dual booting. Those in the second category will ask themselves why they bothered buying a Mac in the first place when they could've spent the same (or less) on a PC designed for Windows.

IMO, the best way to give customers the best of both worlds is to give/sell them Virtual PC or some equivalent. When the different apps can all run at once, you will use them all.

I'm with you for this one, but I think that there will be some interesting possibilities: just imagine running Windoz apps INSIDE Mac OSX.. seems weird? maybe not. See

http://www.hardmac.com/news/2005-08-11/

and

http://www.codeweavers.com/about/general/press/?id=20050809

Probably there could be some slowness deriving from the emulation, but this should be a one-shot just when you just do not have a mac equivalent ready.

I think it sounds promising.

 

[VicMacs]

It was alive to begin with?

so it's NOT coming?

 

[digitalbiker]

Not Really

here dies the powerbook G5

Not really. Most people who think that a G5 PB may materialize in Paris, also believe that the first Intel based PB will use the Merom chip which won't be released until late 2006.

The January release of the Yonah processor has been well known for awhile now but I think many feel it would be better suited for the Mac Mini at MW 2006.

 

[~Shard~]

so it's NOT coming?

Sorry to kill the dream...

 

[Lacero]

here dies the powerbook G5

I wonder what's gonna happen to this guy?

 

[~Shard~]

I wonder what's gonna happen to this guy?

Ah yes, PowerMac G5 - I haven't seen him around much anymore, and he hasn't posted since June I believe. Perhaps all this Intel news has made him distraught and he is contemplating a name change. Wherever he is, I wish him well.

 

[Evangelion]

Yes but DMA is usually used for those types of things so IO flow usually goes from memory to the devices directly. In the case of the G5 the U3 (and U3H) acts as a crossbar switch with DMA programability (and DART remapping). So data flowing from memory to AGP doesn't involve the CPU buses at all, just the memory bus and AGP bus. This is the same for K8 except the memory controller lives on die with the CPU cores. So your point above isn't really a win.

I know that AGP's etc. mem-access doesn't use the FSB. But CPU's mem-access does. And not only does it eat in to the FSB, it has to go through a detour. In Pentium, G4, G5 and the like, the CPU's access to the RAM first goes to the northbridge, then it goes to the RAM. Then it returns to the northbridge, and it then goes back to the CPU. That is slow. In K8 it goes straight to the RAM, eliminating that detour. So there is an advantage, a clear one

The K8 on the other hand has a limited transistor budget because they also have to fit the CPU cores on the same die. So the implementation of the interconnect isn't as capable.

What makes you think that? The interconnect is HyperTransport on both. And many of the stuff traditional northbridges to, are still done by the northbridge in K8-systems. K8-northbridge is just simpler because it doesn't have to handle RAM-access, since that is handled by the integrated mem-controller on the CPU.

Finally the K8 is tied to a given memory bus technology and a limited range of data rates. You have to spin the K8 to change memory technology. (FB-DIMM may change that)

That's not really a problem. There are no real valid alternatives to DDR right now that would offer significantly better performance. DDR2 offers more bandwidth, true, but it has higher latencies. And DDR-RAM is still widely available, so AMD is not handicapped because they focus on DDR. And you CAN use faster RAM than AMD has specced on the K8. Many people overclock their RAM.

Also adding more K8 CPUs requires either fancier multi-ported RAM and related management or separate RAM banks, one per CPU with the use HT to share data from banks among CPUs.

Each CPU usually has a RAM-bank of it's own. So the bandwidht doubles with each CPU (altough in some cases there is only one RAM-bank), but latency increases a bit.

The K8 has low latencies and generally good bandwidth characteristics thanks to the location of the memory controller but as you see above drawbacks exist.

There are some drawbacks. But those drawbacks are mostly theoretical, and I would say that the positives clearly outweight the negatives. People were worried that they have to replace their CPU's if they want to use different RAM. True, but is that REALLY a problem? People don't seem to be lusting after DDR2, they are doing fine with regular DDR.

 

[hjhhjh]

why was conroe running on linux , while all the others were using a form of windows. does this mean it wont work on windows? possibly windows will use the pentium line continued , while apple use conroe , what u think?

 

[rickag]

why was conroe running on linux , while all the others were using a form of windows. does this mean it wont work on windows? possibly windows will use the pentium line continued , while apple use conroe , what u think?

No.

 

[Err]

I for one welcome our new Intel overlord

now maybe ported apps wont be such a joke, with macs userbase growing rapidly maybe some forword thinking developers will drop direct x and support openGL making porting apps to mac simple, things like Doom3 benchmarks on dual G5's hurt, thats our pride and joy, and mid level win boxs eat it up in gaming. I think the switch to intel coupled with a slew of bad hype around windows vista could turn the tides....... IMHO

disclamer - written under the influence of redbull, which tastes..... bad...

 

[madmaxmedia]

Apple has this wonderful ability to leave things hanging on the tech vine slowly withering. They pluck it literally moments before it spoils and really hurts the bottem line.

Like the current PowerBooks...I own one and love it, but when they finally come out with a upgraded CPU PowerBook they will definitely make hay for a few quarters. It feels like 3 out of every 10 MR forum posters are still using their TiBooks waiting for a G5 (now Intel) PowerBook.

 

[lightsout]

64-bit computing has nothing to do with performance. And it never did.

For PPC, SPARC, etc. Sure, it doesn't as the processors behave the same way.

For x86, it isn't the case. More registers, you can assume the processor has SSE2 (MMX, SSE1, etc) at least rather than making it i386 compatabile, the memory space is flat, etc.

I've got an Opteron box at work which speeds-up 25% when running the x86-64 binary over the i386 binary.

 

[heisetax]

Jobs the type to do this though

If it is a software issue, it'll be about the same. Considering my iBook is ready to use but the time I've moved my hand back down from opening the lid, there isn't going to be any benefit to faster speeds. Certainly you won't notice it, ignore the sheeple 'omg jobs is god, now Intel is god and G5 is teh suxs' people*.

If it is hardware, then pray that Intel have sorted their stuff out, because I've had no end of crap from x86 based notebooks and sleep.


* I'm really unimpressed with the ability to think of some of the people who have posted in this thread. Suddenly x86 is god and G5 sucks, etc etc. That's a load of crap. The Apple system controller for the G5 sucks, that's about it. Altivec will still beat out whatever floating point capabilities these new processors have. I think Jobs has been suckered by Intel personally, and will get a solution that is roughly the same, power-wise, as staying with PowerPC. Some applications will benefit, others will lose. Until Merom is out, a 7448 is a better choice than a Pentium-M in terms of power consumption, and a G5 is reasonable, and would be great with a better system controller. I'd like to see how a 65nm dual-core 8641D compares against Yonah or Merom as well, but I guess we will never find out. I don't think it will be a step back necessarily, I just don't think it is as amazing a move as people here think it is. It isn't as if Intel is a master of delivering on time either - where're the 5GHz+ Pentium 4s? I hope Jobs isn't doing this out of some personal vendetta against IBM...

Jobs can act like a baby with companies that do not do exactly what he wants. Remember the lact of ATI video cards from Macs for some time after they talked about a possible new Mac just hours before Steve Jobs did. IBM made him look bad when they could not ship a 3 GHz G5 when Steve had promised.

Just remeber that Steve Jobs is a salsman, not a computer engineer. Because he is such a good salesman piople be him when they really shouldn't. It will take a few years before I will be able to begin to believe in an Intel Mac. Noting said yet means anything as it will be a couple of years before there will be an Intel Mac that I would think about purchasing. Then it would be another year or two bfore there was a track history to know whether the Intel Macs will work as promised.

As far as hardware goes.Apple will loose some sales as people wait to purchase an Intel Mac as they don't want to have an architecture that Apple has quit using. Those that don't trust Steve & want to insure that their future Mac experience is as nice as it has been will purchase the current Macs. But these people will not replace these Macs for a long time.

A lot of what goes on because so many Macs supporters believe that Steve Jobs is a god. Thus they willbelieve that the formaerly best PPC is now inferior to the formerly worst Intel processor.

I wonder how long, if ever before we see a 3 GHz Mac. Remeber that the Pentium M comes in 1.6 & 2.1 Hz ranges. They are advertising the new models to not being as fast with clock speeds as before, but doing more work than the faster clock speed models do. This seems like what we have been told of the PPC in the past. Steve will keep the same ads as before, only he will change Intel rather than IBM & x8l rather than PPC. We'll be told that he was promised 3 GHz, but because of greater than expected problems with going to the 65nm chipe size, clock speeds are actually slower than before, thus we will have to wait until 2010 for a 3 GHz model.

It will be the same story, just different names to learn.

Bill the TaxMan

 

[deejemon]

*

 

[JFreak]

Many people overclock their RAM.

yes. and even more people don't know how much heat those overclocked ram modules generate...

 

[chaos86]

yes. and even more people don't know how much heat those overclocked ram modules generate...

and even more still of those people dont realize how much jeopardy they are putting their hundreds or even thousands of dollars worth of equipment in order to get an extra 5 FPS in WoW.

newsflash: overclocking gives you very little in return for the time, effort, and potentially dollars in replacement parts.

 

[madmaxmedia]

I think Jobs has been suckered by Intel personally, and will get a solution that is roughly the same, power-wise, as staying with PowerPC.

Well, you'd think that Jobs and the best Apple engineers did do their homework on this, considering the ramifications. How likely is it that MR posters have more foresight into Intel's and IBM's upcoming CPU's than Apple?

That being said, no one knows exactly what will happen for sure (including Apple.)

On the desktop side, I cam imagine Intel and IBM being actually relatively close in performance. But where would Apple go with their PowerBook? 3 more 167MHz speed bumps in the next 2 years?

 

[VanNess]

Not necessarily. New buyers may be attracted to Intel Macs if they're capable of running Windows (and Linux) in addition to OS X and competitively priced compared with the alternatives. That could end up being part of Apple's future marketing strategy for selling their computers. Why buy a Dell that only runs Windows (and Linux) when you can buy a Mac that runs those OSs and OS X? If future Intel Macs do indeed have that kind of multi-OS support it's one reason I doubt Apple will license OS X for Intel to other vendors, at least not immediately. If hardware sales turn unprofitable then they'd have more reason to consider licensing OS X but otherwise it doesn't make much sense (to me) for them to do it and risk killing their own profitable computer business.

Sorry for any redundant speculation. I haven't kept up with every post in this thread.

I think it would be more accurate to speculate that some buyers may be attracted to Intel Macs if they're capable of running Windows (and...Linux?) in addition to Mac OS X.

If I had to guess how many buyers, I would say not very many. It may have obvious geek appeal, but not much real world interest beyond that. There is at least empirical evidence (lol) that the majority of average, bread and butter OS X users choose OS X because they want Windows off their machine (and thier lives), not to have Windows back on it again - like some nagging in-law that just won't go away.

And I think Apple would prefer to address Windows compatibility by designing OS X to be easily interoperable with existing Windows systems as opposed to having Apple users install a competing OS (and given the present state of Windows, the security risks and extra OS management that come with it) on an Apple box along side it's flagship OS. And who is going to offer support for Windows on a Mac? Apple? Not a chance. If you go out and buy a Dell box, you can at least get support for Windows by calling Dell. A Mac user with Windows installed on his Intellimac is going to be an orphan; no one to turn to for support should something go wrong with Windows - which is all but guaranteed these days if it's hooked up to the net. Not a good state of affairs for the average user.

Putting Windows (or Linux) on an Apple Intellimac just doesn't jive with Apple's "it just works" mission for the Mac platform, which is a major reason why folks use OS X. And they are Windows-free at last.