Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Another point that I see people missing:

Processors use more power per quantum of performance as they are pushed to their peak performance.

Put another way, compare two similar processors - one running at 100% and one running at 90%. The one running at 100% performance is using a lot more than 10% extra energy over the one running at 90%.

Heat output etc goes off the charts. Input voltage goes much higher as the voltage swings in the CPU cycle are pushed to extremes. See the desktop overclocking / laptop underclocking scenes for confirmation.

Thus, a dual core processor running both cores at 60% could well use less battery power to do the same job as a single core processor running at 100%.

Likewise (and I'm speculating here) a quad core laptop processor running all four cores at 60% might well use less battery power (and output less heat) than a dual core processor running both cores at 100%.

So it's not about shutting off unused cores, it's about running what you have at as low speeds as possible.

Thus, if I'm right, quad (and more) cores are the only existing way forwards for laptop makers to get more power with less battery consumption.
 
RedTomato, I completely agree. When it comes to the car analogy, the cylinder can be either on or off, but whatever is the most efficient for CPUs can be done. In any case, power consumption should not vary too much if someone is surfing on a quad or dual core, if the power management and cpu architechure doesn't specifically counter this...
 
I can't see them putting it in a MacBook for at least 5 years (considering there is no real need to it for the average MacBook user).

5 years ? :rolleyes: Just in time for an ACD update. :D Good grief. Talk about lowered expectations. Apple should spin off the computing division. :D
 
Another point that I see people missing:

Processors use more power per quantum of performance as they are pushed to their peak performance.

Put another way, compare two similar processors - one running at 100% and one running at 90%. The one running at 100% performance is using a lot more than 10% extra energy over the one running at 90%.

Heat output etc goes off the charts. Input voltage goes much higher as the voltage swings in the CPU cycle are pushed to extremes. See the desktop overclocking / laptop underclocking scenes for confirmation.

Thus, a dual core processor running both cores at 60% could well use less battery power to do the same job as a single core processor running at 100%.

Likewise (and I'm speculating here) a quad core laptop processor running all four cores at 60% might well use less battery power (and output less heat) than a dual core processor running both cores at 100%.

So it's not about shutting off unused cores, it's about running what you have at as low speeds as possible.

Thus, if I'm right, quad (and more) cores are the only existing way forwards for laptop makers to get more power with less battery consumption.

RedTomato, I completely agree. When it comes to the car analogy, the cylinder can be either on or off, but whatever is the most efficient for CPUs can be done. In any case, power consumption should not vary too much if someone is surfing on a quad or dual core, if the power management and cpu architechure doesn't specifically counter this...

I see you power point. But you won't get four cores at 60% vs 2 at 100% if the OS and applications aren't written to utilize all four cores. And most Apps currently are not. I'm not sure with multiple apps running if it will run some apps off two cores and others off the other two. Somehow I doubt that. A quad core isn't exactly like having 2 dual cores.
 
I'm not sure with multiple apps running if it will run some apps off two cores and others off the other two. Somehow I doubt that.

Yes it will. Splitting several apps and the OS between different cores is trivially simple. Each app is an independent program in its own right, and isn't dependent on the outcome of other apps.

Don't confuse this with programming a single app to run on several cores, which can be very hard work to do properly. (as the outcome of one function could depend on the outcome of another function which could be running on another core).

Remember we're not looking for raw single-app speed here, just the ability to run several apps (itunes, quicktime, firefox, mail, msn, time machine, the os etc) at the same time and get the job done using as little power as possible.
 
Please don't drag up the "average user" argument. It turns quickly into a pointless hardware specification war.

Apple is very fond of using the fastest processors Intel has to offer while skimping on the remaining hardware.

It all depends on what the competition does and what Apple does in the future. For all we know, the dual core might end up being used in a 'value range'.
 
It all depends on what the competition does and what Apple does in the future. For all we know, the dual core might end up being used in a 'value range'.

It already is. Almost all of Intel's PC processors are now dual core, even the Celerons. Soon, you won't be able to buy a single core laptop / desktop processor.

It's not been possible to buy a single core Apple computer for quite some time now. I'm not including the Apple TV as that's a single use appliance not a general purpose computer. (last I checked, it ran a Pentium M, may have moved onto a core dual, if not now, then will do soon)
 
Problemo is, what software is available to take advantage of all those cores:rolleyes:

There already IS benefit without specially written apps. It's 'somewhat like' having 4 machines to run different apps on. As long as the OS knows how to evenly distribute the app processes to different cores, you will see benefit. See ReTomato's explanation:

Yes it will. Splitting several apps and the OS between different cores is trivially simple. Each app is an independent program in its own right, and isn't dependent on the outcome of other apps.

Don't confuse this with programming a single app to run on several cores, which can be very hard work to do properly. (as the outcome of one function could depend on the outcome of another function which could be running on another core).

Remember we're not looking for raw single-app speed here, just the ability to run several apps (itunes, quicktime, firefox, mail, msn, time machine, the os etc) at the same time and get the job done using as little power as possible.
 
Please don't drag up the "average user" argument. It turns quickly into a pointless hardware specification war.

Apple is very fond of using the fastest processors Intel has to offer while skimping on the remaining hardware.

Completely untrue.
 
I see you power point. But you won't get four cores at 60% vs 2 at 100% if the OS and applications aren't written to utilize all four cores. And most Apps currently are not. I'm not sure with multiple apps running if it will run some apps off two cores and others off the other two. Somehow I doubt that. A quad core isn't exactly like having 2 dual cores.

You are right about one thing... a quad core isn't exactly like having 2 dual cores. Since they share a big cache, rather than splitting it, you get some performance gains in certain situations. You do have to deal with 4 cores sharing the same single pipe off the CPU, which may offset some of those gains though.

But overall, as RedTomato has also pointed out, OSes are pretty good about load-balancing across multiple cores and processors. OS X, Linux, and Windows all do a good enough job. Now, that means any one process is stuck taking no more than 1 core if it isn't multithreaded, but hey, that isn't a bad thing in a multitasking world where you run a dozen apps at once.

Current Core 2 chips can throttle the chip effectively based on load. So to get really good battery life, you want to spread your CPU load across as many cores as possible. Pegging 1 core at 100% and leaving the other at 5% will cause the whole chip to draw nearly peak power. But 52% on each means your CPU can throttle down, or be in a low power state for roughly half of every second. Both of which save power.

I believe Nehalem lets the CPU throttle each core independently, which will improve this even further when there is an imbalance. In general though, performance /and/ battery life improvements from software will be by making apps take little slices of time from every core in a system, rather than all of one. It will take time for APIs and application design to finish catching up though.
 
Well, bad news and good news today:

According to Digitimes and Reghardware, the first wave of Montevina chips will only be available late June instead of early June.
The mobile quad QX9300 is still planned for Q3 at $1,038.
But it seems that there will be another mobile quad in Q4. The Q9100 will have a lower TDP (35W?) and given the part number, it should be a 2.26GHz part.
9300 = 2.53, 9200 = 2.40, 9100 = 2.26?
This could be an option of the 15/17" MBP and maybe the 20" iMac for XMas or MWSF 2009.
 
According to Digitimes and Reghardware, the first wave of Montevina chips will only be available late June instead of early June.

FAB28 is running behind schedule for opening, which leaves FAB32 as the only fully 45nm facility (not counting the development fab). On the plus side, FAB11X should be converted to 45nm production before the end of the year, so that should allow Intel to start really pumping out 45nm Penryns and Atoms in time for the holiday sales.
 
Completely untrue.

Why isn't it? I am using a 17" Penryn right now and from what I am reading it has the fastest mobile chip Intel offers (sans the 2.8 which needs much more cooling) and it has an under-clocked 512MB GFX card.

I kind of agree with Eidorian on this one. Even the MacBook has some light hardware when it comes to what Apple could put into it.
 
I can't see them putting it in a MacBook for at least 5 years (considering there is no real need to it for the average MacBook user).

Well if you look at all the processes that goes on at the same time with a normal mac setup having multible cores is useful Even with running non-multi-threaded applications.

Lets say you are using iTunes, Office 2008 with a rather complex spreadsheet, Photoshop a large image. Already 2 of those cores can be fully utilized at once a third one will at least be partially utalized. and the OS can still run on the last. Doing things like that on a single CPU is painful. Doing it on Duel CPU is workable having more cores will make doing more easier.
 
Now I'm 99% certain Apple will use P9500 & P8600 in MBP (2.53 & 2.4 Ghz)
and P8600 & P8400 in MB (2.4 & 2.26 Ghz).

And as speeds remain virtually the same, big story will be about "real world" speed gains with faster FSB/RAM.
But even bigger will be power consumption decreases!

As to a BTO (2.6Ghz) option, I think they'll either quietly discontinue it (as 2.8Ghz T9600 is becoming too hot : 35W vs. 25W), or relegate only to a 17". Or there is always :rolleyes: option of a "special" processor, crafted exclusively for Apple ;):D:eek:


So ... I'm not saying that you quad-prophesying people won't ever see your day. But not before next spring I feel.
And I'm pretty happy this way as this clears the path for MB Air Pro.
Yeah no DVD/battery door for me.

Or if I couldn't hold my guns, rev.B MBAir with SL9400 (20W/1.86Ghz) will do. Anyone heard of Apple Edition SL9500(20W/2.26Ghz) ?! :p:D
 
I wonder if Apple will want to put the P9500 in the MacBook or limit it to the MacBook Pro. In addition to the slight clock-speed boost over the P8600, the double-sized cache (6MB vs. 3MB) would help put the MacBook Pro on a higher performance level to help leverage the other features (discrete GPU, larger display, ExpressCard, FW800, Dual-DVI).

Then again, all of that does help explain the $650 difference in price, so it might not matter allowing the MacBook to have the P9500, as well.

I believe the T9600 will be a MacBook Pro option for both screen sizes since it has the same 35W TDP that the T7800 2.6GHz unit does in the current 15" and 17". If the MacBook Pro gets a new case design, we might see the 15" also get the X9100 and perhaps the QX9300. I expect the 17" will get both since the larger chassis size should be able to handle the 44 watt TDP.
 
Now I'm 99% certain Apple will use P9500 & P8600 in MBP (2.53 & 2.4 Ghz)
and P8600 & P8400 in MB (2.4 & 2.26 Ghz).

And as speeds remain virtually the same, big story will be about "real world" speed gains with faster FSB/RAM.
But even bigger will be power consumption decreases!

As to a BTO (2.6Ghz) option, I think they'll either quietly discontinue it (as 2.8Ghz T9600 is becoming too hot : 35W vs. 25W), or relegate only to a 17". Or there is always :rolleyes: option of a "special" processor, crafted exclusively for Apple ;):D:eek:


So ... I'm not saying that you quad-prophesying people won't ever see your day. But not before next spring I feel.
And I'm pretty happy this way as this clears the path for MB Air Pro.
Yeah no DVD/battery door for me.

Or if I couldn't hold my guns, rev.B MBAir with SL9400 (20W/1.86Ghz) will do. Anyone heard of Apple Edition SL9500(20W/2.26Ghz) ?! :p:D

:confused:

Do you know that current cpus used in the MB/MBP ARE 35W parts?

Nothing prevents Apple from using:
- 2.26/2.40/2.53GHz parts in the MB/Mac mini,
- 2.53/2.80GHz parts in the MBP, along with a custom 2.66GHz (like the current iMac)
- all of those + the 3.06GHz and future quads in the iMac.

Intel has also a Q9100 (quad 2.26GHz @ 35W) forecasted for Q4, I certainly expect Apple to use it (as an option) in a january 2009 refresh of the MBP.
 
American Technology Research analyst Doug Freedman asserts that Intel did not properly complete the certification for either the new mobile Penryns designed for use with the Montevina chipset or for the 802.11n part of the chipset and he claims means they cannot be sold in the US at this time.

It's a bit difficult to determine what he is talking about, since he mentions "CPUs", but then talks about a slower ramp-up of 802.11n-enabled devices. To my knowledge, Apple does not use Intel's wireless chipset so if it is an issue with that part of the chipset, then Apple could still use the Cantiga chipset with the new CPUs.
 
:confused:

Do you know that current cpus used in the MB/MBP ARE 35W parts?

.... big story will be ....
But even bigger will be power consumption decreases!

So, yes I know what current Penryns are, and what Intel planned for them to be.

Point was with new chips you can continue in the old ways - using T9400/9600 (2.53/2.8) parts with the same TDP & even the prices.

Or they can ... innovate, and produce MBAir Pro ...
No internal DVD/removable battery, hence extremely slim & sturdy (I grew to love it in my iPhone, and I like the feel of MBA body).
With 25W parts, with even better (marginally but still) performance, they finally can pull this machine out.


... so what would you call an innovation, a 5% faster of the same ... with a bit bigger HDD .....
or a completely new body, a new paradigm (as if we didn't see it coming :p) for a portable computing, and a new sacrifices for the spirit of portability/Mac'iness ;):D)
What would you choose if you was Apple/SJ ?


I hate SJ for not seeing gaming, but I'd laud him for not seeing "quad-computing" for laptops ....
And I think he will : no other Mac has "pro" features of a Mac Pro (bays/expandability; interfaces; etc)


Think of it this way : Apple still doesn't want to use chips like everybody else (and they don't: it's only my fantasy that I have T7700 in my MBP, in "reality" it's: "2.4GHz Intel Core 2 Duo processor").
But they hate being behind in Ghz race ....

So, this P series gives them this chance ..... not only to produce less heated laptop ... (and that's what everybody will be doing - shoveling new parts in old machines) .... but use it as the building block for a something new .... pushing their new visions at the same time .... Air ?!


P.S. did you notice that they've changed keyboard layout & added multitouch to MBP ... but "forgot" to enlarge trackpad ....
 
Bear in mind also, that P8600/8400, which is bound to MB eventually ... has exactly the same price as current T8300/8100, exact same 3MB cache, and only difference (8600vs8300) is power consumption (25W vs 35W)

With Intel's aggressive push for Centrino 2, we'll see rapidly diminishing shipments of gen.1 penryns.
That will push for 25W MB's by the back to school or bit later (if they decide to introduce new designs along with cooler CPUs)
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.