Apple won't be exclusive Intel , until the end of 2007 when the transition should be complete. They might get them a bit cheaper , of course, but not yet for being Intel inside across the line.
I am sure that it doesn't really matter that Apple isn't all Intel they will still get the discounts. The apple move to Intel is worth at least that discount to Intel in sheer PR. The discount, if not a deeper one, is probably part of the contract.
That to me says we will see a Powerbook G5 with the Intel 'books following on at 1.6 (iBook) and 2.1 (Powerbook) in 12 months or less.
We won't see a G5 PB because a 1.6GHz G5 is probably quite a bit faster than a 2.1GHz Yonah. That's speculation, but it will be a real feat if the Yonah can touch the G5 as far as vector and FPU, which get heavy use in the only tasks that really tax modern processors. Integer performance is wonderful in games and normal stuff, but the only thing that is really a damper for most professionals buying top-end pb's these days is render speed. Be it 3d or 2d, all good rendering apps use altivec well and are FPU intense. I am not excited about the Yonah. When Intel chips have better FPU's and VPU's, I'll be excited.
Whether Yonah is 9 months away or 19, Apple is not going to design a Powerbook G5 to be built for only one year. Emotionally-driven or not, they are smarter than that.
And as for the contention that Apple will not use Intel chipsets: Is Apple that keen on reinventing the wheel? They should custom-design and custom-build parts that are no better than what they can get off the shelf?
Ok, have at this guy's proclomation of certainty Blackcat.
The only sensless assertion with regards to this matter would be to proclaim that it has to be one way or the other.
While it would seem to make economic sense for apple to use an off the shelf chipset, there are other factors at play.
Apple will likely design slight incompatabilities into the platform so that their OS can't be run on generic hardware. Most of apple's revenue has always been from selling hardware. They could change the company's focus, but have shown no inclination toward this. Nobody has any idea of if or how apple will try to maintain their roll as sole provider of hardware for the mac os.
We won't see a G5 PB because a 1.6GHz G5 is probably quite a bit faster than a 2.1GHz Yonah. That's speculation, but it will be a real feat if the Yonah can touch the G5 as far as vector and FPU, which get heavy use in the only tasks that really tax modern processors. Integer performance is wonderful in games and normal stuff, but the only thing that is really a damper for most professionals buying top-end pb's these days is render speed. Be it 3d or 2d, all good rendering apps use altivec well and are FPU intense. I am not excited about the Yonah. When Intel chips have better FPU's and VPU's, I'll be excited.
Ugh. Why must this FPU nonsense continue to be perpetuated?
The Yonah is essentially a dual-core Dothan on 65 nm process. We obviously don't know what the Yonah speeds are since there aren't any public tests; however, we can look up the available single processor Dothan.
Dothan @ 2.0 GHz: 1261 Spec_Int; 1285 Spec_FPU
G5 @ 2.5 GHz: 1082 Spec_Int; 1361 Spec_FPU
Slightly less than a 6% FPU improvement on a G5 with 25% more clockspeed. Now, if we could have a dual processor G5 powerbook, then there might be a valid comparison.
As for vector units, there are advantages and disadvantages between Altivec and SSE3; however, Altivec is no longer vastly superior to Intel's vector units (compared to crappy MMX and early SSE, it was far, far ahead); however, Altivec hasn't evolved and Intel has caught up. SSE3 at least allows double precision unlike Altivec. Secondly, only the wonderful IBM FSB design of the G5 is capable of keeping the Altivec fed. G4's anemic bus can't even feed data to Altivec fast enough to take great advantage of it.
Yonah is an all new chip. it will handily outperform the mobile G5. Bank on that.
Tonah is desinged around the Pentium M core, with changes.
The floating point performance is a main target point of Yonah. The Pentium M already whips the G5 at integer and next it will beat in in floating point.
The Pentium M already keeps pace with the G5 in applications from Adobe and Macromedia.
The Yonah will only serve to leap ahead.
After Yonah, there are a few others in the works. As for the Powerbooks, they may not even get Yonah. They may have Merom chips. Yonah will probalby be for iMacs, iBooks and Minis.
I for one am very excited about finally having some serious CPU options for my Powerbook.
The mobile G5 is a joke. Anyone who says that the low powre G5 is a Penium M killer is seriously out of touch. A G5 at 2.5 GHZ is, but anywhere below the 2 GHZ mark, the G5 takes a back seat. The main thing the G5 has goign for it right now is that it clocks higher than the M. However, clock for clock, the M is superior on average.
There's NO WAY that a dual-core 1.66ghz Pentium M will be slower than a 1.6ghz G5. For one, a 1.6ghz G5 needs a 800mhz FSB, which even Intel doesn't try to produce in a notebook. It's simply too power demanding and likely can't be cooled properly. If you need any proof that fast buses are hard to come by in notebooks, look at the 533FSB "Sonoma" chipset for Pentium M; it was delayed for about 6 months because of issues, and Intel has a lot of R&D money. So assuming that a 533mhz bus is all IBM could manage, it would certainly leave that G5 starved for bandwidth.
Besides that issue, also consider that the Pentium M is a very strong CPU in most tasks. Its only real weakness is the way it handles multithreading. By having a second core attached through 4MB of shared cache, this problem will disappear. When AMD added the second core to the Athlon64, it turned into an encoding monster. Yonah will likely have very similar success, IMO.
Apple won't be exclusive Intel , until the end of 2007 when the transition should be complete. They might get them a bit cheaper , of course, but not yet for being Intel inside across the line.
I think "exclusively Intel" means exclusively Intel for x86 products. HP sells Alpha stuff and they still had the Intel deal until they went AMD for some products.
Well, for ideal performance, 800FSB is preferred. Even 600mhz is a pretty fast bus for a T&L notebook, and it may not be possible for IBM to make one. Yes, I was aware that the iMac runs the G5 on a reduced bus, but it just proves the point that the G5 must be crippled to go into notebook. Pentium M was DESIGNED for notebooks. Sure, Apple could get a G5 into a Powerbook, but it won't perform like it was in a Powermac. At the least, it won't be able to keep up with a similar clocked Pentium M.
I've got a question. Is Yonah really a laptop chip? I know the Pentium M, which is the lineage of Yonah, was designed for laptops (Centrino). But this is dual-core, and I wonder if that makes sense for a laptop, where the emphasis is usually on low power consumption and small size. Are we really ready for a dual core chip in a laptop? Do we need one, or are single cores still enough? I'd think Apple would use a single-core version of it in their laptops, rather than dual core.
I've got a question. Is Yonah really a laptop chip? I know the Pentium M, which is the lineage of Yonah, was designed for laptops (Centrino). But this is dual-core, and I wonder if that makes sense for a laptop, where the emphasis is usually on low power consumption and small size. Are we really ready for a dual core chip in a laptop? Do we need one, or are single cores still enough? I'd think Apple would use a single-core version of it in their laptops, rather than dual core.
You must be fixating on the people you disagree with and then projecting that disagreement into your interpretation of other posts. Rest assured, relatively few people think that apple will use stock or reference design motherboards.
Well I'm one that does believe that Apple will use Intel Chip sets and will borrow as much as they can on board design. A big part of the reason to go with Intel and not AMD is the "other" chips. There is much more to this than just a new cpu, this is all of the things that we don't know that we cannot live without but Apple has not shown us how really cool that tech is, Intel is the chips and Apple is the software and wrapper.
I've got a question. Is Yonah really a laptop chip? I know the Pentium M, which is the lineage of Yonah, was designed for laptops (Centrino). But this is dual-core, and I wonder if that makes sense for a laptop, where the emphasis is usually on low power consumption and small size.
Yes, Yonah is a laptop chip. Yonah's power consumption will be the same as the current Dothan Pentium-M and the prior Banias Pentium-M chips.
It's not a miracle or anything. It's the fact that Yonah will be fabbed at 65 nm while Dothan is fabbed at 90 nm and Banias was fabbed at 130 nm. Approximately, Dothan has twice as many transistors Banias, and Yonah has twice as many transistors as Dothan, basic Moore's "Law" trends for successive fab generations.
Each new fab generation shrinks transistors to half the size, subsequently doubling the amount of transistors possible per "affordable" die size. Each fab generation also reduces the capacitance of each transistor and reduces the voltage required to power the transistor, both reducing the power consumption to make up for additional power consumed by 2x as many transistors.
Intel also has done something that IBM and Moto/Freescale have refused to do, design and implement engineering solutions to reduces power consumption, especially at 90nm:
2 GHz Dothan P-M
144 million transistors
~25 Watts TDP*
2 GHz Freescale 7448
~90 million transistors
~35 Watts max**
2 GHz PPC 970fx
~52 million transistors
~45 Watts max**
* TDP is Intel's power consumption number for computer makers to design to.
** estimates.
Just look at those numbers. It's quite amazing what Intel has done and what IBM and Freescale have not done.
Quote:
Are we really ready for a dual core chip in a laptop? Do we need one, or are single cores still enough? I'd think Apple would use a single-core version of it in their laptops, rather than dual core.
Apple will use dual-core Yonah chips in their Mac/Intel laptops. Not only that, you can look forward to quad-core laptop chips in 3 years when the 45 nm fab node goes into production.
Going into the future, the 45 nm node is the only thing that is important for Apple to consider from a chip supplier. All this talk about Athlon, Opteron and Cell are beside the point. No one talks about Moore's 2nd "law" which he observed that the development costs for fab development doubles every successive generation.
It is quite possible that only to 2 corporations will have a 45 nm fab. One will be Intel, who will get there first. The other will be a multi-conglomerate of IBM, AMD, and East Asian fabs.
will apple get it's own custom chips that others won't have, will apple with intel be able to maintain a competive edge compared with say dell, ibm sony??? if not apple is selling it's os and external designs. what then will separate apple from the others, or will apple get stuff before others??? will then apple be able to push intels envelope? will apple remain an innovator and in what areas?
Well, most of Dothan's transistor count is taken up by the L2 cache, which the cpu can power down when not in use (which helps a bunch in TDP and power management). The remaining cpu components don't take up that many transistors.
Comments
Originally posted by fieldor
Apple won't be exclusive Intel , until the end of 2007 when the transition should be complete. They might get them a bit cheaper , of course, but not yet for being Intel inside across the line.
I am sure that it doesn't really matter that Apple isn't all Intel they will still get the discounts. The apple move to Intel is worth at least that discount to Intel in sheer PR. The discount, if not a deeper one, is probably part of the contract.
Originally posted by Blackcat
That to me says we will see a Powerbook G5 with the Intel 'books following on at 1.6 (iBook) and 2.1 (Powerbook) in 12 months or less.
We won't see a G5 PB because a 1.6GHz G5 is probably quite a bit faster than a 2.1GHz Yonah. That's speculation, but it will be a real feat if the Yonah can touch the G5 as far as vector and FPU, which get heavy use in the only tasks that really tax modern processors. Integer performance is wonderful in games and normal stuff, but the only thing that is really a damper for most professionals buying top-end pb's these days is render speed. Be it 3d or 2d, all good rendering apps use altivec well and are FPU intense. I am not excited about the Yonah. When Intel chips have better FPU's and VPU's, I'll be excited.
And as for the contention that Apple will not use Intel chipsets: Is Apple that keen on reinventing the wheel? They should custom-design and custom-build parts that are no better than what they can get off the shelf?
Have some sense, people!
The only sensless assertion with regards to this matter would be to proclaim that it has to be one way or the other.
While it would seem to make economic sense for apple to use an off the shelf chipset, there are other factors at play.
Apple will likely design slight incompatabilities into the platform so that their OS can't be run on generic hardware. Most of apple's revenue has always been from selling hardware. They could change the company's focus, but have shown no inclination toward this. Nobody has any idea of if or how apple will try to maintain their roll as sole provider of hardware for the mac os.
Originally posted by cubist
Have some sense, people!
You don't hang around here much, do you?
Originally posted by Splinemodel
We won't see a G5 PB because a 1.6GHz G5 is probably quite a bit faster than a 2.1GHz Yonah. That's speculation, but it will be a real feat if the Yonah can touch the G5 as far as vector and FPU, which get heavy use in the only tasks that really tax modern processors. Integer performance is wonderful in games and normal stuff, but the only thing that is really a damper for most professionals buying top-end pb's these days is render speed. Be it 3d or 2d, all good rendering apps use altivec well and are FPU intense. I am not excited about the Yonah. When Intel chips have better FPU's and VPU's, I'll be excited.
Ugh. Why must this FPU nonsense continue to be perpetuated?
The Yonah is essentially a dual-core Dothan on 65 nm process. We obviously don't know what the Yonah speeds are since there aren't any public tests; however, we can look up the available single processor Dothan.
Dothan @ 2.0 GHz: 1261 Spec_Int; 1285 Spec_FPU
G5 @ 2.5 GHz: 1082 Spec_Int; 1361 Spec_FPU
Slightly less than a 6% FPU improvement on a G5 with 25% more clockspeed. Now, if we could have a dual processor G5 powerbook, then there might be a valid comparison.
As for vector units, there are advantages and disadvantages between Altivec and SSE3; however, Altivec is no longer vastly superior to Intel's vector units (compared to crappy MMX and early SSE, it was far, far ahead); however, Altivec hasn't evolved and Intel has caught up. SSE3 at least allows double precision unlike Altivec. Secondly, only the wonderful IBM FSB design of the G5 is capable of keeping the Altivec fed. G4's anemic bus can't even feed data to Altivec fast enough to take great advantage of it.
Tonah is desinged around the Pentium M core, with changes.
The floating point performance is a main target point of Yonah. The Pentium M already whips the G5 at integer and next it will beat in in floating point.
The Pentium M already keeps pace with the G5 in applications from Adobe and Macromedia.
The Yonah will only serve to leap ahead.
After Yonah, there are a few others in the works. As for the Powerbooks, they may not even get Yonah. They may have Merom chips. Yonah will probalby be for iMacs, iBooks and Minis.
I for one am very excited about finally having some serious CPU options for my Powerbook.
The mobile G5 is a joke. Anyone who says that the low powre G5 is a Penium M killer is seriously out of touch. A G5 at 2.5 GHZ is, but anywhere below the 2 GHZ mark, the G5 takes a back seat. The main thing the G5 has goign for it right now is that it clocks higher than the M. However, clock for clock, the M is superior on average.
Thanks for saving me some legwork and typing!
There's NO WAY that a dual-core 1.66ghz Pentium M will be slower than a 1.6ghz G5. For one, a 1.6ghz G5 needs a 800mhz FSB, which even Intel doesn't try to produce in a notebook. It's simply too power demanding and likely can't be cooled properly. If you need any proof that fast buses are hard to come by in notebooks, look at the 533FSB "Sonoma" chipset for Pentium M; it was delayed for about 6 months because of issues, and Intel has a lot of R&D money. So assuming that a 533mhz bus is all IBM could manage, it would certainly leave that G5 starved for bandwidth.
Besides that issue, also consider that the Pentium M is a very strong CPU in most tasks. Its only real weakness is the way it handles multithreading. By having a second core attached through 4MB of shared cache, this problem will disappear. When AMD added the second core to the Athlon64, it turned into an encoding monster. Yonah will likely have very similar success, IMO.
Originally posted by fieldor
Apple won't be exclusive Intel , until the end of 2007 when the transition should be complete. They might get them a bit cheaper , of course, but not yet for being Intel inside across the line.
I think "exclusively Intel" means exclusively Intel for x86 products. HP sells Alpha stuff and they still had the Intel deal until they went AMD for some products.
Originally posted by Wingnut
For one, a 1.6ghz G5 needs a 800mhz FSB
Funny, a 1.8ghz G5 in an iMac works with a 600mhz frontside bus...
Originally posted by 9secondko
Yonah is an all new chip. it will handily outperform the mobile G5. Bank on that.
Tonah is desinged around the Pentium M core, with changes.
My guess is that the English translation of the Hebrew "Yonah" allows for the "Tonah" spelling, but not knowing Hebrew, it is a complete guess.
According to this:
Wiki
Yonah is derived from the Pentium-M (Dothan) core.
Originally posted by BRussell
I've got a question. Is Yonah really a laptop chip? I know the Pentium M, which is the lineage of Yonah, was designed for laptops (Centrino). But this is dual-core, and I wonder if that makes sense for a laptop, where the emphasis is usually on low power consumption and small size. Are we really ready for a dual core chip in a laptop? Do we need one, or are single cores still enough? I'd think Apple would use a single-core version of it in their laptops, rather than dual core.
Yes.
http://www.intel.com/cd/ids/develope...969.htm?page=6
Originally posted by dfiler
Everybody?
You must be fixating on the people you disagree with and then projecting that disagreement into your interpretation of other posts. Rest assured, relatively few people think that apple will use stock or reference design motherboards.
Well I'm one that does believe that Apple will use Intel Chip sets and will borrow as much as they can on board design. A big part of the reason to go with Intel and not AMD is the "other" chips. There is much more to this than just a new cpu, this is all of the things that we don't know that we cannot live without but Apple has not shown us how really cool that tech is, Intel is the chips and Apple is the software and wrapper.
Originally posted by BRussell
I've got a question. Is Yonah really a laptop chip? I know the Pentium M, which is the lineage of Yonah, was designed for laptops (Centrino). But this is dual-core, and I wonder if that makes sense for a laptop, where the emphasis is usually on low power consumption and small size.
Yes, Yonah is a laptop chip. Yonah's power consumption will be the same as the current Dothan Pentium-M and the prior Banias Pentium-M chips.
It's not a miracle or anything. It's the fact that Yonah will be fabbed at 65 nm while Dothan is fabbed at 90 nm and Banias was fabbed at 130 nm. Approximately, Dothan has twice as many transistors Banias, and Yonah has twice as many transistors as Dothan, basic Moore's "Law" trends for successive fab generations.
Each new fab generation shrinks transistors to half the size, subsequently doubling the amount of transistors possible per "affordable" die size. Each fab generation also reduces the capacitance of each transistor and reduces the voltage required to power the transistor, both reducing the power consumption to make up for additional power consumed by 2x as many transistors.
Intel also has done something that IBM and Moto/Freescale have refused to do, design and implement engineering solutions to reduces power consumption, especially at 90nm:
2 GHz Dothan P-M
144 million transistors
~25 Watts TDP*
2 GHz Freescale 7448
~90 million transistors
~35 Watts max**
2 GHz PPC 970fx
~52 million transistors
~45 Watts max**
* TDP is Intel's power consumption number for computer makers to design to.
** estimates.
Just look at those numbers. It's quite amazing what Intel has done and what IBM and Freescale have not done.
Are we really ready for a dual core chip in a laptop? Do we need one, or are single cores still enough? I'd think Apple would use a single-core version of it in their laptops, rather than dual core.
Apple will use dual-core Yonah chips in their Mac/Intel laptops. Not only that, you can look forward to quad-core laptop chips in 3 years when the 45 nm fab node goes into production.
Going into the future, the 45 nm node is the only thing that is important for Apple to consider from a chip supplier. All this talk about Athlon, Opteron and Cell are beside the point. No one talks about Moore's 2nd "law" which he observed that the development costs for fab development doubles every successive generation.
It is quite possible that only to 2 corporations will have a 45 nm fab. One will be Intel, who will get there first. The other will be a multi-conglomerate of IBM, AMD, and East Asian fabs.