Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
IBM's 90nm PowerPC 970FX
- Thread starter MacRumors
- Start date
- Sort by reaction score
Chryx:
The Pentium-M *IS* based on the Pentium 3 core. (which was based on the Pentium 2, which was based on the Pentium Pro). The Intel Israeli design team actually took the P3 core and modified it for this design, and the pipeline on the Pentium-M is longer than the P3, allowing for frequency scaling, but shorter than the P4 which is an energy hungry CPU design. Banias, the codename for the Pentium-M CPU will be phased out within the next two quarters, and Dothan, the successor to Banias will appear. It will have 2MB of L2 cache, and scale up to fairly well in frequency.
-PPCTech
The Pentium-M *IS* based on the Pentium 3 core. (which was based on the Pentium 2, which was based on the Pentium Pro). The Intel Israeli design team actually took the P3 core and modified it for this design, and the pipeline on the Pentium-M is longer than the P3, allowing for frequency scaling, but shorter than the P4 which is an energy hungry CPU design. Banias, the codename for the Pentium-M CPU will be phased out within the next two quarters, and Dothan, the successor to Banias will appear. It will have 2MB of L2 cache, and scale up to fairly well in frequency.
-PPCTech
Originally posted by Snowy_River
Uh, atoms don't have heads...
And, that's a viola...
Isn't an atom a head?
We'd like to think so , but it is nothing more than a P-3 slightly enhanced and at a 130NM process. The P-3 is still a nice processor and whips the P4 on mhz to mhz basis. The P-M was Intel's response to the awesome reception the G4 notebooks received. You couldn't fit a P4-M into a 1" notebook, they are still predominately in the large bricks. The P3 now P-M had a much longer life for the notebook market.
All we need is just a little patience. I say new PB G5s at Macworld Tokyo in March. Makes good timing after IBMs processor forum in Feb that whets our appetites even moreso.
1) I'm familiar with Intel Codenames for these things
2) I've not seen any actual _evidence_ that Banias/Dothan are of Tualatin/Coppermine/Katmai/Deutchetes /Klamath etc etc lineage, conjecturally it seems plausible, but I remain skeptical.
I'll get in touch with an ex-Intel buddy and see if he can make some enquiries on this.
BTW, Dothan appears to be running hot thanks to Intels leaky 90nm process
You are correct sir, but take a look at how quickly the P-M is moving up in W and heat. Add 2MB to the L2 Cache and boom, voltage goes up, heat accelerates. A deeper pipeline means more energy and heat and an overall slow down. The P-M is more an admission that the P4 is already close to the end of its lifecycle, imho. I think Intel has to have something else up there sleeve. Unless, of course, it was the Itanium-2 which has proved to be a dismal failure for them. You would think such a large firm would have a few projects going as contingencies.
I would hate to see Intel caught with their pants down, I would like to see the pressure continue on IBM to race ahead of Intel. The worse thing that can happen to a company is that they become #1.
Yep, Intel is really struggling with their 90NM process. Can IBM be that far ahead? It really is something to ponder. Will Apple be able not only to leapfrog Intel based computers performance wise but also catch up and maybe even surpass Intel on the MHz marketing front? Intel's seems to be closing really fast. Very interesting change of events.
It would be like SJ to shock us all. These W and heat numbers suggest that Apple is ready to go to 3GHz really soon, much sooner than Summer.
Because that falls within the range of applications for which the PPC970 architecture is well suited, perhaps? Ignoring that "devastatingly quick" isn't saying much (perhaps you should look at some CFD benchmarks for various current systems some time).
There are a lot of people here pulling numbers out of /dev/ass and making claims about all the architectures that are unfounded in reality. I work on high-performance computing systems for a living. We have a bunch of both G5 systems and Opteron systems here, and we've benchmarked both on gobs of different types of supercomputing codes (our application space makes broad use of the capabilities of the architectures, but is all large memory codes).
I've said it before, and I'll say it again: The PPC970 is tops for DSP type codes, particularly if you are running tight, compute intensive loops. If you are running more general purpose or latency bound applications, the Opteron is the clear winner most of the time. Subscribe to one of the many mailing lists dedicated to commodity supercomputing on the 'net; you'll find lots of independent tests at various R&D labs that confirm this.
It reminds me of the G4 Photoshop "benchmark"; such a ridiculously narrow algorithm space that it proves nothing in the general case. See what happens when you smoke Apple marketing copy?
Amazing how many people still rely on benchmarking and marketing hype rather than the real world users and word of mouth.
Sort of like the people who go out and buy the high HP vehicle based on race track tests, then complain that their lightly used vehicle they bought is a gutless lemon in city driving.
Of couse then the mechanic takes the Mustang Cobra out for a high RPM blast throught the streets, talks about it on the net, and then gets fired.
All depends on whether your typical use is similar to the tests or not.
That's funny, it seems that you are better qualified to be a professional troll. If you need a reference for your new trolling position, I would be happy to write one for you.
Maybe you should write NASA and tell them about this huge mistake they are making:
http://www.macnn.com/news/23089
Originally posted by mkaake
me, i'm just happy
i figure, these are going to show up in some product, be it a cube, laptop, or desktop... which means refurbs, old machines and used machines suddenly get cheaper
call me cheap (or poor - that would be more accurate), but i love it when new products come out so i can get what everyone used to think was great at a good price...
matt
I'll call you "smart", because you think like I do!
Re: Re: What happened to the 750VX?
Actually I'm begining to think that Aple may just bypass the VX chip in favor of the 970 in all desk top systems. Lets face it al they need is a lower power system controller (SC) to stick this chip into just about all of the current models. A lower power SC should be easy to botain by eliminating SMP support and maybe reducing or chaning the attached I/O buses.
This would leave the iBook as the only place to stick the VX.
It would be a bit of a marketing trumph to have a entire line of 64 bit desktops. From the engineering side extending the address space is the primary advanatge of the 970 in the low end machines.
Dave
Actually I'm begining to think that Aple may just bypass the VX chip in favor of the 970 in all desk top systems. Lets face it al they need is a lower power system controller (SC) to stick this chip into just about all of the current models. A lower power SC should be easy to botain by eliminating SMP support and maybe reducing or chaning the attached I/O buses.
This would leave the iBook as the only place to stick the VX.
It would be a bit of a marketing trumph to have a entire line of 64 bit desktops. From the engineering side extending the address space is the primary advanatge of the 970 in the low end machines.
Dave
Re: Re: New PowerMacs, sure; iMacs, maybe; PowerBooks, no way.
In my Mr. Miyagi voice I would say
"The past is an excellent way of recording your mistakes so as not to repeat them, however it is lousy as an absolute guide for what will happen in the future."
In my less Zen-like personality, I would argue that a 1.6-1.8 GHz 'modular' Macintosh (with a 3:1/4:1 FSB), providing a common base for the new iMac - which I refer to as 3rd Generation iMac (or 3GiM, for short) - and a new 'Cube'-like business Mac is precisely what Apple needs to answer the critics who claim that a) the 2GiM is getting stale [true] and that b) Apple has no headless offering for the business community, who are up to their eyes in cheap flat-panels and CRTs but fed up of managing their way through Windows Update hell [also true].
Due to SPJ's Barnum-like personality, 2004 is going to see more new and revised product announcements than any other Macintosh year since the early 90s, and the whole IIvx/si/LC/Performa product matrix fiasco.
Unlike the 90s however, I suspect that each product will be carefully placed with laser-like accuracy and in a manner designed to disrupt the competition (a la iPod and iTMS), as well as the market and media perception of Apple as a company.
In my Mr. Miyagi voice I would say
"The past is an excellent way of recording your mistakes so as not to repeat them, however it is lousy as an absolute guide for what will happen in the future."
In my less Zen-like personality, I would argue that a 1.6-1.8 GHz 'modular' Macintosh (with a 3:1/4:1 FSB), providing a common base for the new iMac - which I refer to as 3rd Generation iMac (or 3GiM, for short) - and a new 'Cube'-like business Mac is precisely what Apple needs to answer the critics who claim that a) the 2GiM is getting stale [true] and that b) Apple has no headless offering for the business community, who are up to their eyes in cheap flat-panels and CRTs but fed up of managing their way through Windows Update hell [also true].
Due to SPJ's Barnum-like personality, 2004 is going to see more new and revised product announcements than any other Macintosh year since the early 90s, and the whole IIvx/si/LC/Performa product matrix fiasco.
Unlike the 90s however, I suspect that each product will be carefully placed with laser-like accuracy and in a manner designed to disrupt the competition (a la iPod and iTMS), as well as the market and media perception of Apple as a company.
Re: Re: Re: New PowerMacs, sure; iMacs, maybe; PowerBooks, no way.
Somebody ought to forward this message ot Steve J at Apple. There simple is no good excuse for the way that the imac line has been marketed, they really need to reconsider their attitude with respect to performance of these machines.
I've always like the concept of an all in one, they certainly are great for some applications but lets face it they need to perform. The G5 tower should be where Apple addreesses the professional crowds with SMP and expandability, and the dropped the ball on expandability there. I have to caution though that SMP is quickly becoming an expectation of many hardware purchasers so that will eventually work its way into the all in one line also.
What is really needed in the iMac line is two PC's with significant performance differrences between them. Give the user the option of a truely low cost iMac or something middle of the road. Thus maybe one imac with a 1.8GHz 970 and another with a 2.4GHz procesor. Finally they need to put real memory in these machines.
Thanks
Dave
Somebody ought to forward this message ot Steve J at Apple. There simple is no good excuse for the way that the imac line has been marketed, they really need to reconsider their attitude with respect to performance of these machines.
I've always like the concept of an all in one, they certainly are great for some applications but lets face it they need to perform. The G5 tower should be where Apple addreesses the professional crowds with SMP and expandability, and the dropped the ball on expandability there. I have to caution though that SMP is quickly becoming an expectation of many hardware purchasers so that will eventually work its way into the all in one line also.
What is really needed in the iMac line is two PC's with significant performance differrences between them. Give the user the option of a truely low cost iMac or something middle of the road. Thus maybe one imac with a 1.8GHz 970 and another with a 2.4GHz procesor. Finally they need to put real memory in these machines.
Thanks
Dave
I don't know what criteria Consumer Reports uses but from www.computers.com
I don't think I've seen a single Pentium M that gets under 3 hours of battery life even under harsh conditions like that. I know that the new IBM T40 series that we have just started rolling out here in the office gets at least 3+ hours while playing a movie.
IBM has now fixed that error, the actual results are:
970 1.8 5800 MIPS
970FX 2.0 7584 MIPS
That seems like a pretty big jump for only 200MHz increase, but I know nothing of Dhrystone, so maybe its not.
the new PDF updated today is:
http://www-306.ibm.com/chips/techli...0B287256BF3006FBE54/$file/PPC_QRG_1-22-04.pdf
Re: Re: Re: Re: New PowerMacs, sure; iMacs, maybe; PowerBooks, no way.
That and Apple lives in the high end of the computer market. They do not have a consumer tower.
That and Apple lives in the high end of the computer market. They do not have a consumer tower.
The forum is where you come to find out what is happening in the future, and it is very exciting to see how quickly IBM are moving ahead, no doubt motivated by the $3bn price tag hanging over their heads as they work away merrily in their factory.
Here in the closer present, Apple sold fewer PowerMac G5s than last quarter, and sold about 190,000 Powerbooks, which was up on the previous quarter. The current G4 has some legs in it still, and despite the duff PB specs in the eyes of many here, the PBs are doing very well in the notebook market place,. Whilst Im not saying a G5 Powerbook wont be on the horizon, the PowerMac is the main concern and must surely be the likely recipient of the latest greatest chips.
Here in the closer present, Apple sold fewer PowerMac G5s than last quarter, and sold about 190,000 Powerbooks, which was up on the previous quarter. The current G4 has some legs in it still, and despite the duff PB specs in the eyes of many here, the PBs are doing very well in the notebook market place,. Whilst Im not saying a G5 Powerbook wont be on the horizon, the PowerMac is the main concern and must surely be the likely recipient of the latest greatest chips.
Marketing 101
Speaking of heads up the anal orifice, would you like to enlighten the rest of us as to how Apple's tiered product line differs from any other company in existence? Dell doesn't sell just one product -- they offer a choice, depending on the buyer's budget and needs. This is basic marketing, kiddies: your product line should segregate the market into those who want to pay more, and those who want to pay less.
So here's the deal: If you want the best, and can pay for it, get a G5. If you don't, or can't, get an iMac or eMac. If you can't afford either of those, buy a Dell, or build your own. Apple (and SJ) have stated over and over and over that they're not trying to sell commoditized boxes. They simply aren't interested. So get over it.
Look, you can disagree with Apple's strategy all you want, but to accuse them of "playing games" or being idiots is nothing more than a childish cheap shot.
Speaking of heads up the anal orifice, would you like to enlighten the rest of us as to how Apple's tiered product line differs from any other company in existence? Dell doesn't sell just one product -- they offer a choice, depending on the buyer's budget and needs. This is basic marketing, kiddies: your product line should segregate the market into those who want to pay more, and those who want to pay less.
So here's the deal: If you want the best, and can pay for it, get a G5. If you don't, or can't, get an iMac or eMac. If you can't afford either of those, buy a Dell, or build your own. Apple (and SJ) have stated over and over and over that they're not trying to sell commoditized boxes. They simply aren't interested. So get over it.
Look, you can disagree with Apple's strategy all you want, but to accuse them of "playing games" or being idiots is nothing more than a childish cheap shot.