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.
Intel's Plans for Low-Power Chips
- Thread starter MacRumors
- Start date
- Sort by reaction score
lvnmacs said:Hi, I'm a new Mac user and forgive my naiveness. The best thing so far I love about my Powerbook is when I wake it up from its sleep, it responds really quick. Will it be the same with the Intel based processors?Thanks in Advance
even faster.
I finally read all these messages.
A few considerations:
If you want to compare processors and their architectures, you should do so comparing equal or similar MHz ratings and moreover, close-in-time releases.
Just because a 3 GHz chip is slightly faster than a different architecture 2 GHz chip, doesn't mean that it is a better processor. Higher clock speeds are closely related to die thickness (as far as I can remember).
The PowerMac G5, although arguably the fastest personal computer in the world when it came out, was still largely superior in performance to an x86 based computer with same clock speed; even if you don't buy Apple comparisons, the computer was still at par with x86 based computers running higher clock speeds. And it did that without taking advantage of the 64 bit ability.
It would be interesting to see what the G5 chips will be capable of by the time we get all these yonah and Prescott (or whatever they're called) in the Mac.
Everyone here has been ranting about the processors, but I didn't see anyone asking what follows: what exact speeds will the front bus have? And if those speeds are at par with the G5 (1 GHz front bus), what will be the overall power consumption of those boards w/ processor?
Here is a question for those that have time to check this out: how much slower is a 1.67 Powerbook G4 compared to a similarly spec'd 1.67 Centrino equipped PC Laptop?
This could be my imagination, but I believe that Mac compatible applications take up less space on a HD and in the computer RAM. Is this (very appreciated) quality going to disappear?
As far as all the enthusiasm around these upcoming releases, I would like to say this: I have been using G3, G4, Pentium and AMD PCs for the last 6 years. I have not owned/used a single Mac that did not feel perfect, reliable, smooth, rock solid, hiccup free machine. As a teacher, I wish I could say the same about the PCs. So, these great Pentium PCs are the most unreliable and annoying machines in the world. You are probably ready to jump and say that it's because of Windows. Well, I am not so sure Billy is the only one to blame. We'll see what happens.
Lastly, as many frequently like to point out, Apple is a computer manufacturer. I purchased their computers because they were made differently (and dare I say, better). Now, with Intel making their chips, boards and -for the low end, all-in- one design machines- possibly the video and the sound card, what will it really mean to have a "Mac"? Because if it is (once again, like many like to point out) to run Mac OS X, then why should I not be able to just run that OS on a PC?
And we can all say CIAO to RISC and AltiVec. Isn't that great?
A few considerations:
If you want to compare processors and their architectures, you should do so comparing equal or similar MHz ratings and moreover, close-in-time releases.
Just because a 3 GHz chip is slightly faster than a different architecture 2 GHz chip, doesn't mean that it is a better processor. Higher clock speeds are closely related to die thickness (as far as I can remember).
The PowerMac G5, although arguably the fastest personal computer in the world when it came out, was still largely superior in performance to an x86 based computer with same clock speed; even if you don't buy Apple comparisons, the computer was still at par with x86 based computers running higher clock speeds. And it did that without taking advantage of the 64 bit ability.
It would be interesting to see what the G5 chips will be capable of by the time we get all these yonah and Prescott (or whatever they're called) in the Mac.
Everyone here has been ranting about the processors, but I didn't see anyone asking what follows: what exact speeds will the front bus have? And if those speeds are at par with the G5 (1 GHz front bus), what will be the overall power consumption of those boards w/ processor?
Here is a question for those that have time to check this out: how much slower is a 1.67 Powerbook G4 compared to a similarly spec'd 1.67 Centrino equipped PC Laptop?
This could be my imagination, but I believe that Mac compatible applications take up less space on a HD and in the computer RAM. Is this (very appreciated) quality going to disappear?
As far as all the enthusiasm around these upcoming releases, I would like to say this: I have been using G3, G4, Pentium and AMD PCs for the last 6 years. I have not owned/used a single Mac that did not feel perfect, reliable, smooth, rock solid, hiccup free machine. As a teacher, I wish I could say the same about the PCs. So, these great Pentium PCs are the most unreliable and annoying machines in the world. You are probably ready to jump and say that it's because of Windows. Well, I am not so sure Billy is the only one to blame. We'll see what happens.
Lastly, as many frequently like to point out, Apple is a computer manufacturer. I purchased their computers because they were made differently (and dare I say, better). Now, with Intel making their chips, boards and -for the low end, all-in- one design machines- possibly the video and the sound card, what will it really mean to have a "Mac"? Because if it is (once again, like many like to point out) to run Mac OS X, then why should I not be able to just run that OS on a PC?
And we can all say CIAO to RISC and AltiVec. Isn't that great?
I just can't get excited about this... There's just not enough information. Speeds without power numbers and power numbers without speeds. Remember how Prescott was supposed to be this low power wonder at 90nm and wound up higher power than the previous generation? Remember the Itanic-- would have been a great architecture if it came out when it was supposed to. Intel has been stuck in the 3GHz range for as long as IBM has been stuck in the 2s.
I don't think Intel is going to change our fortunes much. What they will do though is stop all the petty bickering about whether Windows is faster that OSX or this Pentium is faster than that Gx.
The other thing this does is kill half the excitement leading up to major events. We know the roadmap years in advance. The software will still be exciting, but I'll miss wondering what the next step in hardware is.
What will be interesting is how Apple handles system design. Intel is always pushing some initiative or another and issuing reference designs for new systems that never see the light of day because Dell et. al. cut out all the new stuff to save a buck. Apple is the kind of company that could bring Intel's dreams to production.
I don't know if that's good or bad, to be honest. I think it could be good if the ideas flow both ways.
I don't think Intel is going to change our fortunes much. What they will do though is stop all the petty bickering about whether Windows is faster that OSX or this Pentium is faster than that Gx.
The other thing this does is kill half the excitement leading up to major events. We know the roadmap years in advance. The software will still be exciting, but I'll miss wondering what the next step in hardware is.
What will be interesting is how Apple handles system design. Intel is always pushing some initiative or another and issuing reference designs for new systems that never see the light of day because Dell et. al. cut out all the new stuff to save a buck. Apple is the kind of company that could bring Intel's dreams to production.
I don't know if that's good or bad, to be honest. I think it could be good if the ideas flow both ways.
I'm not quite following the conventional wisdom (on this thread at least) that Yonahs will be the same speed as G4s...
Not only are the clock speeds faster, but they're supposed to be more efficient. Their FSB is also several times quicker, so where exactly do we draw the conclusion that they're no faster than G4s?
I don't buy it, unless somebody has an in-depth explanation.
Not only are the clock speeds faster, but they're supposed to be more efficient. Their FSB is also several times quicker, so where exactly do we draw the conclusion that they're no faster than G4s?
I don't buy it, unless somebody has an in-depth explanation.
~Shard~ said:What happened to "H"?
I doubt they would do something with an "I" in it. Afterall, the "G" has nothing to do with the manufacturer of the chip (Motorola, IBM),so why use "I"? And then why would it be "I6"? Nah, Apple will come up with something else...
I suggest the "G-More Powerful Than Superman, Batman, Spiderman and the Incredible Hulk Put Together".
or even: hi5 ???
Well, Apple has never called their laptops "notebooks". Their first was called a Portable but all other laptops has been named PowerBook. (Well except for the DuoDock.)hkhaskell said:I'm curious to see if they even call these new computers Powermacs and Powerbooks at all, considering the fact that they are getting away from the PowerPC platform... Somehow I don't think they are going to bother giving the chip a fancy marketing name because the chips are going to be fairly ubiquitous, unlike the PowerPC chips where really nobody else had them.
Maybe we'll just go back to calling it a Mac and a Mac notebook...
Regardless of the CPU.
For a great resource on all Macintosh computers, download Mactracker:
http://www.mactracker.ca/
Pure guess
I actually think that it makes sense for Apple to release a low volume Power Mac to iron out bugs before general release. I'm guessing that a single processor P-4 PM, essentially an upgraded developers box with a real Mac Mother Board will be released early 2006.
Hopefully it will include an exchange/upgrade program for the current developer boxes. Possible because it would only require a MB swap and some changes in the interior plastic.
PowerBook 1.0
PowerMac dual 2.0
PowerMac Cube 550 (over-clocked)
PowerMac AGP 500
Developers P-4
I actually think that it makes sense for Apple to release a low volume Power Mac to iron out bugs before general release. I'm guessing that a single processor P-4 PM, essentially an upgraded developers box with a real Mac Mother Board will be released early 2006.
Hopefully it will include an exchange/upgrade program for the current developer boxes. Possible because it would only require a MB swap and some changes in the interior plastic.
PowerBook 1.0
PowerMac dual 2.0
PowerMac Cube 550 (over-clocked)
PowerMac AGP 500
Developers P-4
I disagree, if you're looking at speed you should compare the fastest chip from each manufacturer. The fastest Opteron against the fastest Xeon, the fastest Turion against the fastest Pentium M.gio64 said:
MHz is part of the design - some chips push the clock rate, others try to get the best IPC. The fastest POWER5 chips are 1.9 GHz, it wouldn't make sense to run a test to compare POWER5 to 1.9 GHz Xeon or 1.9 GHz PPC970.
The PPC970 is no faster at 64-bit than at 32-bit, except for one datatype that is not important for most applications.gio64 said:
The x64 chips, however, are quite a bit faster at 64-bit than at 32-bit. The 64-bit mode has several architectural changes beyond wider integer registers.
Don't focus on one specific attribute like the FSB - look at the overall system design.gio64 said:
The G5 isn't a well-balanced system - it has very fast FSB's (21.6 GB/sec for the dual 2.7) connected to a 6.4 GB/sec memory system. (http://www.apple.com/powermac/architecture.html) Most of the FSB speed is unusable.
Even today's Centrino has faster memory - dual channel 533 MHz, as well as multiple PCI Express connections, SATA,...
How about 1.3 to 1.3?gio64 said:
http://www.barefeats.com/al15b.html
(Of course, the Intel laptop comparable to the 1.67 G4 would be a 2.26 GHz Pentium M with the 533MHz bus, DDR2 memory and PCI Express chipset....)
Get your HD ready for FAT binaries, when the Mac application has to support:gio64 said:
- 32-bit PPC with and w/o AltiVec
- 64-bit PPC
- 32-bit Intel x86
- 64-bit Intel x64
Promisses, yet to be seen
It is a pity that, as it seems, the Apple-IBM partnership ended, so-far, given the
know-how of the later company.
It would have made sense for Apple to have Intel / AMD for their middle products and IBM / Sony etc.
for what one might call state of the art.
The next iteration of the Cell processor, new versions of Linux / KDE will hopefully provide an
alternative to the Mactel.
The incentive and the flair to buy a MACTEL are not very high anymore, that special thing is gone.
It is a pity that, as it seems, the Apple-IBM partnership ended, so-far, given the
know-how of the later company.
It would have made sense for Apple to have Intel / AMD for their middle products and IBM / Sony etc.
for what one might call state of the art.
The next iteration of the Cell processor, new versions of Linux / KDE will hopefully provide an
alternative to the Mactel.
The incentive and the flair to buy a MACTEL are not very high anymore, that special thing is gone.
Evangelion said:Partially off-topic: it's very interesting to see the reaction in here. Not long ago everybody told how Intel/x86 sucks, and how PPC is the future. Now everyone is telling how x86/Intel kicks ass. I guess dictionarys should have an entry called "Doing a 180-degrees change in opinion in a relatively short time: see "Apple's userbase""
Thanks for generalizing....
gio64 said:
All other qualities being the same, it would be better
yes, they are somewhat related to the process they are made on. What's your point?
I know few Opteron-users who would disagree with that claim. Computers based on Opteron were available long before G5 PM's were available.
Depend on the app. If the app relies lots of memory-bandwidth and low latencies, Opteron would have propably crushed the G5.
In many cases, "64-bitness" actually slows the computer down. In some cases it speeds it up. With x86-64, the performance usually goes up, thanks to architecture-improvement that "64bits" offer (twice the number of GP-registers, twice the number of SSE-register)..
Well, there would propably be dual-core G5's, running at about same clock-speed as G5 is running now. But if single-core G5 doesn't destroy it's single-core x86-counterparts, I fail to see how dual-core G5 could destroy it's dual-core x86-counterparts.
AFAIK, Conroe will have 1066MHz bus, Merom will be somewhat slower.
Considerably less that with G5
. No, make that "humungously less than with a G5".Quite a bit slower.
Uh, I don't think there's any notable difference.
Macs will still be Macs, that will not change. The CPU might be different, but it will still be a Mac.
For the most part, RISC and Altivec were just buzzwords that were thrown around. Your computing-experience is not magically better because the CPU you use is a RISC CPU. And for all intents and purposes, modern x86-CPU's ARE RISC-CPU's. As to Altivec... Well, Intel has SSE, SSE2 and SSE3, so it's not that bad.
What differences will you notice when using Macintel? Well, it will be faster. It might consume less power and run cooler. That's about it. Are those BAD qualities in your book?
shawnce said:Thanks for generalizing....
Do you dispute my claim? Yes, SOME people might still think that PPC is the best thing since sliced bread. But before the Intel-announcement, just about ALL mac-users thought that PPC rules, and x86 sucks. And now everyone (excluding few exceptions) is raving how x86 kicks ass and how they can't wait to get their hands on some "Intel Inside" Macs
.Yes, that is a complete 180-degrees change in opinion. And the speed and totality of that change is downright humungous.
Yes, RDF is strong with Steve Jobs
.This first PM G5 was generally a well balanced system. The FSB bandwidth is generally inline with memory throughput and other data targets connected to the "north" bridge (including the other CPU). As CPU clock rate have increased to 2.7 I agree that things have grown a little unbalanced, at least on paper. I have not found a good test that shows the PPC970 running at 2.7GHz to be badly memory bandwidth starved (not factoring in other data sinks). In general it may be little of an issue and so Apple has decided not to bump things up much.AidenShaw said:
Keep in mind that the FSB of G5 is bidirectional so the bandwidth listed is divide between up and down. So for a 1GHz FSB (2GHz CPU) you get 4GB/s up and 4GB/s down (due to overhead you get around 80-90% of that maximum). On a 2.7GHz system you get 5.4 GB/s up and 5.4GB/s down. That is per CPU of course.
PowerMac G5 first generation...
PowerMac G5 current generation...
More found on this page.
Ok, ok, I understand.
Steve Jobs (which makes more money than all of us put together 1000 times over) is a flippin' idiot. All these years he could have had this great system architecture, but as always idiots do, continued to use crap. Makes sense. It also makes sense that the switch came after having introduced a new chip just a couple of years back (when the switch would have been certainly more timely). I think you should all call him up and give him: a) some hints on how to make these new machines, b) some crap for all these years that he has made us use this garbage.
Finally, I see that nobody answered my last question.
If Apple makes PC...
BTW, I won't go back and argue about some statements that I saw (RISC being more instruction that x86... good one).
I just hope somebody figures how to use OS X on a regular PC, so I don't have to spend top dollars to buy just another PC with a pretty logo on it.
Steve Jobs (which makes more money than all of us put together 1000 times over) is a flippin' idiot. All these years he could have had this great system architecture, but as always idiots do, continued to use crap. Makes sense. It also makes sense that the switch came after having introduced a new chip just a couple of years back (when the switch would have been certainly more timely). I think you should all call him up and give him: a) some hints on how to make these new machines, b) some crap for all these years that he has made us use this garbage.
Finally, I see that nobody answered my last question.
If Apple makes PC...
BTW, I won't go back and argue about some statements that I saw (RISC being more instruction that x86... good one).
I just hope somebody figures how to use OS X on a regular PC, so I don't have to spend top dollars to buy just another PC with a pretty logo on it.
Evangelion said:Do you dispute my claim? Yes, SOME people might still think that PPC is the best thing since sliced bread. But before the Intel-announcement, just about ALL mac-users thought that PPC rules, and x86 sucks. And now everyone (excluding few exceptions) is raving how x86 kicks ass and how they can't wait to get their hands on some "Intel Inside" Macs
Yes, that is a complete 180-degrees change in opinion. And the speed and totality of that change is downright humungous.
Yes, RDF is strong with Steve Jobs
Maybe its because in the past 5 years Pentium's have sucked booty in a big way. I'm sorry but the P4's the first and second iteration
sucked. ( When benchmarks prove that clock for clock the P3 kicked the snot out of the P4 something is smelly in Intelville.) Only now is Intel getting the P4 to be a somewhat OK chip with Prescott. It wasn't until the Pentium M came onto the scene that anything out of Intel really hit the sweet spot. Why do you think Apple will be waiting until 2007 to release some of their wares. Current desktop chips don't get the job done. The P4 with its netburst microarchitecture is NOT where Apple want to go with their wares. Who would want to go with a CPU that was designed around marketing which is the only real reason to have such insanly large pipelines. (e.g. You can ramp up the clocks on the chip.)
Review again what the next gen desktop line of chips is doing and compare it to the P4. Its like night and day. I do believe that the G3/4/5 all kicked the snot out of the Pentium at certain points in time. But as always Moto couldn't keep up or at the very least didn't want to dump the $$$ into R&D to keep up. Many things can be said about x86 but the biggest is that Intel vs. AMD keeps the market fresh. Well burp the tupperware Flo. Apple's about to get a kick in the pants in the freshness department and all I can say is WOOT!
Of course Apple users are going to go nuts now. What happens now if Apple releases updates in a more timely manner? It would be like The Simpsons episode where Mayor Quimby turned up the city's lights so it was day 24 hours a day. Drove all the animals insane. Apple users are use to mild to moderate updates. Slow and steady. I fear for Apple user's sanity if Apple updates their wares once every few months.
I think a 10 step program will have to be started by Apple for people who can't take that their laptop is not the latest greatest 4 months after they get it. J/King guys.
That's the problem with switching to Intel
I'm afraid this will be the principle attitude once Apple has migrated to Intel.
In the past, Apple had unique hardware, unique design, and unique OS.
After the switch, Mac users really will get to compare Apples to Apples, and Oranges to Oranges. The specs and hardware Apple sites for their latest Intel PB will be the same specs and hardware that Dell sites for it's Inspiron.
The principle difference will be price and OS. If the OS becomes available on x86 or people like the features of Vista near the same as Mac OSX, then the only difference will be price.
I'm sorry, I like Apple, but they will never be able to compete with Dell solely on price. I also don't think Mac fans will pay 30 to 40% more for a prettier package, logo, and form factor.
gio64 said:
I'm afraid this will be the principle attitude once Apple has migrated to Intel.
In the past, Apple had unique hardware, unique design, and unique OS.
After the switch, Mac users really will get to compare Apples to Apples, and Oranges to Oranges. The specs and hardware Apple sites for their latest Intel PB will be the same specs and hardware that Dell sites for it's Inspiron.
The principle difference will be price and OS. If the OS becomes available on x86 or people like the features of Vista near the same as Mac OSX, then the only difference will be price.
I'm sorry, I like Apple, but they will never be able to compete with Dell solely on price. I also don't think Mac fans will pay 30 to 40% more for a prettier package, logo, and form factor.
thanks
There's a lot of "FSB MHz myth" floating around these parts!
BTW, I wonder if the Intel switch means the end of HyperTransport on Apple motherboards. My guess would be yes - it would be rather strange to glue an AMD HT Tunnel off the Intel chipset - especially since Intel is putting PCIe off the northbridge!
Thank you for a nice *system* view of how the components work together.shawnce said:
There's a lot of "FSB MHz myth" floating around these parts!
BTW, I wonder if the Intel switch means the end of HyperTransport on Apple motherboards. My guess would be yes - it would be rather strange to glue an AMD HT Tunnel off the Intel chipset - especially since Intel is putting PCIe off the northbridge!
shawnce said:
It boils down to this. The G5 platform didn't expand in any meaningful way after the first itteration. This isn't new for Apple. The same thing happened to the G4. 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. (In the case of the G4 PowerMac I think it not only spoiled but really started to stink up the place but what recourse did Apple really have?) Apple really doesn't STAY at the cutting edge of tech. And please guys. Don't give me the same old line about what about fiber optic keyboards in the PowerBook or dual layer burners and such nonsense. Its no different then Ford adding bling bling hubcaps to last years Focus. In the end its the same old car warmed over. Now WHY this is the case. *shrugs* Who knows. It really doesn't matter at this point does it?
As for the person saying the G5 system isn't balanced. Wow. What are you huffing? The G5 had one of the most balanced systems out there with very few real bottle necks anywhere in the system. The G5 was a peice of art tech wise AFAIC. If only IBM and Apple could have continued to push the system. It really is too bad. Every time I look at the diagrams of the G5 I really think Apple and IBM did one heck of a fine job on that system.
Agreed (however I believe the next PM using PPC will bring things current-ish, god I hope!... please at least PCIe for video adapter).SiliconAddict said:
Why it happens... well likely as long as things sell well enough from Apple's point-of-view why not ride the gravy train using an existing chip set as long as you can. It isn't cheap to develop chip sets and mother boards (chips set development and production startup is the larger cost, the biggest cost for Apple).
Now with the Intel switch I believe development/production of chip sets will be removed from the equation (at least to some extent). This will likely help accelerate Apple's revision rate (what they bump in a revision) but again it depends on sales and costs.
Personally that is why I usually buy the first generation systems (possibly waiting for the first revision) of any new architecture from Apple since it likely bring up close to parity in the market and you can ride that system for at least a couple years until the next major overhaul since often the interim releases don't track the current tech in market well.
Again the Intel switch will likely change this some... which is a good thing.
Agree. It was one of the better overall system designs at the time and one of the best Apple has ever released IMHO (putting EE hat on).
I don't know if this is old/off topic, but those Steve Balmer videos are all you need to convince yourself that a Mac is better than a Windoze PC.
Thanks for making them readily available, they are empowering.
How can anything good come from a company that has guys like that running the show?
Thanks for making them readily available, they are empowering.
How can anything good come from a company that has guys like that running the show?