But Uncle Steve, I thought Intel chips were faster?
The biggest problem I can see with Apple's switch to Intel is the fact that Apple will have to work against much of its own marketing. Take, for example, this graphic (source):
According to this chart the PowerPC smokes the Intel chips, and does it handily. Therefore, when Apple is selling Macs with Intel in them AND Macs with PowerPC's in them how are they going to deal with the Megahertz Myth (in reverse)?
According to this chart the PowerPC smokes the Intel chips, and does it handily. Therefore, when Apple is selling Macs with Intel in them AND Macs with PowerPC's in them how are they going to deal with the Megahertz Myth (in reverse)?
![TUAW [Cafepress]](http://www.blogsmithmedia.com/www.tuaw.com/media/tuaw-cafepress-promo.png)
Reader Comments (Page 1 of 1)
Motor City Rollie said 4:19PM on 6-16-2005
I think the comparisons will be directly made to the G4. The first products to incorporate the Intel Chip setting will be in the General Consumer Entry Arena.
Reply
Piko said 4:19PM on 6-16-2005
Keep in mind, Jobs was referring to the future performance of Intel processors that we have not yet seen. Remember that chart showing Performance vs Watts. The PowerPC chip was up against a glass ceiling of 15 while the Pentium was projected at 70.
Intel has a year to prove to themselves they can crank out a CPU that is faster than the G5 2.7Ghz. Now, if IBM should happen to whip out a 3 or 3.5 Ghz PowerPC in the meantime, you think Apple would think twice about switching (you bet they would)?
The biggest thing that Intel has going in their favor is the performance to heat ratio. Their chips are running amazingly cooler than the PowerPC which is why we still haven't seen a G5 in a laptop.
Piko
Reply
aj said 4:19PM on 6-16-2005
Of course SPECfp is a test that takes advantage of the G5's amazing number of Altivec and general-purpose floating-point registers. The G5 in integer, ehhhh, not so much. That's why His Steveness pointed up the "70 integer operations per watt" on the Intel roadmap. In real application terms, integer ops are a better guide to overall performance.
If the next generation of Intel chips includes things to satisfy the media-creation and sci-tech crowd - like a hefty amount of SS3 registers for instance - then we'll get a good balance of general and FP performance, for once.
Reply
Oli said 4:19PM on 6-16-2005
You do realize that this is the performance PowerPCs and Intels deliver NOW. Apple's reason for switching is the performance these chips will deliver next year or the year after that. Hint: Intel might get faster.
Plus there is no graph showing the G4's performance against Intel's mobile processors :-)
Reply
Ben said 4:19PM on 6-16-2005
Who knows what Apple/Intel have up their sleeves and whether a new processor will arrive by next year to even the masses. We can't assume that Apple will use Intel processors like the currently available ones.
Reply
nezromatron said 4:19PM on 6-16-2005
There are lies, damned lies, then there is marketing.
I think Apple will have little trouble convincing people that the Intel chips are better. As stated the chips that the Apple machines will use are still a year our, so we might see some sort of quad-core chip busitng out by then.
Reply
Chad Johnson said 4:19PM on 6-16-2005
There was a very in-depth article, http://www.anandtech.com/mac/showdoc.aspx?i=2436 but it found some drawbacks with the PPC with respect to kernel level locking and threads. The PPC had better memory throughput and faster integer calculation, though it sucks power like its going out of style. It did well with single-threaded tasks, like rendering a movie, but it did very poorly with things like Apache and MySQL which spawn 50 or 60 threads. The PPC uses some kind of virtualization that abstracts processor threads from programmer threads. So, there was much larger penalty for creating and destroying threads. The x86 compilers had more direct control over the kernel and threads.
Then, at about that point, I start getting confused
Reply
Blue Balloon said 4:19PM on 6-16-2005
My sources
Mac OS X running different than Window XP over speed
Intel will design different CPU, unlike P4 or Xeon
More staff to be hiring for Intel... for short time. *giggle*
Reply
Maurice said 4:19PM on 6-16-2005
I was planning on buying a Powerbook G4 in the next few weeks (my first apple!) and I've been reading posts & comments here and at Engadget for the past while now.
The only thing that I want to know is, is it worth buying a powerbook now or when they come with Intel chips? I'll be using it for web development and graphics, so what impact will the Intel chip have on that? Thanks.
Reply
Brian said 4:19PM on 6-16-2005
Well according to that graphic, divide the dual G5 score by two and you get 7.85. That's lower than the speed of the 3GHz P4. With the new dual-core Pentiums I think Apple will find a way to do amazing things with Intel.
Reply
Maurice said 4:19PM on 6-16-2005
I was planning on buying a Powerbook G4 in the next few weeks (my first apple!) and I've been reading posts & comments here and at Engadget for the past while now.
The only thing that I want to know is, is it worth buying a powerbook now or when they come with Intel chips? I'll be using it for web development and graphics, so what impact will the Intel chip have on that? Thanks.
Reply
Mike Steinbaugh said 4:19PM on 6-16-2005
This chart compares apples and oranges. Wow, the dual G4 system wins out over a single Intel chip system. Big surprise. However, if you had a dual Intel chip system, Intel would win out. So I fail to understand why this is a significant graphic.
Reply
djones said 4:19PM on 6-16-2005
I think there's something bigger in store at Cupertino. Steve was pushing the idea that power consumption was more important to them than performance. That message came across to me much heavier than places are reporting it.
This makes me think that Steve has something quite revolutionary and culture-changing up his sleeve for the way we use computers. Maybe that tablet-looking thing that they secured a design patent for? I think Apple's about to transform our concepts of computers and mobility.
As for new PowerMac performance running on Intel? Don't worry about it. I'm sure by the time they get to those models, they will be using very fast dual-core Intel CPUs in the desktop models.
Reply
Sean Ouimet said 4:19PM on 6-16-2005
Maurice: I've been using my Powerbook since fall 2004 (when I 'made the switch') and I couldn't be happier with my decision. Not only does it handle my (similar) web/graphic development needs, the OS stability, iLife extras and new OS X 10.4 features are worth the switch alone. (I've managed to convince at least 3 other individuals + my office to go mac since that time.)
In regards to Intel, all >new< software developed "for" the Intel platform Macs should be 100% compatible with PPC basec macs for several years past 2007. Remember, the transition itself will take two years, and that is the point at which they want 100% of their developers and new computers operating with Intel support.
Thanks to Dual Binaries in XCode, support for both architectures in any newly developed software should continue far past this point. Remember, they'll still be releasing NEW models of PPC based macs within this time. I'd say your more than safe in making the investment. Make the switch, you'll be glad you did!
Reply
Dogzilla said 4:19PM on 6-16-2005
I have to echo Mike Steinbaugh's comments above: this chart shows a dual 2-ghz G5 compared against 3Ghz Xeons (which are designed for servers, not general use) and a single 3ghz P4. What does this prove, exactly? How about comparing 2ghz G5s against 2ghz P4s?
Reply
Sam Kass said 4:19PM on 6-16-2005
1. These look like the original Pentium 4 processors, probably on anemic motherboards compared to what's available today. A single processor 3.6GHz Pentium 4 today clocks in at over 22, and dual Xeon at 30-40 on the SPEC CPU2000. The fastest machine you can buy in the Intel world gets over 100. Same with SPECfp. The whole point of this switch is that now Apple has options.
http://www.spec.org/osg/cpu2000/results/res2005q2/#SPECint_rate
2. SPEC is very compiler dependent. One thing I didn't note was whether the Intel compiler, which spanks gcc's booty, will be available to Mac developers. That could make a huge difference.
3. If today's PowerBooks give you what you need today, there's no reason not to buy. They'll be supported for years, likely as long as their useful life would have been, anyway. That being said, next year's Pentium M (which is going dual-core at year-end,) should be rather amazing.
Reply
Mark Wingad said 4:20PM on 6-16-2005
Hi
Er rember those tests were made against P4's running windows.
the biggest question becomes how much does Windows suck
CPU wise.... We will find out which OS eats more CPU cycles and
the biggest CPU hog. This will really Show whats going on for the
1st time ever.
Reply
Maurice said 4:20PM on 6-16-2005
To: Sean
Cheers mate. To be honest I was still thinking of buying one, but wanted a mac user's point of view. I don't really see the big difference in the Intel switch myself, especially for what I have in mind for the pb, but hey it's worth finding out. Thanks again.
Reply
Harry Richman said 4:20PM on 6-16-2005
I was on that same website wondering the same thing; why has apple left this up, and is it true. http://www.apple.com/powermac/performance/ , look at the titles, almost all of them are some sort of heavy rendering, which a 64 bit is perfectly suited to do, not many people are rendering bibng files on a day to day basis.
Beside in a year or two, I'm sure Intel will have 64 bits throughoout their line.
Reply