PDA

View Full Version : What's so great about the Intel Macs?




EHUnlucky7x9@ao
Jan 20, 2006, 09:37 PM
Hey guys, I've been reading everyone's comments about the Intel Macs and I still can't get ground on WHY the Intel Macs are so great? I'm not sure if this has been discussed before, so if you can show me the thread, I"ll gladly do my reading. But I've searched and searched online and I can't find the REASONS why it's better. I always thought that the PowerPC processors were better? I've own macs since the G3's first came out and everyone always praised the Macs over the Pentiums. What's changed now? Does the Intel processors run faster than dual-core PPC's? Let me know your thoughts and/or share your knowledge.

Oh yeah, by the way... how does the Macbook beat out the Powerbook? They look the same, probably do the same thing... boy...this whole Intel thing has me confuzzled. :-D :p



Avarus
Jan 20, 2006, 09:59 PM
Hey guys, I've been reading everyone's comments about the Intel Macs and I still can't get ground on WHY the Intel Macs are so great? I'm not sure if this has been discussed before, so if you can show me the thread, I"ll gladly do my reading. But I've searched and searched online and I can't find the REASONS why it's better. I always thought that the PowerPC processors were better? I've own macs since the G3's first came out and everyone always praised the Macs over the Pentiums. What's changed now? Does the Intel processors run faster than dual-core PPC's? Let me know your thoughts and/or share your knowledge.

Oh yeah, by the way... how does the Macbook beat out the Powerbook? They look the same, probably do the same thing... boy...this whole Intel thing has me confuzzled. :-D :p

Duo Core's run faster and cooler than a single PPC proccerssor. And good luck trying to fit one of those into a Power Book or iMac, Duo Core's are better and more efficient technology.

MacsomJRR
Jan 20, 2006, 10:12 PM
I always thought that the PowerPC processors were better? I've own macs since the G3's first came out and everyone always praised the Macs over the Pentiums. What's changed now? Does the Intel processors run faster than dual-core PPC's? Let me know your thoughts and/or share your knowledge.


Not all Intel chips are created equal. The chips that the new Macs are using are just like the above dude said. They are faster, run cooler etc... They are just better chips. What else do you want?

They aren't pentiums also.

Anonymous Freak
Jan 20, 2006, 10:29 PM
Ars Technica (http://arstechnica.com) has good 'Processor Review (http://arstechnica.com/articles/paedia/cpu.ars)' articles whenever a new architecture comes out. Look for one on the CoreDuo processor soon. (He had FOUR more-detailed-than-former-Intel-employee-me-can-handle articles on the G5 and history of PPC, when the G5 came out.)

But the essence is that the PowerPC became less efficient per clock in recent revisions, while the Pentium-M/CoreDuo architecture became more efficient per clock. And the Core Duo is MUCH more efficient 'per Watt', as Jobs now likes to point out. This derives from the fact that the Core Duo (and the Pentium-M predecessor) was designed from the ground up for good performance with the minimum power usage.

PowerPC's big selling point early on was that it was a 'simple' RISC architecture, where most of the hard work was actually in software. So if you wrote your program efficiently, it would run fast. And because the processor was simple, it wouldn't be hard to ramp up processor speeds. Unfortunately for Apple, Intel had more pressure to increase speed (in AMD) than Motorola or IBM did. The main market of the PowerPC really isn't PCs, it's as 'embedded controllers' in things like Cable modems, DirecTV receivers, traffic light controllers, satellites, and other non-computer devices, where all-out processor power isn't the main selling point. So Intel ramped up processor speeds in their architecture much faster than the PowerPC camp did. (In fact, Intel did this partly by having the core of the processor actually be RISC, with internal translation.) Intel's might in developing new manufacturing processes also helps here. IBM is good, Intel is better. (i.e. if Intel were to take the G5 design, and tweak it for their manufacturing plants, they could probably get 3 GHz out of the design no problem.)

Later, the addition of the excellent 'Altivec' vector instruction set allowed the G4 (and later the G5) to perform amazingly fast calculations on sets of repeating data. (Things like photo, audio, and video processing make heavy use of this.) At the time, Intel's counterpart, MMX, was woefully underpowered. However, over time, Intel introduced SSE, SSE2, and the latest, SSE3; which makes their vector processing at least equal to Altivec, if not better. (I haven't seen a true unbiased comparison between SSE3 and Altivec yet.)

So between clock speed increases going better for Intel, and the leveling of the core architecture; the Core Duo's base architecture SHOULD be about equal to a G5 at the same GHz rating. That means that the dual-core in the Core Duo should be about twice as fast as the single-core G5. Yes, Apple always blows performance improvements out of proportion. They always have. But realistically, a 2.0 GHz iMac Core Duo should perform about equally well (on well optimized native apps,) as a dual-processor 2.0 GHz Power Mac G5.

DrEasy
Jan 20, 2006, 10:46 PM
PowerPC's big selling point early on was that it was a 'simple' RISC architecture, where most of the hard work was actually in software. So if you wrote your program efficiently, it would run fast. And because the processor was simple, it wouldn't be hard to ramp up processor speeds. Unfortunately for Apple, Intel had more pressure to increase speed (in AMD) than Motorola or IBM did. The main market of the PowerPC really isn't PCs, it's as 'embedded controllers' in things like Cable modems, DirecTV receivers, traffic light controllers, satellites, and other non-computer devices, where all-out processor power isn't the main selling point. So Intel ramped up processor speeds in their architecture much faster than the PowerPC camp did.
Thank you, this really helped me get the "big picture" I was missing.

MisterMe
Jan 20, 2006, 10:51 PM
....

PowerPC's big selling point early on was that it was a 'simple' RISC architecture, where most of the hard work was actually in software. So if you wrote your program efficiently, it would run fast....You are confusing RISC with EPIC. The selling point of RISC is that it eliminated those instructions that were not used by compilers. This allowed the designer to decrease the decoding stage and increase the number of registers. The resulting processor allowed for relatively simple compilers to produce code that was nearly as efficient as assembly language. EPIC processors have very little onboard intelligence for such things a code-ordering and branch prediction. Instead, responsibility for these functions were the responsibility of very sophisticated compilers.

EHUnlucky7x9@ao
Jan 20, 2006, 10:52 PM
Wow, I see...so here's another question...I know people say that when you purchase a Mac and it's processor speed, you have to double it in order to get it's PC-equivalent. With these new Mac's, does that thought still apply or no more?

DeSnousa
Jan 20, 2006, 10:57 PM
Wow, I see...so here's another question...I know people say that when you purchase a Mac and it's processor speed, you have to double it in order to get it's PC-equivalent. With these new Mac's, does that thought still apply or no more?

It never really went that way, its just that the os operated apps well enough and in a way that made the computer experience better.

I would not even place a speed difference on the chips between Mac and Windows. Keep in mind they are 2 os. I guess you could run photoshop tests and etc to compare. But we should wait for better optimised native apps first.

maclamb
Jan 20, 2006, 11:05 PM
I got the intel imac for these reasons:
1. Dual core cpu in a non-pro machine (i don't need a powermac)
2. Ability to do extended desktop to my 20" LCD (so I could get the 17 and save $)
3. Quiet and cool (quieter than anything else inlcudes my powerbook)
4. When UBs are fully available it will scream - and it ain't bad now...

Anonymous Freak
Jan 21, 2006, 12:07 PM
You are confusing RISC with EPIC. The selling point of RISC is that it eliminated those instructions that were not used by compilers. This allowed the designer to decrease the decoding stage and increase the number of registers. The resulting processor allowed for relatively simple compilers to produce code that was nearly as efficient as assembly language. EPIC processors have very little onboard intelligence for such things a code-ordering and branch prediction. Instead, responsibility for these functions were the responsibility of very sophisticated compilers.

I was referring to the original idea of RISC and PowerPC. Over time, PowerPC processors became 'smarter', and therefore, more complex. Yes, EPIC and VLIW are the latest incarnations in the 'move the work to software' approach. Even then, Itanium's core processor (not including extravagant amounts of cache,) is larger than the P4's core. I think only Transmeta, with their Crusoe, got it right. And we see where that got them.

superbovine
Jan 21, 2006, 02:22 PM
i've seen real world benchmarks on dual processor PPC vs dual processor amd ...video encoding PPC wins hands down, however things like audio encoding and gaming benchmarks x86 wins. I think it shows, that certain application PPC might be better suited for, but perhaps not in the bigger picture down the road.

janstett
Jan 21, 2006, 02:47 PM
Hey guys, I've been reading everyone's comments about the Intel Macs and I still can't get ground on WHY the Intel Macs are so great? I'm not sure if this has been discussed before, so if you can show me the thread, I"ll gladly do my reading. But I've searched and searched online and I can't find the REASONS why it's better. I always thought that the PowerPC processors were better?

You bought into Steve's reality distortion field.

Interesting that he used the Spec2000 tests to show how much faster the Intel Macs are. The truth is, the Spec2000 tests have ALWAYS showed the Intel processors to be faster than their PPC equivalents (with small exceptions). All that "fastest PC in the world" stuff was just bunk. So you have to remember, a lot of this evangelism around the Mac is and has been disposable hype, rabid marketing.

The Intel Macs are better for Apple because they free themselves from the whims of Motorola/IBM/FreeScale. Since they are in such a tough market, they have a harder time competing with Intel. And as of the last few years it's become more and more obvious that they didn't want to fight the battle anymore.

Also, look at the trend Apple followed, even before the Intel switch. They started using PC standard components (IDE, PCI, etc.) and now they're using Intel processors, chipsets, and probably motherboards. As opposed to developing all this stuff themselves and begging Motorola to try harder. The days when computer companies used to be able to do all this proprietary engineering are over (see Silicon Graphics). Apple, let's not forget, is the last holdover from the days when there really were different computers and computer companies and nobody had more than 20% of the market (I'm talking the days of Commodore, Atari, Coleco, TI, Timex Sinclair).

Let's all pause to think what the Mac today really is when distilled to its basic elements -- a fancy UI on top of BSD Unix running on Intel processors.

Apple (and Steve) have always refused to admit that SOFTWARE, IE the Operating System and applications like QuickTime, are their core talent. Steve always had to have the vice grip lock on the hardware, he has a romanticized view of the hardware being special. The move to Intel shows that maybe Steve is learning that software is what's important not hardware -- he's not cutting off his nose to spite his face anymore. If he decides to open up Mac OS on any PC, he can let go of his last obsession and give a big middle finger to Bill Gates at the same time.

Another thing is that it was clear the G5 was never going to be in a notebook, and IBM didn't care apparently. So the Powerbooks were dead in the water while Intel was doing interesting things with Centrino re: good power consumption and power. Steve said so in the keynote. Now throw in efficient DUAL CORE processors in a laptop, and watch out, this is good stuff. Intel suddenly isn't the evil empire anymore. They're a hard-working supplier that actually works on improving their chips.

Basically this switch allows Apple to get more standardized components and work with a processor line that will actually get some money spent on it and see rapid development change.

janstett
Jan 21, 2006, 02:59 PM
Wow, I see...so here's another question...I know people say that when you purchase a Mac and it's processor speed, you have to double it in order to get it's PC-equivalent. With these new Mac's, does that thought still apply or no more?

You can't even say that with PCs anymore. As someone said above, the Pentium M chips really changed the game, and a 1.4 GHz Pentium M is not the same as a 1.4 GHz Pentium IV.

Intel was playing a game of more complex processors and faster clock speeds all the time up to the Pentium IV. They took a sidetrack with the Pentium M and that side track ended up being the future. Right now I don't see a future for the Pentium IV (topped out at 3.6 GHz), 64-bit Itanium is dead too.

AmbitiousLemon
Jan 21, 2006, 03:25 PM
Technically speaking the PPC chips still are much better than any Intel chips (even the CoreDuo). The old PC fan boys around here (switchers) will try to argue that the Intel chips were always faster and point to things like spec. Truth is Intel chips cheat at these benchmarks requiring you to look at real world performance to assertain true speed. And in nearly all real world benchmarks PPC wins over Intel (AMD is another story).

PPC chips almost always have been faster than the best offerings from Intel (including the current moment in time). The latest battles have mostly been between AMD chips and PPC chips with various processors released by each camp leap frogging the competition. The reason the PowerMac isn't Intel yet is that Intel doesn't have a chip that would match the current performance of the PowerMacs, let alone anything that would constitute an upgrade.

The advantage to Apple moving to Intel isn't speed or power per watt (as the marketing guys would like you to think). The main reason for the switch is that IBM embarassed Steve - by not delivering on the promise of a mobile G5 and a 3Ghz G5.

And thats where the advantage of Intel comes in. The switch makes Macs slower in the short term, and relegates us to probably never having the fastest processor (AMD has taken over that crown in the x86 crowd), but what it does give us is parity with PC vendors (they will be using the same stuff) and a more reliable source of chips.

This doesn't necessarily mean no more broken promises. Intel has a long history of not meeting roadmap deadlines, watering down features on planned chips, and grossly overstating the expected performance of planned processors. But we know we will be getting the same crap as everyone else. And supply problems should be a thing of the past (although currently we are having CoreDuo supply problems). But we know Intel is dedicated to improving processors for PC use, while the PPC camp seems to have lost interest in PC processors. So Intel represents a more dedicated (if promiscuous) partner.

This has nothing to do with technical features, speed, or watts - this has to do with business, and finding a partner who is dedicated to the product they are supplying you.

MattG
Jan 21, 2006, 03:31 PM
All I know is...I want mine ;) :D

Catfish_Man
Jan 21, 2006, 05:30 PM
Technically speaking the PPC chips still are much better than any Intel chips (even the CoreDuo). The old PC fan boys around here (switchers) will try to argue that the Intel chips were always faster and point to things like spec. Truth is Intel chips cheat at these benchmarks requiring you to look at real world performance to assertain true speed. And in nearly all real world benchmarks PPC wins over Intel (AMD is another story).

PPC chips almost always have been faster than the best offerings from Intel (including the current moment in time). The latest battles have mostly been between AMD chips and PPC chips with various processors released by each camp leap frogging the competition. The reason the PowerMac isn't Intel yet is that Intel doesn't have a chip that would match the current performance of the PowerMacs, let alone anything that would constitute an upgrade.

Obligatory credibility establishment: I've been a Mac user literally my entire life and I write Mac software.

You're incorrect, unless you count POWERx as PowerPC. Intel pulled ahead of PowerPC around 2000-2001, except in certain carefully chosen benchmarks (RC5, some photoshop filters, etc...).

The G5 brought PPC back to being quite competitive; faster in many cases, slower in others. However, it stagnated after that (pick a theory as to why; my theory is a combination of shoddy support chipset and lack of interest from non-Apple companies).

Currently the G5 is competitive at streaming floating point stuff like Photoshop and video apps, and gets destroyed at latency sensitive and int heavy stuff (compiling, for example). The quad G5 is obviously very fast, but if you compare it to a quad Xeon system, it's again just competitive (however, Intel artificially inflates prices for Xeons, so the quad G5 would likely cost less than the quad xeon)

Re: SPECcpu. It's actually very difficult to cheat at SPEC. There are rigorous submission rules, and potential cheats are investigated when discovered. That said, there are a few known "legal but nasty" compiler optimizations, such as Sun's 179.art hack. I would regard SPECcpu as more reliable than any other synthetic benchmark for the tasks it measures (mostly engineering and scientific computing), but not as accurate as a *well done* (that means non-Apple. Apple cheats horribly and blatantly) application-specific benchmark.

Apparently the new low end iMac is 256% faster than a 1.6GHz G5 PowerMac at compiling using GCC, according to a test on the arstechnica Mac forums. For most apps, though, the speed increase is in the 10-30% range, and for some the new machine is slower. I'm fairly curious about the compile numbers though. I've heard from other sources that compiling is "very fast" on the Intel machines, but I haven't tracked down anyone to confirm just how fast it is, so I'm taking the number above with a big grain of salt until I do.

<edit>
One theory I've heard about why GCC is so much faster on the mactels is that its x86 version has just had waaaaay more attention paid to it by the legions of open source x86 hackers. Probably at least somewhat true.
</edit>

AmbitiousLemon
Jan 21, 2006, 05:54 PM
Well Catfish man you have an odd way of looking at things. What you said in most part completely agrees with what I have said. You had to go to an extremely expensive high end Intel cpu to find something comparable to a PowerMac. But that isn't playing fair. If you want to look at a higher class then you need to do so on both ends. Compare it to a Power cpu.

Independent benchmarks consistently show the G5 outperforming Intel chips (even holding its own against Xeons and Opterons, which isn't exactly a fair comparison). You also mentioned that Intel pulled ahead around 2000. This again shows the PPC has been ahead most of the time and it was just a short span of time during early 2000s before the G5 came out that the G4 wasn't competing. With the G5, PPC has pulled ahead again and even with the CoreDuo the G5 remains ahead (hence why no Intel PowerMac). Current independent benchmarks confirm this.

You also mention cost, mentioning the Xeon system is much more expensive. Also of note is that the CoreDuo is also a more expensive chip than the G5. Apple is having to go with a more expensive cpu for the iMac and Pro laptop in order to garner just a 10% improvement in real world benchmarks. A bump up to a faster G5 probably would have provided a similar improvement in speed (and perhaps at lower cost).

PPC technology is simply hands down superior technology to x86, and if you think different you are fooling yourself. The switch to Intel is probably a very good move, but lets not confuse people by trying to pretend the benefits of the switch are technological. Its as I said before, the switch is a good idea because developers of PPC are no longer focused on creating CPUs for personal computing. While the PowerMac remains competive Apple has not been provided good upgrade options for the laptops or low end desktops. Intel chips provide good cometitive products across the spectrum and also provides a more reliable source of chip (since Intel is very much interested in creating cpus for personal computers).

EHUnlucky7x9@ao
Jan 21, 2006, 09:46 PM
Wow, all this info is interesting. So IBM hasn't made a mobile G5...it takes time. So Intel will be doing that job now?

AmbitiousLemon
Jan 21, 2006, 09:54 PM
Wow, all this info is interesting. So IBM hasn't made a mobile G5...it takes time. So Intel will be doing that job now?

Well the G5 is a PowerPC processor. It is being abandoned. We are switching to x86 processors made by Intel. The CoreDuo (cpu in the new MacBook Pro and iMac) IS a mobile chip. So in essence it is the mobile G5 that IBM had no interest in creating.

I said a few times the reason the PowerMac wasn't updated was that its still faster than any Intel chip. The other end of that question is why the iMac and Macbook Pro then? The answer would be that these are the best selling Macs. No doubt Apple wanted to move its most popular lines to Intel to send a strong message to developers that its time to get those Universal Binaries out. I think thats also why these machines (great machines don't get me wrong) seem a bit half-baked by Apple standards. They are a warning shot. Priming the development environment for Intel Macs.

NeuronBasher
Jan 21, 2006, 10:28 PM
I said a few times the reason the PowerMac wasn't updated was that its still faster than any Intel chip. The other end of that question is why the iMac and Macbook Pro then? The answer would be that these are the best selling Macs.

The fact that the G4 PowerBooks are extremely slow by todays standards and have been showing their age for 18 months didn't hurt the decision to do the laptop refresh early on either, I wager.

AmbitiousLemon
Jan 21, 2006, 10:38 PM
The fact that the G4 PowerBooks are extremely slow by todays standards and have been showing their age for 18 months didn't hurt the decision to do the laptop refresh early on either, I wager.

:) indeed.

SnarkMan
Jan 21, 2006, 11:28 PM
I updated my old 1Ghz G4 iMac to a new 20" Intel iMac this week and I'm lovin' it. I realize it's only a little faster (or a little slower, depending upon what you're doing) than the G5 machines, but it's a lot faster than what I had.

I'm looking forward to the same thrill ride next month when my MBP replaces my used-to-be-top-of-the-line Powerbook. Until then, my Powerbook now feels really slow. That's the only real downside of my new iMac. :-)

EHUnlucky7x9@ao
Jan 22, 2006, 12:06 AM
How soon do you guys think that they will come out with faster speeds on MBP and how short/long is the horizon for the PowerMacs to have these processors?

shrimpdesign
Jan 22, 2006, 01:12 AM
I updated my old 1Ghz G4 iMac to a new 20" Intel iMac this week and I'm lovin' it. I realize it's only a little faster (or a little slower, depending upon what you're doing) than the G5 machines, but it's a lot faster than what I had.
That's awesome! All I have right now is a G4 PowerMac and iBook. I'm going to be blazing with my new Core Duo iMac. It'll probably be evern faster than my Dual 1Ghz G4 in Rosetta if I upgrade to 1.5Gb or RAM. From what I hear, Rosetta likes RAM as much as I like Macs.

benbondu
Jan 22, 2006, 01:15 AM
How soon do you guys think that they will come out with faster speeds on MBP and how short/long is the horizon for the PowerMacs to have these processors?

I would imagine that many professionals would like to see more universal apps released before they trade in their Quad G5 for an Intel machine. I'm not a professional user, but I still don't think it makes much sense to get an Intel Mac until most of the programs I use run native. Hopefully for many applications that will coincide with an Intel PowerMac launch. Maybe by then Intel will have a 64-bit chip ready, although I suppose you can argue whether or not that's necessary.

EHUnlucky7x9@ao
Jan 22, 2006, 08:13 AM
Hey, if you guys ever get rid of your gual G5's, let me know...i'll take them off your hands.

So to stick with one of the thoughts mentioned before.......I primarily do video editing, Photoshop, and DVD authoring.... my best bet is to stay with the PPC since they work better, right?

The programs to keep in mind for me to be constantly using is iMovie + Final Cut, Photoshop, iDVD + DVD Studio Pro.

dwd3885
Jan 22, 2006, 08:44 AM
Hey, if you guys ever get rid of your gual G5's, let me know...i'll take them off your hands.

So to stick with one of the thoughts mentioned before.......I primarily do video editing, Photoshop, and DVD authoring.... my best bet is to stay with the PPC since they work better, right?

The programs to keep in mind for me to be constantly using is iMovie + Final Cut, Photoshop, iDVD + DVD Studio Pro.

i have one, check my thread

godrifle
Jan 22, 2006, 09:42 AM
The switch makes Macs slower in the short term, and relegates us to probably never having the fastest processor (AMD has taken over that crown in the x86 crowd), but what it does give us is parity with PC vendors (they will be using the same stuff) and a more reliable source of chips.

How does moving over to x86 via Intel make it anything but more likely that we'll see the use of AMD CPUs? I'd argue that the wholesale move of OS X to x86 from PPC makes it more likely, since the heavy lifting from PPC to x86 has been done. Adding AMD processors to the product mix is nowhere near as complex.

Catfish_Man
Jan 22, 2006, 04:13 PM
Well Catfish man you have an odd way of looking at things. What you said in most part completely agrees with what I have said. You had to go to an extremely expensive high end Intel cpu to find something comparable to a PowerMac. But that isn't playing fair. If you want to look at a higher class then you need to do so on both ends. Compare it to a Power cpu.

POWER cpus are in an entirely different class. The support chipset for a POWER5 costs more than an entire 2-way Xeon or G5 machine. Comparing them is ludicrous, and the benchmarks reflect this (>3000 SPECfp for a POWER5+)


Independent benchmarks consistently show the G5 outperforming Intel chips (even holding its own against Xeons and Opterons, which isn't exactly a fair comparison).

Actually Xeons and Opterons are exactly the fair comparison. They're identical to the Pentium-4 and Athlon64, except they support dual processors like the G5 does (I'm not saying compare against the large-cache high end Xeons. I agree that would be unfair).


You also mentioned that Intel pulled ahead around 2000. This again shows the PPC has been ahead most of the time and it was just a short span of time during early 2000s before the G5 came out that the G4 wasn't competing. With the G5, PPC has pulled ahead again and even with the CoreDuo the G5 remains ahead (hence why no Intel PowerMac). Current independent benchmarks confirm this.

I agree that PPC was ahead for much of its lifespan; it's an excellent ISA and had some excellent people working on it. However, I have issues with the statement that the G5 is ahead again. CoreDuo is not a good chip to compare to, as it's intended as a laptop chip. Intel currently lacks a chip really suitable for the PowerMacs, but that doesn't mean that its high end desktop chips aren't as fast or faster than the G5. Faster G5s in the iMac would have been as fast or faster than switching to CoreDuo, but I believe they would have run into heat issues. The iMac is oddly laptop-like.


You also mention cost, mentioning the Xeon system is much more expensive. Also of note is that the CoreDuo is also a more expensive chip than the G5. Apple is having to go with a more expensive cpu for the iMac and Pro laptop in order to garner just a 10% improvement in real world benchmarks. A bump up to a faster G5 probably would have provided a similar improvement in speed (and perhaps at lower cost).

I said the CPU was more expensive, not the system. Switching to Intel allows Apple to get rid of its highly custom (and presumably very expensive) northbridge design. Whether the net cost ends up being cheaper or more expensive, you can't just compare $$$/cpu to figure it out.


PPC technology is simply hands down superior technology to x86, and if you think different you are fooling yourself.


PPC is not a technology, it's an instruction set. The actual technologies behind PPC chips vary in how advanced they are. Freescale is *clearly* behind the curve, having only just moved to 90nm manufacturing fairly recently. IBM is more up to date, but appears to pay a heavy cost in manufacturing reliability on their bleeding edge process. Design-wise PPC has the major advantages of Altivec and fused multiply-add instructions, and a number of smaller advantages (easier instruction decoding, for example). The question is whether Intel's engineers and manufacturing are *enough* better to outweigh the downsides of the x86 ISA. I think that has been demonstrated quite well, although the somewhat disastrous Prescott Pentium-4 design muddies the issue somewhat. Hopefully they've learned their lesson.


The switch to Intel is probably a very good move, but lets not confuse people by trying to pretend the benefits of the switch are technological. Its as I said before, the switch is a good idea because developers of PPC are no longer focused on creating CPUs for personal computing. While the PowerMac remains competive Apple has not been provided good upgrade options for the laptops or low end desktops. Intel chips provide good cometitive products across the spectrum and also provides a more reliable source of chip (since Intel is very much interested in creating cpus for personal computers).

Agreed with everything except the first sentence. Core Duo is not a major upgrade over the G5 (yet), but its a fantastic replacement for the G4 in a way that the G5 couldn't manage. Even with the G5, reports of the new iMacs being quieter lead me to believe that there have been significant savings in heat and power.

Basically, for the PowerMacs, we're waiting for Conroe in the second half of this year. Then do the comparison again. Either Intel has been BSing us (a definite possibility) or it's going to be an awesome chip.

macOSX-tastic
Jan 22, 2006, 05:32 PM
i also think the intel chip has significant potential. i just want to wait a few revisions before i even consider it a likely candidate replacement for my current machine, which i expect to last a few years yet.

definitely the way forward.

janstett
Jan 22, 2006, 06:04 PM
Technically speaking the PPC chips still are much better than any Intel chips (even the CoreDuo). The old PC fan boys around here (switchers) will try to argue that the Intel chips were always faster and point to things like spec. Truth is Intel chips cheat at these benchmarks requiring you to look at real world performance to assertain true speed. And in nearly all real world benchmarks PPC wins over Intel (AMD is another story).

Sigh... I suppose that comment is directed at me. I'm not a switcher, I love computers and have since the Carter administration (my first was the TRS-80). I've learned to let go of past hatred for Apple and just enjoy the alternate platform the Mac provides, along with Linux and Windows. I don't care to be a zealot on either side of the issue anymore.

Let me translate for those who haven't participated in the OS wars. "Real World Performance" = photoshop, and specifically hand-picked tests that showed Macs to be faster than Windows at the same tasks. Apple always cherry-picked bizarre tests that showed them to be faster, when they lost on objective tests that work on ALL processors, i.e. Spec2000.

If the PowerPC were so inherently superior, it's pretty clear that Apple wouldn't be following the path it is now. Just to start with their cost for the PPC itself was a fraction of what they're paying Intel. So if they're really cheaper AND faster, what the heck are they switching for? I mean, what more do I have to say... Apple switched to Intel Processors, the entire line will be switched by the end of 2006. Do I have to go further and bury the point into the ground?

As to performance, the Intel Macs already do certain things faster than the PPCs and wait until we see native apps. But what's the point, PPC is dead unless you want to buy a game console.


PPC chips almost always have been faster than the best offerings from Intel (including the current moment in time). The latest battles have mostly been between AMD chips and PPC chips with various processors released by each camp leap frogging the competition. The reason the PowerMac isn't Intel yet is that Intel doesn't have a chip that would match the current performance of the PowerMacs, let alone anything that would constitute an upgrade.


Or perhaps that switching the entire line over in one fell swoop is a herculean task (see the MacBook Pro's uncreative design), and Apple is waiting for Intel's next generation of processors for the desktop. Not to mention stalling for more native application support.

I mean, we can test this now once we get past the EFI issue. Tell you what, once someone cracks this issue I'll install OSX86 10.4.4 on my dual core hyperthreading P4 3.6 GHz and run some native benchmarks, we'll compare it against the best dual core G5 and let's see which is really faster.

I agree with your observations at the end of the post. Intel is a hard-working supplier that listens to its customers. After all, I'm sure Apple negotiated with AMD as well and in the end went with the "evil empire".

Catfish_Man
Jan 23, 2006, 01:47 AM
But what's the point, PPC is dead unless you want to buy a game console.

I'm guessing you're considering POWER and PowerPC to be different? IBM doesn't really do that anymore, afaik, so I'd say you can add servers (and possibly reaaaally high end workstations) to the list. and high end networking gear, and cars, and a ton of other really small stuff. There's just this consumer-computer sized hole in the middle of the line that no one seems interested in making chips for (Both the G4 and the G5 are adapted from other markets). Kind of a pity, really.

Naimfan
Jan 23, 2006, 10:42 AM
I have to say that your explanation of the differences among chips was terrific, notwithstanding the nitpicking by some folks. Thanks for posting it!

Best,

Bob

EHUnlucky7x9@ao
Jan 23, 2006, 11:30 AM
Wow, this is all fascinating. I'm surprised at the length of the comments left behind. I guess there isn't truly a simple, short and sweet reason. So now I know it varies upon the different uses that I will have with my computer. Since I'm sticking to video editing, I guess the PPC is the way to go. If not, it doesn't matter because my machine has enough juice to use Final Cut + DVD SP.

Isn't it best to wait to see how the Intel Macs are before buying them. Look at the iMac G5...they keep changing/ adding features to it. I think it's like only a few months between the revisions too... so maybe I should wait for a newer MacBook Pro than the ones they have now. Is there anything really that CAN be added as a feature to the laptop down the road? Just faster processors and faster VRAM right?

Arcus
Jan 23, 2006, 11:45 AM
I just wanna surf porn faster. Which one is for me..


j/k

notjustjay
Jan 23, 2006, 11:48 AM
You bought into Steve's reality distortion field.

(Sorry, a bit OT)

Joe Anstett?

Were you at one time an alt.video.dvd regular? (Back in the days of DIVX etc.)?

jacobj
Jan 23, 2006, 12:08 PM
Well the G5 is a PowerPC processor. It is being abandoned. We are switching to x86 processors made by Intel. The CoreDuo (cpu in the new MacBook Pro and iMac) IS a mobile chip. So in essence it is the mobile G5 that IBM had no interest in creating.

I said a few times the reason the PowerMac wasn't updated was that its still faster than any Intel chip. The other end of that question is why the iMac and Macbook Pro then? The answer would be that these are the best selling Macs. No doubt Apple wanted to move its most popular lines to Intel to send a strong message to developers that its time to get those Universal Binaries out. I think thats also why these machines (great machines don't get me wrong) seem a bit half-baked by Apple standards. They are a warning shot. Priming the development environment for Intel Macs.

AmbitiousLemon, you have argued that the move wasn't a technical one, but I beg to differ - slightly.

I agree that the PPC desktops are fast machines. I don't care about benchmarks, just using one is proof enough for me. But the laptop market is growing faster than the desktop market and as such Apple needed to ensure that they could compete. That was never going to happen with the PPC. Apple couldn't switch laptops to x86 and leave the desktops on PPC as the developers would never play ball.

If I were buying a desktop now I would still buy a PowerMac G5. But I'm not. I need a laptop and will probably continue to use laptops exclusively for many years to come. So for me the intel switch is a good one and purely a technical one at that.

disconap
Jan 23, 2006, 12:26 PM
So, having read most of this thread (and learned quite a bit, I was under the impression that the x86 chips were just not supported in the OS much yet, and that was the main reason that the PPC chipsets were still faster for many applications), what I'd like to know is is Apple abandoning the PPC altogether, or will future systems incorporate PPC chips, at least in the PowerMac line?

And have they abandoned the 64-bit angle entirely?

AmbitiousLemon
Jan 23, 2006, 02:41 PM
Good article taht sheds some light ont his topic here (http://www.reghardware.co.uk/2006/01/23/intel_macs_25pc_faster/)


Don't say we didn't warn you. But when the world's last great computer company decided to tie its fortunes to the world's slowest chip company, the reality was never going to match the hype.

...

Only once in the past two decades has Intel been able to claim the performance crown, very briefly in late 1995 when its Pentium Pro knocked DEC's Alpha chip off the top of the benchmarks. On desktop performance alone, Intel has been bested for several years by AMD's far more competitive Athlon chip. Intel's next generation 64bit processor Itanium is a billion dollar dud, and it failed to crank much advantage out of the deep pipelined P4, which always ran hotter, and more inefficiently, than generations of Athlon or RISC processors. So last year Intel finally tore up its roadmaps, abandoning its Athlon-killer P7 core for future desktops, and leaving us to look forward to derivatives of third generation mobile chips. These will be powering Microsoft PCs - and now Apple computers, too - for the next few years.

When Microsoft chose a next generation chip for its Xbox 360 console - something expected to have a life of five years - it chose a dual core PowerPC processor, the platform Apple was abandoning.

...

So Intel makes a lot of chips, but they're never the best. Tell us something new, you're thinking.

...

Why did Apple move to Intel, then, really?

Intel justifiably remains one of the most lauded companies on the planet not for the quality of its chips, but for its consistent innovation in production. It's a manufacturing company first and foremost, and its R&D is geared towards keeping its facilities full.

What falls off the end of the Intel production doesn't really matter.


I don't agree with their final analysis (basically saying Apple did this because they don't care about computers anymore), but the facts of their story quoted above support what I had been describing earlier int his thread.

janstett
Jan 23, 2006, 03:46 PM
(Sorry, a bit OT)

Joe Anstett?

Were you at one time an alt.video.dvd regular? (Back in the days of DIVX etc.)?

Yep, same guy... I'm sitting out the alt.video.bluray and alt.video.hddvd forums :)

janstett
Jan 23, 2006, 03:50 PM
Apple couldn't switch laptops to x86 and leave the desktops on PPC as the developers would never play ball.


Actually, says who? I was going to make this argument but you inadvertently beat me to it. If the G5 were so superior on the desktop, why couldn't Apple adopt Intel for the laptops/iMacs and stay with the G5 on the desktop?

After all, applications from this point on are going to be Universal (both PPC and Intel binaries). It won't matter what you have "under the hood".

So, if the processor inside doesn't matter with new Universal binaries, I ask again why Apple doesn't keep the G5 on the desktop?

AmbitiousLemon
Jan 23, 2006, 03:55 PM
because its not that simple. Universal binaries are a temporary thing. eventually (probably sooner than most think) apps will be intel only. And long before that happens applications will be optimized for the intel platform.

Universal binaries are a transition technology, not something permanent.

NeuronBasher
Jan 23, 2006, 04:36 PM
Developers do not want to have to optimize for two architectures, and this particularly hits those types of apps that benefit from Altivec or SSE. The first round of Universal Binaries that are released for the pro apps are probably going to be much less than optimal since the SSE support will be immature.

The Adobe products might be an exception to this since they may be able to leverage some of the work that they've done for the Windows ports, but the apps that are Mac-only are going to take more than a little while to be completely optimized on the new platform. That doesn't mean they won't be quite performant, just that we can expect the performance to improve "for free" as later versions are released taking more advantage of the processor features of the Intel.

SSE is no Altivec, but it's still better than nothing, and it'll be nice once it's being taken advantage of everywhere.

Catfish_Man
Jan 23, 2006, 04:48 PM
Good article taht sheds some light ont his topic here (http://www.reghardware.co.uk/2006/01/23/intel_macs_25pc_faster/)



I don't agree with their final analysis (basically saying Apple did this because they don't care about computers anymore), but the facts of their story quoted above support what I had been describing earlier int his thread.

FUD. Alpha competed with PA, SPARC, and MIPS, and then at the end of its life it halfheartedly competed against POWER and Itanium. Comparing it to desktop* chips is just as silly as comparing POWER5 to desktop chips. The Pentium 4, Intel's biggest desktop-market flop in quite a while, spent much of the middle of its lifetime running slightly faster than anything AMD produced (pretty much the whole time the Northwood core was in production). It only really lost traction with the double blow of the Athlon64 and the remarkably deluded Prescott core. The replacement for Prescott appears to be a return to sanity, with a pipeline that's less than half as long.

*I'm using desktop here to refer to laptops, desktops, and the low end of workstations.

<edit>
NeuronBasher: many Altivec-ized apps will take advantage of SSE immediately, as they use Apple's Accelerate framework rather than using Altivec directly. It's true, though, that the ones that using Altivec directly will take a performance hit.
</edit>

bigfib
Jan 24, 2006, 04:41 AM
Did anyone wonder *why* ibm has failed to deliver the agreed (with apple) roadmap??
Could it be because Microsoft is now a far bigger customer of G5 chips (for the xbox) than apple??
Maybe steve walked out because IBM has a new organ grinder.... Gates...

andrewag
Jan 24, 2006, 05:42 AM
The advantage to Apple moving to Intel isn't speed or power per watt (as the marketing guys would like you to think). The main reason for the switch is that IBM embarassed Steve - by not delivering on the promise of a mobile G5 and a 3Ghz G5.


Finally, someone who gets it! :D

bugfaceuk
Jan 24, 2006, 06:03 AM
Well the G5 is a PowerPC processor. It is being abandoned. We are switching to x86 processors made by Intel. The CoreDuo (cpu in the new MacBook Pro and iMac) IS a mobile chip. So in essence it is the mobile G5 that IBM had no interest in creating.

I said a few times the reason the PowerMac wasn't updated was that its still faster than any Intel chip. The other end of that question is why the iMac and Macbook Pro then? The answer would be that these are the best selling Macs. No doubt Apple wanted to move its most popular lines to Intel to send a strong message to developers that its time to get those Universal Binaries out. I think thats also why these machines (great machines don't get me wrong) seem a bit half-baked by Apple standards. They are a warning shot. Priming the development environment for Intel Macs.

That's an interesting perspective on the reasons for chosing the iMac and MacBook. I think perhaps it is more relevant to the timing than the model selection. However, still a very interesting point.

As for the various processor architectures... all I know is my Mac Mini does more than it should be able to when you compare it to my 3200+ AMD 64 with 2gigs of super fast memory. All but direct number crunching.... the mini wins.

I'm hoping that's more OS than CPU....

macker
Jan 24, 2006, 06:11 AM
Check this out:

http://www.reghardware.co.uk/2006/01/23/intel_macs_25pc_faster/

It doesn't make good readinh.

Catfish_Man
Jan 24, 2006, 06:46 AM
Did anyone wonder *why* ibm has failed to deliver the agreed (with apple) roadmap??
Could it be because Microsoft is now a far bigger customer of G5 chips (for the xbox) than apple??
Maybe steve walked out because IBM has a new organ grinder.... Gates...

The Xbox doesn't use the G5. A triple-core G5 would be far far too power hungry for it. It uses a three core CELL variant called Xenon.

My theory on why IBM has failed to deliver is that Apple has failed to fund (as well as the fact that the 90nm transition was just brutal for the whole industry). A high end chip line is *expensive* to design and produce, and Apple's sales volumes just aren't big enough to justify significant investment. The G5 has done fairly well by piggybacking off the design work for the POWER4, but that doesn't help much when aiming at laptops.

janstett
Jan 24, 2006, 07:00 AM
Developers do not want to have to optimize for two architectures, and this particularly hits those types of apps that benefit from Altivec or SSE. The first round of Universal Binaries that are released for the pro apps are probably going to be much less than optimal since the SSE support will be immature.

I don't think that's a valid excuse.

Apple *could* pursue a dual pronged strategy -- Intel in notebooks and iMacs, G5 on the desktop. Since the G5 is meant for the high end users, pros, and developers, target the G5 optimizations as a priority and let the low-end users on Intel "suffer" with unoptimized (or less-optimized) code. (Besides, the compilers do a good deal of optimization for each platform).

After all, "pros" aren't supposed to rely on notebooks or iMacs, right? Portable users can't expect desktop performance anyway.

The fact that they aren't doing it means something. It may be that neither Apple, IBM, or Intel wanted half a loaf or a mixed line of product. It may be that Apple told IBM to piss off despite the fact that they'd be better off with desktop G5s.

NeuronBasher
Jan 24, 2006, 07:15 AM
(Besides, the compilers do a good deal of optimization for each platform).

The compilers do a pretty good job of optimization, but I honestly don't know if they do any automatic optimzation for the Altivec or SSE engines.

After all, "pros" aren't supposed to rely on notebooks or iMacs, right?

There's where your premise goes awry. The MacBook Pro has Pro in it's name for a reason. It's meant to be a portable workstation for "pro" users.

With that said, I wasn't aware of the Apple Accelerate framework since my programming is at a higher level than that. With an extra layer of abstraction built in, it means that the developers only have to code against the Accelerate libraries and trust Apple to do the right thing with optimizations. There will be some hand-coded SSE and Altivec code, certainly, but hopefully it will be far less than I had originally assumed.

Catfish_Man
Jan 24, 2006, 02:37 PM
The compilers do a pretty good job of optimization, but I honestly don't know if they do any automatic optimzation for the Altivec or SSE engines.

ICC, XlC, and GCC4+ will autovectorize, but it's pretty primitive compared to what a competent human can do.

Krevnik
Jan 24, 2006, 03:47 PM
The Xbox doesn't use the G5. A triple-core G5 would be far far too power hungry for it. It uses a three core CELL variant called Xenon.

My theory on why IBM has failed to deliver is that Apple has failed to fund (as well as the fact that the 90nm transition was just brutal for the whole industry). A high end chip line is *expensive* to design and produce, and Apple's sales volumes just aren't big enough to justify significant investment. The G5 has done fairly well by piggybacking off the design work for the POWER4, but that doesn't help much when aiming at laptops.

Uh, not quite here... The Xenon is a tri-core 970 variant. It loses OoO execution, and some of the more complex parts of the branch prediction, but gains in pipeline width for the VMX unit (more vector processing units on each core), gains SMT and the hypervisor.

Cell uses a single 970-derived core with a lot of the same tweaks (losing the complex branching and OoO, while gaining the SMT), but seems to lack the hypervisor and the VMX additions. The rest of the SPEs on the die are completely custom pieces of work using a FlexIO bus out to IIRC, an onboard DMA controller which managed each core's communication with the outside world and through their cache.

The only reason why Xenon was able to handle three cores is because they removed an awful lot of stuff from the core design which takes a fair amount of space, and they consolidated the entire cache into a single 1MB shared cache. The 970 dual-core design allocates 1MB to each die, which accounts for a chunk of space. By dropping down to 1MB and sharing it, they almost get enough room for a third core from that alone.

I would like to mention that the 360 /is/ very power hungry, drawing peak wattage of about 200W.

Catfish_Man
Jan 24, 2006, 04:44 PM
Uh, not quite here... The Xenon is a tri-core 970 variant. It loses OoO execution, and some of the more complex parts of the branch prediction, but gains in pipeline width for the VMX unit (more vector processing units on each core), gains SMT and the hypervisor.

Cell uses a single 970-derived core with a lot of the same tweaks (losing the complex branching and OoO, while gaining the SMT), but seems to lack the hypervisor and the VMX additions. The rest of the SPEs on the die are completely custom pieces of work using a FlexIO bus out to IIRC, an onboard DMA controller which managed each core's communication with the outside world and through their cache.

The only reason why Xenon was able to handle three cores is because they removed an awful lot of stuff from the core design which takes a fair amount of space, and they consolidated the entire cache into a single 1MB shared cache. The 970 dual-core design allocates 1MB to each die, which accounts for a chunk of space. By dropping down to 1MB and sharing it, they almost get enough room for a third core from that alone.

I would like to mention that the 360 /is/ very power hungry, drawing peak wattage of about 200W.

The CELL PPE is in no way 970 derived, except possibly the altivec unit. Going from 5 issue out-of-order to 2 issue in-order with SMT is more than enough to qualify as "completely different" in my book. More different than the 604e and 970, certainly, which most people consider unrelated. The PPE
also has a different pipeline length, and makes extensive use of custom-designed logic, while the 970 is built primarily using IBM's automated layout tools. realworldtech.com has a series of excellent articles on the subject; well worth reading for those interested. http://www.research.ibm.com/arl/projects/rivina.html is also likely related.

I admit I'm not 100% sure on how similar Xenon is to the CELL's PPE (the cache structure is quite different, as you noted), but it's definitely quite close. I would dig up some references to back that up, but it'll have to wait until I'm on a more reliable net connection.