I'd say close to 0.1%
Someone's forgotten about high K transistor technology
If not next year, then the year after.
4+GHz Quad core MBPs!
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.
Did Apple Do The Right Thing in Moving To Intel From PowerPC???
- Thread starter rashdown_online
- Start date
- Sort by reaction score
Someone's forgotten about high K transistor technology
If not next year, then the year after.
4+GHz Quad core MBPs!
If nothing else, then for marketing
I don't know if anyone has mentioned this point yet, but I think Apple's move to Intel chips is important and correct for one reason - marketing. Whether there is or isn't a megahertz myth, whether altivec is or isn't better than SSE, whether Motorola did a better job with graphics or not are all non-issues from a marketing perspective. The vast majority of potential customers don't understand those arguments and more importantly aren't affected by the answer. Because Apple can now tell the majority of the computing universe exactly how they compare to your average Dell eliminates the whole requirement to justify the hardware. Now, Apple can focus on what truly makes it great - software! Its marketing machine can accentuate all of the wonderful things that make the Apple experience unique. The "I'm a PC, I'm a Mac" ads make this point specifically. For the most part, these ads don't compare hardware. They focus on usability and software.
Intel may not provide the absolute best platform for graphics or sound, but Apple's switch to Intel satisfies most user's requirements and allows Apple to market software without customers asking, "Yeah but..."
I don't know if anyone has mentioned this point yet, but I think Apple's move to Intel chips is important and correct for one reason - marketing. Whether there is or isn't a megahertz myth, whether altivec is or isn't better than SSE, whether Motorola did a better job with graphics or not are all non-issues from a marketing perspective. The vast majority of potential customers don't understand those arguments and more importantly aren't affected by the answer. Because Apple can now tell the majority of the computing universe exactly how they compare to your average Dell eliminates the whole requirement to justify the hardware. Now, Apple can focus on what truly makes it great - software! Its marketing machine can accentuate all of the wonderful things that make the Apple experience unique. The "I'm a PC, I'm a Mac" ads make this point specifically. For the most part, these ads don't compare hardware. They focus on usability and software.
Intel may not provide the absolute best platform for graphics or sound, but Apple's switch to Intel satisfies most user's requirements and allows Apple to market software without customers asking, "Yeah but..."
It's impossible to know for sure what the development of PPC would have been had Apple stuck with it. For all the claims of low-power G5s being unavailable and other similar wisdom, we don't know what would have been made had there been a buyer for them.
I'm disappointed with the switch to Intel. I remember the wonderful late eighties/early nineties where hardware manufacturers competed to produce the most elegant systems, not just elegant cases. Today it's the same cocktail of cutting-edge buses and hack-upon-hack legacy-compatible designs on the inside, with only the case and, with Apple, operating system, delineating the Dells, HPs, and Apples.
And I can't see how anyone can argue that's anything other than a crying shame.
I'm disappointed with the switch to Intel. I remember the wonderful late eighties/early nineties where hardware manufacturers competed to produce the most elegant systems, not just elegant cases. Today it's the same cocktail of cutting-edge buses and hack-upon-hack legacy-compatible designs on the inside, with only the case and, with Apple, operating system, delineating the Dells, HPs, and Apples.
And I can't see how anyone can argue that's anything other than a crying shame.
Because it didn't make business sense. Doesn't matter how elegant the solution if you can't make money on it. Steve learned that lesson at NeXT.
Yep, it's a shame the Macintosh wasn't profitable back in 2005. And, of course, it still bewilders everyone that a computer priced at $10,000 (1988 dollars) and sold only to academics wouldn't achieve mass market sales.
It's also a shame a comment saying it's a shame there's a lack of choice can engender an immediate apologia claiming that a lack of choice is a good thing because it makes (good) "business sense". I'm not sure what to make of that. The two things have little to do with one another. Does the fact it makes good business sense mean it's not a crying shame? Of course not.
It's depressing a loss of choice would be so enthusiastically loved, with 90% of voters in this poll apparently excited by it. Just think how great it'll be once Apple switches to Vista!
What Steve learnt from his NeXT experiences was to not let another company exclude him from a profitable market. The settlement with Apple forced NeXT out of the desktop market and into the workstation market... a market that was already starting to shrink due to the increasing abilities of desktop systems.
Basically, Steve learnt that it is one thing to be late to a party, it is quite another to show up as all the guests are leaving.
Why is it that people can't look at the actual facts... while the first cubes may have been $10,000, that wasn't where the price of NeXT systems stayed. Lets (once again) look at the actual prices of some NeXT systems... comparing the prices between machines from NeXT and Apple (with comparable features) from 1991:peharri said:
NeXTstation (68040 at 25MHz, 8 MB of RAM, 105 MB hard drive, 2 bit (black & white) 17" display, Ethernet) $4,995.00
Macintosh IIsi (68030 at 20MHz, 5 MB of RAM, 80 MB hard drive, 8 bit 12" display, LocalTalk) $5,097.00
NeXTstation Color (68040 at 25MHz, 12 MB of RAM, 105 MB hard drive, 16 bit (color) 17" display, Ethernet) $7,995.00
Macintosh IIci (68030 at 20MHz, 4 MB of RAM, 80 MB hard drive, 8 bit 13" display, LocalTalk) $7,897.00
You would have had to have bought a '040 based Cube with a Dimension graphics board to push the price of a NeXT above $10,000... and there wasn't much on the market for under $20,000 that could match it's performance.Macintosh IIsi (68030 at 20MHz, 5 MB of RAM, 80 MB hard drive, 8 bit 12" display, LocalTalk) $5,097.00
NeXTstation Color (68040 at 25MHz, 12 MB of RAM, 105 MB hard drive, 16 bit (color) 17" display, Ethernet) $7,995.00
Macintosh IIci (68030 at 20MHz, 4 MB of RAM, 80 MB hard drive, 8 bit 13" display, LocalTalk) $7,897.00
By the time the NeXTstation Turbos were released (with 68040 processors at 33 MHz), these systems were on par (performance wise) with similar products by Silicon Graphics (the IRIS Indigo workstations using R3000 processors at $7995), Sun (the SPARCstation IPX at $13,495) and Apple (the Quadra 900/950 at $8500). Both Silicon Graphics and Sun had better brand recognition in the workstation market and as pointed out, NeXT couldn't compete directly with Apple in the desktop market.
If you guys are going to argue over this stuff, the least one of you could do is argue based on the actual history of these events.
It's not entirely irrelevant, particularly when you realize my point was that I happen to like the PowerPC architecture. Performance-wise though, the cleanliness of a design indicates what its future growth prospects are. When you compare the amount of resources Intel applies to keeping x86 current with the amount of resources IBM applies to Power, you see the benefits of elegant design.
Much like I'm not impressed with what billions of dollars have bought Microsoft, I'm not impressed with what billions of dollars have bought Intel.
Didn't get NetBurst powerbooks either... Nor did we get Core powerbooks with a 1.5GHz memory bus. As I've mentioned, the G5 wasn't a mobile CPU. That doesn't mean that PowerPC wouldn't have been appropriate. G4 was quite efficient, but was limited by the memory bus at the time. A multi-core G4 would have been quite competitive, methinks.
Depends on what your usage model was. G5 was faster at multimedia apps because Altivec outstripped SSE, for example.
Then why did the Mini pricing go up by more than that? The rest of the design was essentially the same aside from the fact that the Intel mini saved itself a separate GPU. The "Intel is cheaper" argument is quite misinformed-- just the die sizes make that clear, and product pricing supports it.
Why would it suck for general purpose computing? It hasn't even been attempted... Nobody's bothered to port a mainstream OS to it. Even the PowerPC portion alone would do quite well at over 3GHz. The architecture itself is also much more scalable than anything else that is shipping at those volumes.
The complexity in coding for it is why we won't see it on the desktop for a little while. Heterogeneous cores is the future though. Once the compilers catch up expect to see this kind of design used everywhere. Rather than bundle all of your execution units with a single control pipeline, it is much more efficient to break them out individually and not waste all that idle silicon.
The point is that Cell is advancing the state of the art. That's what I mean by R&D leader. If you look at Intel's contributions, they're all in semiconductor processes. They aren't the R&D leader in CPUs that some have claimed. That's my point. Cell is one example. ARM was far beyond what Intel could do at the time in the portable space which is why Intel licensed it for their own use. Itanic is based on an HP architecture. Core borrows liberally from the Alpha IP that Intel now owns.
Everyone is under the misconception that the G5 is the Power5 but its actually something entirely different. The G5 is a PPC 970 cpu which is based off the POWER4 cpu. So even if the POWER6 is faster than C2D and all it would never see the light of day on a mac.
I've always disliked intel processors, even now the G5 is still the only TRUE 64bit computer out there. (intel and amd use an archatecture that is only semi-64 bit)
But, as for getting a POWER series based CPU in a mac? nah, I'll pass...
Cell 2...now thats what I want in my mac (a dual cell 2) or.... The D-wave 16 qubit QPU (quantum processing unit) if you recall along with the annoucnment of the POWER6 series, IBM showed off the next-gen version of the cell processor clocked @ 6Ghz.
instead of measuring computer performance using proprietary stress tests why not just measure in terms of Gigaflops? its universal, its very exact, and easy to tell which is faster (heres a hint, the bigger the number the better)
But, as for getting a POWER series based CPU in a mac? nah, I'll pass...
Cell 2...now thats what I want in my mac (a dual cell 2) or.... The D-wave 16 qubit QPU (quantum processing unit) if you recall along with the annoucnment of the POWER6 series, IBM showed off the next-gen version of the cell processor clocked @ 6Ghz.
instead of measuring computer performance using proprietary stress tests why not just measure in terms of Gigaflops? its universal, its very exact, and easy to tell which is faster (heres a hint, the bigger the number the better)
People also forget that even though OS X is theoretically cross-platform and the individual apps could be cross-compiled, maintaining another set of device drivers is a big challenge. Different chipsets, different devices, etc. There's a lot of "invisible" software there that has to be rewritten and maintained, not just by Apple but by third-parties also.
Then why didn't those growth-prospects materialise? It's not good if something COULD happen if it never does. Hey, I COULD win in the lottery.
Um, IBM is am massive semiconductor-company. It's not like Intel spends billions while IBM spend peanuts.
Does NetBurst indicate that x86 is crappy?
There has neven been a Mac or PC with a 1.5GHz memroy-bus. Fastest Memory-bus in the last PowerMac G5 ran at 533Mhz. Are you by any change confusing memory-bus with front-side bus?
Multi-core G4 would have been utterly starved for bandwidth.
Depens on the app, not all apps took good advantage of Altivec.
Because they increased it's capabilities? Because it got twice as many USB-ports as before? Because it got faster CPU? Because it moved to SO-DIMM instead of standard RAM? Because it got two memory-slots as opposed to just one? Because it got a remote? Because it got new multimedia-software?
In case you didn't know: the internals of the Mini got a total makeover when they moved to Intel. The CPU is only a part of the equation.
Um, no it wasn't.
Faster CPU's usually are more expensive that slower CPU's. Had they used as slow CPU as the G4 was in the Mini, they could have shaved maybe 20 bucks off the new Intel-price. But the CPU's they ended up using weren't really that much more expensive. 30 bucks maybe at the low-end.
Cell is an absolute screamer when it comes to streaming computation and single-precision floating-point. And that's the kind of stuff consoles need. General-purpose computing relies heavily on integer, not floating point. Sure, you can run a general-purpose OS and apps on it (as demonstrated by the PS3 Linux-kit) and they would run fine. But the performance would be quite mediocre. You could get a better performance from a CPU that costs a fraction of what the Cell costs.
Yes it has.
Linux runs fine on it.
It's an in-order CPU. It would run OK-ish, but it would get spanked by just about any other general-purpose CPU that cost half as much.
It was jointly designed by HP and Intel, but PA-RISC is quite different than Itanium.
Core borrows heavily from Pentium 3. If you want a CPU that has it's roots in the Alpha, AMD's Opteron/Athlon64 would be one contender. And there's nothing wrong with that, Alpha was a kick-ass CPU.
I totally agree that laptops were the driving factor behind the switch to Intel. And it was a good move.
I'd like to say that this thread was great reading. Too bad some posters didn't read preceding posts before writing their own. If you're going to post, you have to read everything else first.
BTW Cassie, you rock.
I'd like to say that this thread was great reading. Too bad some posters didn't read preceding posts before writing their own. If you're going to post, you have to read everything else first.
BTW Cassie, you rock.
regarding laptops, yes.
But those [who use workstation class computers and] whose primary software relies heavily on Altivec, the move is far from good. Single G5 core had two Altivec units and because even one unit is far more powerful than Intel's the difference is *huge* in favor of G5 systems.
On the short-term, maybe. But once that Altivec-optimized code becomes SSE-optimized, things will change considerably.
Altivec used to be faster. That was due to Altivec being 128bit, whereas SSE was 64bit. But these days SSE is 128bit as well. And SSE was always better at 64bit arithmetic than Altivec was, Altivec shined at 128bit arithmetic and 32bit arithmetic. But these days even that benefit has all but vanished.
If G5 is SOOOO much better, why aren't G5-systems walking all over Intel-systems? Could you provide some tangible benchmarks that show this "huge" benefit G5 has over Intel-CPU's?
You honestly think the OLPC program can stand up to more modern CPUs? Core 2 Duo would eat those things alive. Sure low end computer do things, just not as much or fast as higher end computer.
No they wouldn't, but new ones would slow down, and one day stop because devs would have maxed out all power on the computer. Plus if your idea would have happened 20 years ago, we would have less apps...new app wil appear as more power is there to be used. Stopping this increase of power would result in less apps.
Wow...why? Your computers do everything you want right? So why upgrade them? Sure other people will be able to do more stuff faster, but your computers do everything you want. And since you want to stop CPU increase, why would you need to update your system?
Good for you!
I really don't think computer that you got when you were 5 had much effect on you. I don't think you did much on them.....surely didn't try and fix the problem on them. And how many files did you need to take from them to your next computer?
So you had 1 computer die on you, and you bought a new machine, and moving files was hard, so you want the whole computer world to change? You have had 2 main computer(when you we're old enough to run them yourself), and you think moving things around is too hard?
My advice, use target firewire disk mode when you get your next Mac, and make it a high end Mac, so it will last for a while, if upgrade is the much of a problem for you.
I suggest rethinking them
Please back up your conjecture.....
That will be news to all of the SPARC, Alpha, PA-RISC, POWER, MIPS, and IA64 users out there, not to mention the Cray and NEC and other true 64-bit systems.
Can you please explain this statement, both as to how the x64 architecture is "semi 64-bit" and the particular ways that this alleged "semi 64-bit" discrepancy gives the G5 an advantage over Core 2 Duo or Xeon running with a 64-bit operating system?
I think that this is FUD - there's no difference in 64-bitness between x64 and other true 64-bit CPUs.
That will be news to all of the SPARC, Alpha, PA-RISC, POWER, MIPS, and IA64 users out there, not to mention the Cray and NEC and other true 64-bit systems.
Can you please explain this statement, both as to how the x64 architecture is "semi 64-bit" and the particular ways that this alleged "semi 64-bit" discrepancy gives the G5 an advantage over Core 2 Duo or Xeon running with a 64-bit operating system?
I think that this is FUD - there's no difference in 64-bitness between x64 and other true 64-bit CPUs.
Intel is the right processor. Macbook/Pro is not the right computer. They need a hardware re design.
well, sparc and others have 64bit address-space, whereas x86-64 "only" has 48bit. although that limitation is 99.9% theoretical, and i'm not sure that is g5 better in that regard
What they mean is there is currently no system for sale that offers 64-bit memory addressing. While "64-bit" removes the "32-bit" limitations, there are not yet any "true 64-bit" implementations available for consumers — not that the maxx available memory would be a bottleneck (for consumers), but if the architechture isn't "fully optimised" and "fully loaded", one can always whine...
For now, I'm perfectly happy with 1GB RAM per CPU core, which means 4GB RAM in Quad PowerMac and 2GB RAM in a MacBook Pro. This gig-per-core rule has been in use since 2002 and it has been working very well.
Core is actually derived from what some people think is Intel's best cpu, the Pentium Pro. Its why they say the intel went back to its roots when they designed the core architecture.
Till when????
How much more can Apple grow without running into problems????
Will Apple license 3rd parties to make Machines, yes, clones?
Does Apple want to be in Microsoft position regarding market share?I don't think so, but what if things get out of hand and people start buying Macs like crazy, well like they are doing.
I've seen people that don't even know what a Pc is yet they are getting Macs cause they look cool.
Apple has to be very careful not to swallow itself
For old apps, perhaps... maybe. But for any new audio apps written to take advantage of SSE3 + anything PPC less than a quad-G5 come up short of the Core 2 Duos according to Steinberg's chart...