It looks like it is now official, folks.
fyi
It looks like it is now official, folks.
It's official, their CEO just announced it, running the presentation off of an Intel PC.
http://www.engadget.com/entry/1234000137045772/
Does this mean IBM will be freed to sell its Power PC chips to clone manufacturers who will again be able to offer budget legacy Apple compatibles?
I guess Mac heads will have to stop touting the superiority of the PowerPC platform over x86...
Wonder if Apple is going to dump the dual CPU computers now?
ping
Apple better make this happen quickly. No one wants to be the last buyer of a PowerPC-based Mac. Fewer people will by a PowerPC Mac knowing that a faster, cheaper Pentium version is on the horizon.
Ouch!
This shouldn't be a big surprise - Jobs did the same thing when he ran NeXT in the 80's/90's. Went from Motorola to Intel CPUs. Yawn.
1. Expensive hardware - looks like that is being taken care of.
2. Proprietary hardware - ditto
3. Lack of software titles - which brings me to my main point.
If Apple wants to put a world of hurt on Mircrosoft, they would make their machines dual boot-able. They could even build an advertising campaign around it. Present it like "We know you still HAVE to use a Windows-based computer for SOME things, but for everything else..."
They could even tout how when spyware slows down your Mac running Windows, just start the Apple Operating System and get back up to speed...
Since they now have a reputation for quality peripherals, they could easily compete in that arena and take even more market share away from Microsoft.
Competition is good. I'm glad Apple is finally seeing the light.
If you want on or off the Mac Ping List, Freepmail me.
Pinch me I must still be asleep.
****************************
Its official. After years of speculation and proclamation, Apple has announced its impending shift to x86. Once upon a time, I'd have hailed this as a tremendous step forward for the company--now I'm not so sure.
Using Intel processors will probably cut Apple CPU costs since they'll (presumedly) be buying only-Intel hardware, and will have access to Intel rebates and cost discounts. It may even be a selling point Apple can leverage to increase the ease-of-switching--Apple has announced that their PC's will run Windows perfectly, though they won't ship the OS themselves. There are, however, significant issues and problems I'm not sure how Apple will overcome.
The Emulation Issue:
According to the keynote, Apple will provide a 'revolutionary' code-translation system named 'Rosetta', which will enable PPC software on x86 hardware, with only a minimal performance penalty.
Right. I'll believe it when I see it.
If we step back from Apple's marketing claims and examine the reality of emulation, the evidence is undeniable. Performance hits from emulating other CPU architectures are always significant. True, the architectures involved make a difference--the PPC is much better at emulating x86 and maintaining an acceptable level of performance than Itanium or Transmeta's Crusoe--but "maintaining an acceptable level of performance" and "minimal performance penalty" are not at all the same thing.
Apple's transition to PowerPC from 68K processors is typically held up as an example of an architecture transition where emulation smoothed the gap, but that comparison isn't accurate in this case.
Compare the fastest 68K system Apple ever shipped and the first PowerPC (the Power Macintosh 6100, according to Apple-History.
Macintosh LC 575:
CPU: Motorola MC68LC040
Speed: 33 MHz
FPU: None, Available via Upgrade
Bus Speed: 33 MHz
Data Path: 32 bit.
L1: 8K
L2: None.
Macintosh 6100:
CPU: PowerPC 601
Speed: 60/66 MHz
FPU: Integrated
Bus Speed: 33 MHz
Data Path: 64-bit.
L1: 32K Unified
L2: Optional, 256K standard on 66 MHz version.
By May of 1995 (just over a year after the LC575 was released), Apple had released the Power Macintosh 9500, based on the PowerPC 604 core. This new core offered improved branch prediction, a split L1 cache, improved pipelining, and higher frequencies--it made its debut at 120 MHz, with a 40 MHz FSB.
The point to this comparison is that Apple's PPC transition enabled the company to offer CPUs and systems that dramatically surpassed what was available from 68K, not just in terms of frequency, but across multiple factors. Not only were PPC CPU's faster, they were more efficient, increased effective memory bandwidth, and increased the amount of available L1 / L2 cache--and they did all this within one year of the transition.
Transitioning to Intel now, however, can't possibly offer these types of improvements, and there's no word yet on how the lack of AltiVec capability will impact performance. I'm sure some operations handled by AltiVec can be handled by SSE / SSE2 / SSE3 instead, but again, there's a translation penalty there. Will "Rosetta" be capable of deconstructing PPC code and reconstructing it into x86-compatible code while simultaneously vectorizing it on-the-fly? Again--I'll believe it when I see it.
I don't see any way for Apple to transition its software easily. Rosetta or no Rosetta, this move is going to be rough.
The Marketing Factor:
Ever since the G3 hit a frequency wall, Apple has been pushing twin concepts of CPU efficiency and "Intel's MHz myth." For nearly five years, they've publicly attacked and attempted to humiliate x86 in general and the Pentium 4 in particular--and Mac users ate it up.
Evidence of this anti-x86 bias is clear, even in our previous discussion on the topic. Hit with news that Apple really *was* switching to Intel hardware, Apple faithful denied the possibility, floating increasingly-unlikely alternatives where Apple licensed Intel to make G5's, Apple commissioned Intel for a non-x86 processor, or Apple was going to adopt XScale for desktop / high-end notebook use.
Its not something I can easily quantify on paper, but I can't see Mac faithful taking well to this announcement. I'm not going to predict a mass exodus or uprising, but I'm also not at all sure how Apple's going to deal with its unhappy PPC users.
The G5: From Awesome to Anchor:
So what happens to the G5 now? Even Tiger, aka OSX 10.4, hasn't completed Apple's gradual transition to a fully 64-bit operating system, and there are still notable limitations on what can and can't be done in 64-bit. (Ars has more details on this here).
How much further enhancement and improvement will we see for the G5? Apple has announced that an x86 version of OSX has always been maintained and kept running, but is that version fully optimized for SSE / SSE2? Is it fully 64-bit native?
On the one hand, Apple needs to reassure PPC owners that their investments are not wasted, and that full G5 support and development will continue. On the other, its impossible to underestimate the need to begin working on x86 software development. Again, you can't trust marketing here. AMD boasted to us years ago about how Microsoft was able to boot Windows in a 64-bit environment after only a few days of work--but it took years for the 64-bit edition of Windows to hit the market. Even now, performance lags in some areas, simply because of driver and software support.
The amount of code that'll need to be updated, recompiled, and re-optimized is mind-boggling, and while much of it may lie outside Apple's purview, end-customers aren't going to care.
What impact will this have on present-day G5 sales? Its impossible to say. But it does leave me wondering if Apple has any intention of bumping Powerbook speeds again over the next year, or if they've simply given up on any hope of a low-power G5. I suppose a 1.8 GHz G4 might be possible, but its hard to see IBM putting out much effort to bring out products that Apple isn't planning to use.
Conclusion:
Two years ago, when the G4 was stalled and the G5 was AWOL, I thought it made sense for Apple to jump for x86. Now I'm no longer sure. Steve Jobs has always been willing to do the unexpected and take risks in order to stay one step ahead in the industry, but the timing of this announcement feels off, and it runs directly counter to everything Apple has preached about its own hardware for years.
Those of you who own Macs now, or who were thinking about buying them--how does this affect your plans?