The Sony E3 megaton is ... Apple?

[archie4oz]

The Mac rumor sites speculate that IBM is trying to make dual-core processors for Apple. I would think dual-core G5s are in IBM's processor road map.

That is why you wonder about IBM's ability to produce enough triple-core CPUs at 3.2 Mhz for MS in volume.

Roadmap? The dual core 970MP already exists... They accidentally let that slip last month...

As far as IBM's ability to produce CPUs, they've got capacity to spare, and MS is licensing the CPU so they can have other foundries manufacture the CPU if need be.

 

[ndoogoo]

Does mean the PS3 will crash alot? I almost finished DevilMayCry VIII and the game crashed with lots of little bombs on the screen.

 

[PC-Engine]

But Apple choosing to leave IBM completely? For whom? Freescale? AMD? Sun? Too little benefit for too much work unless IBM is truly ripping them off. I doubt Apple would trust Freescale to tie their own shoelaces, much less bet the computer side of the company on them. No, IBM it is. Both sides stand to gain from the relationship.

All I can tell you is that new versions of certain IBM software for MacOS X, relating directly to its PowerPC chips, was unexpectedly cancelled, even though they were 95% complete and released on other platforms already.

http://news.yahoo.com/s/ap/20050523/ap_on_bi_st_ma_re/wall_street

 

[BlueTsunami]

But Apple choosing to leave IBM completely? For whom? Freescale? AMD? Sun? Too little benefit for too much work unless IBM is truly ripping them off. I doubt Apple would trust Freescale to tie their own shoelaces, much less bet the computer side of the company on them. No, IBM it is. Both sides stand to gain from the relationship.

All I can tell you is that new versions of certain IBM software for MacOS X, relating directly to its PowerPC chips, was unexpectedly cancelled, even though they were 95% complete and released on other platforms already.

http://news.yahoo.com/s/ap/20050523/ap_on_bi_st_ma_re/wall_street

WOW...so that would be a huge blow to IBM......I can see how that would make Apples PC less expensive though...

 

[Glonk]

This would be anything but a huge blow to IBM. Apple's business with IBM was extremely marginal. In fact, that's one reason Apple would like to jump ship.

Apple computers are more or less commodity hardware, which is something IBM has been actively getting out of over time.

IBM uses its PowerPC 970 chips in its own servers, that's what they were originally designed for.

 

[Carl B]

The whole thing is pretty strange, and would require a pretty serious software-shift from Apple. But, there are certainly cost advantages to be had. Anyway for myself, I'll wait for more to either materialize or dematerialize with regard to this particular rumor.

 

[Glonk]

That wouldn't be a megaton for sony. it would be a megaton for IBM, whom already works with apple.

IBM won't work with Apple forever.

Hint hint...

8)

 

[Kolgar]

Yes, I heard the news today... Apple's transitioning to Intel chips and away from IBM.

All right, spill it!!!

 

[Entropy]

This would be anything but a huge blow to IBM. Apple's business with IBM was extremely marginal. In fact, that's one reason Apple would like to jump ship.

Apple computers are more or less commodity hardware, which is something IBM has been actively getting out of over time.

IBM uses its PowerPC 970 chips in its own servers, that's what they were originally designed for.

Well, if IBM isn't interested in making competitive chips for Apple, Apple has to move on.
Quoting myself

I guess I'm just stumped as far as good options go. Freescale has been talking about chips that might be interesting to Apple due to their low power draw, where the 970 and co-logic never really panned out satisfactorily. And that's very important to Apple with their roughly 50% portion of laptops, and their interest in form factor flexibility for desktops.

As far as I can understand, this seems to be the crux of the matter. Steve Jobs didn't say that he couldn't get the performance out of PPCs that he wanted, he said that he couldn't get the performance per Watt that he wanted. Though I'm sure price plays into it as well.

Even though I guess it shouldn't have been such a surprise, it sure was. I can understand Apple not trusting Freescale to deliver, and I can see their frustration with the lack of anything PowerBook/iBook/Mac mini worthy out of IBM, but I can't help bemoaning that Apple didn't cough up the money necessary to fund development of Cell/Xenon derivatives for their needs.
I really like the IBM FPUs. Drats.

 

[Bobbler]

... ...
As far as I can understand, this seems to be the crux of the matter. Steve Jobs didn't say that he couldn't get the performance out of PPCs that he wanted, he said that he couldn't get the performance per Watt that he wanted. Though I'm sure price plays into it as well.

Anyone see the irony in that? (I know this is what Apple said)

Switching to Intel for performance per watt on desktops?

I have a feeling it's more about the money -- the G5 is partly Apple's fault, not just IBM. IBM is fully capable of handling the tiny load of Apple and G5s performance per watt is a lot higher than the P4s. They are going to P4s for a while at least, so I can't imagine performance per watt actually matters much to Apple. However, for laptop chips they have a point, but it made me chuckle to see any company switching to Intel because they want high performance per watt.

 

[Entropy]

... ...
As far as I can understand, this seems to be the crux of the matter. Steve Jobs didn't say that he couldn't get the performance out of PPCs that he wanted, he said that he couldn't get the performance per Watt that he wanted. Though I'm sure price plays into it as well.

Anyone see the irony in that? (I know this is what Apple said)

Switching to Intel for performance per watt on desktops?

I have a feeling it's more about the money -- the G5 is partly Apple's fault, not just IBM. IBM is fully capable of handling the tiny load of Apple and G5s performance per watt is a lot higher than the P4s. They are going to P4s for a while at least, so I can't imagine performance per watt actually matters much to Apple. However, for laptop chips they have a point, but it made me chuckle to see any company switching to Intel because they want high performance per watt.

Stating what should be obvious - Apple is unlikely to ever release a Mac using the NetBurst Architecture. The first machines will be released in a year, and are reputed to be portables. Desktops will follow, and judging by Intels roadmaps, the NetBurst architecture will be phased out by then.

Both the statements of Steve Jobs, and those of Paul Otellini were clear - this is about power consumption, and its implications.

Money is always a factor, but if Apple had seen a good future with IBM, I doubt they would have taken this risk. Simply getting somewhat cheaper processors (if they are) would hardly motivate the work and the risk involved.

 

[Acert93]

Anyone see the irony in that? (I know this is what Apple said)

Switching to Intel for performance per watt on desktops?

Intel is pushing 65nm quickly. They have said it will help fix some of their leakage/power consumption issues. And it looks like the Apple deal dovetails with the time Intel will be switching over.

And with some PC makers putting Pentium M processors on dekstops, maybe Apple will do the same? Yonah is looking like a pretty nice chip imo. The only thing it lacks is 64bit EMT64 which I am not sure is a big loss in the first place (at least not right now for common desktop apps).

 

[chachi]

Money is always a factor, but if Apple had seen a good future with IBM, I doubt they would have taken this risk. Simply getting somewhat cheaper processors (if they are) would hardly motivate the work and the risk involved.

To a large degree this seems to be about IBM's inability to come up with a reasonable low-power G5 (and I just read an article about laptop sales now outpacing desktops), it's been two years now, but it does seem to have some tangential reflection on Cell as a workstation processor. If it was a clear performance win (and you'd think - 218 Gflops, blahblahblah) then Apple would be nuts not to do it, but maybe real-world tests aren't quite so impressive. That's not to say that for some things Cell won't be great, and I expect gaming to be one of them, but for a workstation maybe not.

 

[Fox5]

Hmm, I'd imagine Apple's desktops and high end laptops won't switch over to Intel until a 64 bit pentium M is released. Hmm, does this mean apple will do a 32bit x86 operating system, and then a 64 bit one?

BTW, wouldn't the slower G5s, like the 1.8ghz and 2ghz, be suitable for a laptop? They have power consumption ratings that I think are around centrino levels, though performance isn't.

 

[one]

That's not to say that for some things Cell won't be great, and I expect gaming to be one of them, but for a workstation maybe not.

Apple Powerbook = workstation?

 

[chachi]

That's not to say that for some things Cell won't be great, and I expect gaming to be one of them, but for a workstation maybe not.

Apple Powerbook = workstation?

They're not switching just the portables.

 

[one]

That's not to say that for some things Cell won't be great, and I expect gaming to be one of them, but for a workstation maybe not.

Apple Powerbook = workstation?

They're not switching just the portables.

If their main concern is a server like Xserve which is just a mini-G5, then they had no reason to switch from Power in the first place.

According to Apple they've been preparing x86 for the last 5 years, Cell is too late for use in Apple products considering that selling Cell outside of STI starts in 2007 and there's the burden of software optimization for a new massively-multicore architecture.

 

[Glonk]

Money is always a factor, but if Apple had seen a good future with IBM, I doubt they would have taken this risk. Simply getting somewhat cheaper processors (if they are) would hardly motivate the work and the risk involved.

To a large degree this seems to be about IBM's inability to come up with a reasonable low-power G5 (and I just read an article about laptop sales now outpacing desktops), it's been two years now, but it does seem to have some tangential reflection on Cell as a workstation processor. If it was a clear performance win (and you'd think - 218 Gflops, blahblahblah) then Apple would be nuts not to do it, but maybe real-world tests aren't quite so impressive. That's not to say that for some things Cell won't be great, and I expect gaming to be one of them, but for a workstation maybe not.

It's not the inability to come up with a reasonable low-power G5 -- it was that IBM didn't even try. The G5 is what it is because IBM could also use it in its own products -- blade servers/low-end severs. There was some incentive to the R&D involved.

IBM wouldn't have much use for a laptop chip for itself at all really, that'd just be Apple -- and Apple's market just isn't big enough to justify the use of human resources (which are already spread thing at IBM Micro.).

 

[cthellis42]

IBM wouldn't have much use for a laptop chip for itself at all really, that'd just be Apple -- and Apple's market just isn't big enough to justify the use of human resources (which are already spread thing at IBM Micro.).

As seen in an NYT article:

Even as a chip maker, I.B.M. has moved aggressively beyond the PC industry, focusing on making the processors for video game consoles from Nintendo, Microsoft and Sony, and specialized chips for other uses, like the Internet router computers made by Cisco Systems and cellphone technology by Qualcomm. I.B.M. also uses its Power microprocessors in many of its own server computers, which run corporate networks.

By contrast, the chips I.B.M. makes for Apple represent less than 2 percent of chip production at its largest factory in East Fishkill, N.Y. And while the microelectronics business as a whole is strategically important for I.B.M., it is a small part of the revenue of a company that increasingly focuses on services and software. A. M. Sacconaghi, an analyst for Sanford C. Bernstein & Company, estimates that the company's technology group - mostly microelectronics - will account for less than 3 percent of I.B.M.'s revenues and 2 percent of its pretax income this year.

I.B.M. supplies about 50 percent of the microprocessors used by Apple, providing them for desktop and server computers. Freescale makes the processors used in Apple's notebook and new Mac mini computers.

2% of East Fishkill produces 50% of Apple's microprocessor use. Pretty small indeed.

For years, according to industry analysts, the work for Apple has been barely a break-even business for I.B.M. When the two companies were negotiating a new contract recently, Mr. Jobs pushed for price discounts that I.B.M. refused to offer. For I.B.M., "the economics just didn't work," said one industry executive who was briefed on the negotiations. "And Apple is not so important a customer that you would take the financial hit to hold onto the relationship."

I assume that is mainly the reason for the split. IBM wasn't going in the directions Apple wants or needs, their R&D is focused in other directions, and Apple is too small a partner for them to spend the extra time specifically on their not-much-more-than-break-even procs. The biggest loss for IBM is just a high-profile customer (and brand-wise, Apple comands FAR more visibility than they actually pull in business-wise, so IMHO even if they lost some money on Apple it would still have been better to keep them under their wing) and a shining point to look at for PowerPC chips (though of course they've gained a few comers to counteract it), but if Apple is trying to push against the ocean to get IBM to do what they need, then they probably are much better off moving to Intel (provided it's no headache for their developers or consumers) as they likely get better performance in the directions they need for their machines, and cheaper chips to boot. I guess they also just see more future to Intel's chips than Freescale's possibilities, as well. (I mean, the potentially could have picked up Freescale themselves to make a lot of chips in-house, but I think there would have been many more headaches associated with that and a lot more risk.)

Offhand, I'm mainly interested to see what Intel will be making for them. I assume that although OSX can run on stock Pentiums, they may try to get a bit of unique design in there as well. (Especially since it will make it easier for them to control their OS instead of seeing it leak to any x86 machine by the hackers and coders who just wet themselves in anticipation today. )

At any rate, the future should be interesting as well, as Apple and their developers will likely be supporting multiple formats (and keeping up the possibility of more, mayhap?) in the years to come, so they may be able to go to various different parties for the best deals and most appropriate chips for each of their specific products.

I'm still waiting to see if they can finagle DirectX out of all this, though.

 

[chachi]

It's not the inability to come up with a reasonable low-power G5 -- it was that IBM didn't even try. The G5 is what it is because IBM could also use it in its own products -- blade servers/low-end severs. There was some incentive to the R&D involved.

IBM wouldn't have much use for a laptop chip for itself at all really, that'd just be Apple -- and Apple's market just isn't big enough to justify the use of human resources (which are already spread thing at IBM Micro.).

For laptop read low-power which would of course be great for blades and clustering. The thing is, IBM didn't have a market for the 970 that didn't cannibalize sales of more expensive systems, or commodity type servers, both of which probably had higher profit margins. It also seemed like IBM was dragging their heels on improvements in order to keep those higher end sales safe (i.e. 512K L2 in the 970FX, 1MB should have been a no-brainer, especially when the memory latency is comparitively poor and you need water cooling to get 500 MHz.)

It's too early to say what entirely Apple's motivations are, but it's interesting that one of the big slides that Jobs used to explain the switch...

... specifically says integer performance, which is something that has been the source of chatter about for the new console processors.