Originally posted by mcs37
There best bet is to blitzkrieg the world with dual PPC970 machines, including dual PPC970 17" PowerBooks. If they wait around for their older machines to sell, they'll only hurt themselves more because people will be waiting for a PPC970 PowerBook (why buy a 1 Ghz G4 PB when you can wait a few months and get a 2 GHz PPC 970 PB?). In order to make a big splash this year ("of the laptop"), they need to throw everything they've got at us immediately. It worked for the Germans!
PPC970 from the information presented on it's specs was going to be too hot for notebooks. I think the laptops will remain G4 for the time being, likely with speed bumps until a future PPC 9xx variant is ready. Likely when they shrink the die size and make it more efficient than it is already (it's almost 1/2 way there now from the specs I saw), which for being a more "desktop-sized" processor (G4 is a small embedded-style processor with a SIMD core on the chip and some other goodies for supporting SMP), I pretty well expect that future variants won't be anywhere near as hot as the G4 laptops which have been ridiculously warm.
I do kind of expect the laptops to get a speed bump to approx. where the desktop G4's are in speed. I do think this is the end of the road for Motorola in Macs after the G4 does it's course. If IBM at any point fails... I figure Apple will move to Intel/AMD. I just don't see IBM failing...
Oh and on Xeon vs. PPC 970 pricing. Are you crazy? If the PPC 970 costs anywhere near what the Xeon does, the base machine will be coming out for around $2k, with the top of the line up around $3-4k for a single processor machine!!! LoL Not only do you have to factor the Xeon processor at $200+ a piece (and that's a rough guesstimate, the low-end quote we saw earlier was around $285), but you have to figure on a Xeon motherboard suitable to the task. Then fans for cooling it. Then the case. It's not just the processor price... it's the total cost of everything involved in making a machine at a factory. Parts aren't the only expenditures involved here... and that's why a low end Xeon isn't dirt cheap.
That isn't to say that a PPC 970 machine won't be damn near as expensive (hoping not)... if not more if all the rumored inclusions and goodies hold true as the motherboard technologies sounds quite amazing IMHO. Yet... my beliefs is the 970, which IBM is hoping to have shipping in it's own line of workstation/light server class machines, will be able to ship these machines in somewhere between June-September. Considering Apple has often dealt with September launches... that's a safe guesstimate, but... considering that the G4 desktop sales have been stagnant, I half-expect to see PPC 970's in the desktops and XServes, while we see all of the iMac/eMac/Powerbook/iBook get speed-bumped G4's... with the consumer desktops going as fast as they can go (with heat withstanding), much as the laptops. I don't foresee dual G4's in laptops unless Motorola has something up their sleeves, and I don't think a PPC 9xx laptop will happen before 2004.
It also makes even more sense (Apple moving to PPC 9xx) when you consider that the main reason IBM didn't sign on for the PowerPC G4 roadmap with Motorola was that they had "NO" interest in SIMD technologies. Linux and AIX don't take advantage of said technologies (very little point, none of the apps would either, and they're geared more towards servers where ::gasping:: as noted earlier, most server processors like Power4 and Itanium aren't focused on; comparing them to Pentium/Celeron, Athlon/Duron, and PowerPC is like comparing Hummer H2's to Ducati 999's; different applications, and both are performers in their classes at a specific pricepoint), and place more of their focus on ramped up processor speeds and focuses in other areas in terms of hardware and chip design. IBM's own 64-bit G3 (yes G3; IBM's had 64-bit chips dating back to the PowerPC 615, which was around the time of the 601, 602 [used in the Bandai Pippin @world], 603, and 604 computer processors and the various 5xx embedded chipsets) machines were more of a testament to this, as the 32-bit G3 was more for embedded... which if you really do your research, the G3 is one of the faster embedded-class processors you can find in terms of processor speed. Not that ARM is slow for it's application... it's just designed for a different application of the embedded market.
So the PPC 970 to me sits between the Pentium 4 and the Pentium 4 Xeon (faster than the P4, but slower than the Xeon at least at the initial Mhz, but working favorably for servers where the Xeon is not as strong; which works to IBM's and Apple's favor), while providing tremendous ability to scale, and likely at a more affordable pricepoint. I don't expect a PPC 970 PowerMac to sell for significantly more than the current Pro line. If not, I don't see it being viable until the prices come down.
The SIMD points an Apple tie is "IMMINENT", and pointed that as soon as I read that IBM was including one in the PPC 970. It was like a flag went off and it was obvious that Apple was involved. Since there's not a single "CONSUMER" Linux variant that uses SIMD, and there's not much inkling to support said technologies within the core of the system when you're dealing with a codebase designed to be ported for more workstation/server level stuff (across multiple platforms)... it kind of sticks out like a sore thumb.
The information released on MacAddict months ago added to the credibility of the rumor, when an IBM engineer spoke of faster yields and a summer launch for IBM machines. The information was "SWIFTLY" removed from the site... but by that time it was caught and printed. Just about every Mac rumors site was hopping with the info... and considering the information, it was a no-brainer. Apple cannot afford to make a major transition to anything right now, it's too soon. The fact the PPC 970 will read legacy code, and can be modified to be 64-bit savvy works well with Apple's XServe range of light servers, and works well as a workstation/desktop processor. Over time, I feel it'll make a great laptop processor as well.
To me it's not a question of "if" but "when". I figure before Christmas (after all, the shopping season leading up to Christmas is a big time for Apple, and could bolster their sales by setting Apple sailing into 2003 on a huuuuuuuuuuuge positive note, spurred by the new laptops from earlier this year, the iPod updates, the new iTunes Apple music store, and into 2004 with new PPC 9xx laptops, PPC 9xx consumer desktops, speedbumps to the new PPC 970 desktops, any other iDevices, and iTunes for Windows with Apple's iTunes Music store), but I'm not going to go out on MacBidouille's limb and say "Next day". Then again... as long as we've sat without a major change, I'm sure Apple's been doing more R&D than you can shake a stick at to get this out there. It might release then... it might release later... as long as "SOMETHING" releases though, it'll be good, no matter when it happens.
My guess is... if Apple indeed moves to PPC 970, which I expect, unless the hardware is ready by WWDC, we may not hear a whisper of it... but we may see remnants in early Panther builds (if Apple gives one out, which I'm skeptical on) that leak out. Apple wouldn't want to announce the move to PPC 970 before it happens, for the hopes they can milk a few more G4 sales up until the machines actually launch. Which could be late August or September.
So what I'm saying is... don't get too ramped up or wound up to get disappointed, but just go in knowing that this is more than likely the direction. It's just a matter of "when", not "if" it will happen, as I said earlier. If it happens by the end of the year... it's a "GREAT" thing. If it happens by Q2 of 2004, it's a "GOOD" thing... any later... and we'll be playing catch-up still, but at least we're playing catch-up with someone that has some interest in doing so. We don't need to be as fast or faster... just as long as we're "close", that's the important detail. Apple could easily rework OS X to be more efficient via threading (comparing OS X or XP to BeOS shows what I mean)... and the speed difference would go in our favor. It's not just raw processing power... it's how you maximize the power that's the big issue. OS X still has some areas it could be honed... and that could make all of the difference.
Hopefully the Panther/PPC 9xx combo brings us to the forefront.
🙂
