Intel Launching Woodcrest, Conroe and Merom Early?

[Dr.Gargoyle]

My view on the release cycle is this: You have two options:

a. Have releases far appart. When a new release is comes out, your computer is surpased by an insane amount.

b. Have close releases relatively close and have the computer depreciate gradually....


Personally, I bought a Pentium M laptap even though the Solo/Duo cores came
out. I suppose I don't have to worry about "his !@#$ is bigger than mine!" syndrome.

First of all, since I believe that Apple could make a higher profit on alternative b., something tells me that we will see MUCH shorter release cycles.
Secondly, the release cycles on GPU's became shorter with the ADC->DVI transition when many more and "better" GPUs became available to Apple. I believe that the PPC->x86 transition in the same manner will result in an even shorter release cycle since Intel's release cycle is much shorter than that of IBM.

 

[Bern]

Ahoy Mates. Kinda boring in here what? I'll see you both and raise you:

"And with Leopard On Board."

I'll take your Leopard and raise you a larger capacity (ie: 120-160GB Perpendicular) hard drive

 

[p0intblank]

This is a very good thing!

 

[manu chao]

First of all, since I believe that Apple could make a higher profit on alternative b., something tells me that we will see MUCH shorter release cycles.
Secondly, the release cycles on GPU's became shorter with the ADC->DVI transition when many more and "better" GPUs became available to Apple. I believe that the PPC->x86 transition in the same manner will result in an even shorter release cycle since Intel's release cycle is much shorter than that of IBM.

Powerbook cycles were roughly nine months. So a new MBP in September when Merom might come out makes, September (09) minus January (01), eight months. I see, a much shorter cycle thanks to the switch to Intel.

 

[AidenShaw]

forget the historical release cycles

a. Have releases far appart. When a new release is comes out, your computer is surpased by an insane amount.

This isn't really an option any more - when Intel releases newer, faster chips all the other Intel vendors will use them. If Apple waits months to incorporate the current technology, their computers will look "obsolete" compared to everyone else.

b. Have close releases relatively close and have the computer depreciate gradually...

Powerbook cycles were roughly nine months. ... I see, a much shorter cycle thanks to the switch to Intel.

If you look at the rest of the PC industry, there tend to be "major" releases and "minor" releases - Apple will be forced into the same strategy.

"Minor" releases are a CPU speed bump, better graphics, brighter screen, more memory (RAM and/or VRAM), bigger disk.... These will happen whenever Intel or another vendor has a better or cheaper component. (For example, as soon as the 2.33 Yonah is announced it will become available in the MBP. Maybe as a second BTO option, maybe the whole line will move up a notch.) Typically in the Intel world these aren't even called new models, since it's usually just a new BTO option or a slight change to the standard config.

"Major" releases will be changes in form factor (a new sub-2 kilo notebook, mini-tower) or chip/chipset. (You can trumpet "Now 64-bit with SATA and PCI Express", you can't do the same for "Now 7.8% (133 MHz) faster".)

For an example, look at the Dell Latitude D6x0 series - D600 first with Pentium M, went through a number of speed bumps, changed to D610 with PCIexpress, couple of bumps, now D620 with Yonah dual core. Very similar form factor, same docking station and drive bay options - just a continual improvement in specs over 2 1/2 years or so.

So, minor releases will "keep up with the Joneses" and will happen much more frequently. Major releases will be the "one more things", and might not even coincide with an Intel announcement. (For example, a 1.6 Kilo notebook could be announced now, last month, or next month - it doesn't depend on an Intel release.)

 

[Dr.Gargoyle]

Powerbook cycles were roughly nine months. So a new MBP in September when Merom might come out makes, September (09) minus January (01), eight months. I see, a much shorter cycle thanks to the switch to Intel.

I guess AidenShaw already answered this, but since I am a complete nag.
I can't remember when PB went G3->G4, but the PB has been equipped with a G4 proc. for years. Yonah->Merom is a proc transition very much like G3->G4 ("major" release (see AidenShaw)). The Powerbook cycles you are taking about is a speedbump ("minor" release). According to your info we would see "major" releases (diff proc) at the same speed as we have seen "minor" releases (speedbumps) during the PPC era.

 

[Multimedia]

I'll See Your Larger Capacity Hard Drive And Raise You Two Inside & A SuperMultiDrive

I'll take your Leopard and raise you a larger capacity (ie: 120-160GB Perpendicular) hard drive

I'll See Your Larger Capacity Hard Drive And Raise You To Two Inside - 300 GB net total raid or not raid inside AND a SuperMultiDrive.

 

[manu chao]

"Minor" releases are a CPU speed bump, better graphics, brighter screen, more memory (RAM and/or VRAM), bigger disk.... These will happen whenever Intel or another vendor has a better or cheaper component.

Bigger harddrives, brighter or higher resolution screen, better GPU, more RAM etc., all these changes could have been improved in small steps by Apple even while it was using the PPCs.

It might follow the Intel CPU quite closely but that is no guarantee that it always offer the biggest harddrive available etc..

And one may note that Core Duos are already sold in laptops at 2.26 GHz, but not by Apple. I wonder why nobody here complained about this yet.

 

[manu chao]

I can't remember when PB went G3->G4, but the PB has been equipped with a G4 proc. for years. Yonah->Merom is a proc transition very much like G3->G4 ("major" release (see AidenShaw)). The Powerbook cycles you are taking about is a speedbump ("minor" release).

Sure if you see it that way. But everybody talking about Powerbook releases, including the Macrumors Buyers Guide, and all those people talking about having a revision C 12" Powerbook meant these nine months cycle releases.

It is not about the release cycle of Apple accelerating, it is the frequency Apple will present new processor types that is increasing. If 7410, 7445, 7447, 7448 had come out with a nine month difference, Apple would have used them.

 

[Multimedia]

This Year Is Transition Time - Next Year We Can Forgetaboutit

Bigger harddrives, brighter or higher resolution screen, better GPU, more RAM etc., all these changes could have been improved in small steps by Apple even while it was using the PPCs.

There were. How do you see it that they were not?

It might follow the Intel CPU quite closely but that is no guarantee that it always offer the biggest harddrive available etc..

And one may note that Core Duos are already sold in laptops at 2.26 GHz, but not by Apple. I wonder why nobody here complained about this yet.

That's because what is being announced and shipped right now is merely a stop gap to get us from weak PowerBooks to 64-bit serious MacBook Pros with Leopard next year. This is totally a transition year that will see 32-bit mobile models VERY TEMPORARILY filling a gap Apple and frankly all of us wish we didn't have to go though. By next Summer, this will all be a distant memory we will all be glad to forget except the code writers who will have to deal with 32-bit mobile versions of their software for years to come thanks to the speed wars.

 

[Multimedia]

History Tells Us You Can't Guess Accurately What Will Come Next When

I guess AidenShaw already answered this, but since I am a complete nag.
I can't remember when PB went G3->G4,

November 10, 1997 - January 9, 2001 is your G3 era.

Over 3 years. 233 MHz to 500 MHz - 2x improvement over 8 speed bumps, 10 models

January 9, 2001 - January 10, 2006 is your G4 era.

5 years. 400 MHz to 1.67 GHz - 4x improvement, USB 2, FW 800, GB Ethernet, 11 speed bumps, 24 models

but the PB has been equipped with a G4 proc. for years. Yonah->Merom is a proc transition very much like G3->G4 ("major" release (see AidenShaw)). The Powerbook cycles you are taking about is a speedbump ("minor" release). According to your info we would see "major" releases (diff proc) at the same speed as we have seen "minor" releases (speedbumps) during the PPC era.

Seems like most changes have been incremental to me. But when the 1.25 GHz 15" Aluminum PB G4's came out on September 16, 2003, that was quite an improvement and change. Its 17" pre-cursor had ruled all year since the January 7, 2003 MacWorld Expo unveiling while 15" 1 GHz USB 1 Titanium PowerBooks continued to lag behind the new Aluminum form for over 8 more months. Blah Blah Blah Blah Blah...

So you see, there is no rhyme nor reason to how these products will roll out when.

But I think it's safe to say, thanks to multi-cores, the future track record will be much better than those of the G3 an G4 eras.

 

[manu chao]

There were. How do you see it that they were not?

Apple had the tendency to be roughly half a year behind any announcements in 2.5" disk sector (7200 rpm drives anyone?). Similar things can be said for the GPU.

My point is that Apple will not introduce every two month a minor modification of its products, like in January a faster CPU, in March a bigger disk, in May a better GPU, in July a new port (e.g. e-SATA). They will continue to bundle theses minor upgrades to create a more compelling new product, to reduce the number of versions they have handle internally and that they have to support, etc.
The switch to Intel might alter that policy somewhat, but as I said before Apple has not offered the 2.26 GHz Core Duo yet, it seems they will wait at least until the 2.33 GHz chip becomes available and maybe even until Merom.

 

[Multimedia]

2.26 GHz Yonah In the 17" MacBook Pro NAB Announcement Sunday

Apple had the tendency to be roughly half a year behind any announcements in 2.5" disk sector (7200 rpm drives anyone). Similar things can be said for the GPU.

My point is that Apple will not introduce every two month a minor modification of its products, like in January a faster CPU, in March a bigger disk, in May a better GPU, in July a new port (e.g. e-SATA). They will continue to bundle theses minor upgrades to create a more compelling new product, to reduce the number of versions they have handle internally and that they have to support, etc.
The switch to Intel might alter that policy somewhat, but as I said before Apple has not offered the 2.26 GHz Core Duo yet, it seems they will wait at least untill the 2.33 GHz chip becomes available and maybe even until Merom.

I imagine they are releasing the 2.26 GHz Yonah In the 17" MacBook Pro to be announced Sunday at NAB to differentiate it from the 15" model. I also expect Expresscard/54 in this new model to further differentiate it and maximally satisfy the "port gap" problem. Even eSATA may be a part of this new mobile flagship.

 

[BlizzardBomb]

FW800? I'll bet we NEVER see it in an apple laptop again. What other laptops include it?

Soon, you'll be able to use FW800 on the MacBook Pro thanks to it's ExpressCard Slot

I imagine they are releasing the 2.26 GHz Yonah In the 17" MacBook Pro to be announced Sunday at NAB to differentiate it from the 15" model. I also expect Expresscard/54 in this new model to further differentiate it and maximally satisfy the "port gap" problem. Even eSATA may be a part of this new mobile flagship.

You mean 2.16Ghz, with a 2.33Ghz BTO

 

[THX1139]

It'll be called the MacMac.

MacMac and MacTop are stupid names. I hope Apple thinks of something simple and not as silly as that.

Why not just call the Intel desktops "Mac-Pro"?

 

[Photorun]

MacMac and MacTop are stupid names.

You may want to turn on your humor detector and see if you get a read.

 

[Multimedia]

MacMac and MacTop are stupid names.

You may want to turn on your humor detector and see if you get a read.

I certainly did from your post Photorun. lol.

 

[Norse Son]

WWDC in August, Merom/Conroe around same... Coincidence?

Could I get a nice long hit of whatever it is that you're smoking?


Nothing less that true 64-bit throughout would make sense.

Making 10.5's 64-bit support "a little less lame" than 10.4's support would be a big mistake in the long term. Apple can't expect developers to keep porting to different APIs on OSX.

Was that a cheeky slap I felt, or was it just tongue in cheek? I look at it this way:
• Apple says the Intel-transition will be complete (approx.) a year ahead of schedule...
• WWDC was moved roughly 2 months later than normal...
• Surprise! Microsoft is delaying Vista/"Vestage" 'til January (write that in pencil)...
• Apple announces Bootcamp...
• Then Intel "advances" the release schedules for Merom/Conroe/Woodcrest...
• Then I read this today - and I know, Cringley is not without his doubts or faults, but, still, it is intriguing: http://www.pbs.org/cringely/pulpit/pulpit20060420.html

Speaking of APIs, that Cringley blog, if ("if") true, would represent the real Holy Grail - no need to even buy & install Windows on our Macs in order to use certain apps we can't find in the Mac software aisle.

 

[scottlinux]

All the people with concern about their newly purchased intel iMacs:

The intel iMacs have a socket. You can pop the CPU out, and place a newer one in. So in a couple of years if you want to upgrade, just pop in a newer chip. Your purchase should last MANY many years; just upgrade when needed.

 

[AidenShaw]

Was that a cheeky slap I felt, or was it just tongue in cheek

Tongue in cheek, or course - no smiley needed when things are so obvious.

• Then Intel "advances" the release schedules for Merom/Conroe/Woodcrest...

The Intel schedule change is all about AMD - not Apple/Dell/HP/MS or anything else.

However, August is close enough to the earlier schedule that Apple may have decided that they could do an announcement of "shipping next month".

It's also quite likely that even though the schedule is driven by Intel's competiton with AMD, that partners like Apple (and Dell/HP/IBM...) were told under non-disclosure that things were going well and that the schedule was firm.

Availability of chip samples (remember the guy who has a Merom chip that he plugged into a MacIntel) is also good evidence. If samples are easy to find and work well, you have more confidence in what Intel's saying that if they are scarce as hen's teeth or only "mostly" work.

Considering samples, the August date might also have been chosen so that Apple can show a true 64-bit version of OSX running on a "prototype" system with a Merom or Conroe chip. (They may have been developing on the Pentium 4 64-bit chips, but didn't want to show a Pentium 4 again at WWDC. Remember that the Intel DTK systems from last year's WWDC were actually 64-bit Pentium 4s running in 32-bit only mode.)

Speaking of APIs

Interesting, but only if the licensing terms (or renewal) give Apple the right to Vista APIs, IMO.

I can't see Apple doing a big "XP (only) on OSX" around the time that Vista launches.

There's also the big "OS/2" question - would such a move mark the end of any development of OSX-native software outside of Cupertino itself?

 

[shawnce]

• Then I read this today - and I know, Cringley is not without his doubts or faults, but, still, it is intriguing: http://www.pbs.org/cringely/pulpit/pulpit20060420.html

Humm... my opinion for Cringely just went down all the more after reviewing that article. For example...

Speed. Quite simply, a monolithic kernel like the one used in Linux or most of the other Open Source Unix clones is inherently two to three times faster for integer calculations than the Mach microkernel presently used in OS X 10.4

Ummm what? 1) the performance of integer operations are very much decoupled form kernel implementation, 2) Mac OS X doesn't use a micro-kernel (it isn't a Mach micro-kernel... it is a monolithic kernel with a Mach based schedular and IPC system with IOKit and BSD mixed in), 3) two main sources for deltas in system call performance exist between Linux and Mac OS X exist (this is the main source of performance claims you see thrown around).... the first comes from a difference in the focus / philosophy of optimized pathways and the other from incurring a more expensive context switch since on Mac OS X the kernel isn't mapped into application space like it is on Linux (a few pages are for higher speed user land / kernel communication).

Anyway it is possible that 10.5 will not be universal and maybe only support Intel and in so doing allowing them to revamp more of the kernel (not that it needs it... xnu is a good kernel... just add in x86-64 support... and on such systems map some of the kernel into every application to boost system call performance) but all signs from Apple so far have implied that 10.5 will be universal.

 

[AidenShaw]

...all signs from Apple so far have implied that 10.5 will be universal.

That doesn't mean that PPC and x64 builds will have identical features - there will almost certainly be new capabilities that can only be used on x64 processors.

A feature like true 64-bit support, for example.

There are at least three good reasons for bringing features to x64-only:

1. Some hardware features are not available on PPC (like VT "Virtualization Technology"), or only on a subset of the PPCs (like 64-bit), or not reasonable on many PPCs (like 64-bit on an iMac G5 with a 2 GiB memory limit)
2. Apple may find it hard to justify expensive engineering efforts to bring features to the dying PPC platform - do you spend millions to ensure parity between PPC and OSX, just as you are killing off all the PPC systems?
3. The more goodies that are "Intel-only", the more Intel boxes Apple will sell. (The flip side of "Intel switchover ahead of schedule" is "PPC obsolescence ahead of schedule".)

 

[MrCrowbar]

[...]
[*]The more goodies that are "Intel-only", the more Intel boxes Apple will sell. (The flip side of "Intel switchover ahead of schedule" is "PPC obsolescence ahead of schedule".)
[/LIST]

Good Point. Steve promised support for the PPC platform for a few years but support != developement. So this means mostly bugs will be fixed for the PPC platform whereas the real developement will go into the x86 platform. And as the upcoming Intel processors will all have 64 (correct me if I'm wrong), it's very important for the OS to make some use of that. The G5 may be 64 bit, but that was seldom used by the OS to make things faster.

PS: Sorry if I write about the PPC stuff in the past tense sometimes. Face it, the PPC is dead for Macintosh. Once the Intel PowerMacs (with the right apps) are out, no one would praise the G5s anymore. Besides, some apps out there are frozen for PPC, they develop for x86 and that is a good thing. Tecnology is beginning to develop much faster than we (the users) can adapt to it. There is exponential growth in evolution and it's shifting from carbon based creatures to mashines. Raymond Kurzweil (http://www.kurzweilai.net/) says around 2030 we will have mashines equally intelligent and creative as humans. The mashines will outsmart us (paradigm shift) very fast because they can influence their intelligence directly which we mere mortals cannot. So stop mourning the PPC please.

 

[Norse Son]

Thanks, guys, for the Leopard info, but...

As usual, my head hurts now from trying to fathom all the "geek-speak" (no offense).

I know that 64bit will be the future - once MacOS X is fully 64bit and once major apps add 64bit support - but the interest I had in Cringley's "ramblings" - unsubstantiated or not - was the WinAPI stuff. Personally, I don't want to have to install XP or Vista on my Mac just to use the few odd apps or games unavailable for MacOS X. And if I can use the apps, then Microsoft only receives indirect revenue from it - 3rd party developers do pay a royalty to MS for licensing the APIs, don't they?

And, in my opinion, most, if not all, of the apps and/or games I would want to buy have never been available for the Mac, so it lessens the impact on Mac developers.

 

[shawnce]

That doesn't mean that PPC and x64 builds will have identical features - there will almost certainly be new capabilities that can only be used on x64 processors.

I wasn't attempting to imply otherwise just that I don't see a radical xnu replacement in 10.5 as Cringely was implying (and none of this switch to Linux kernel or Windows crap some are throwing out). I see an evolution of xnu that maintains PowerPC support while growing to enabled / leverage capabilities that Intel chips bring to the table. I think things top on the likely hit list are x86-64 support (at a minimum like what we see on the G5 but more likely well beyond that... kernel using x86-64 and/or common/all frameworks providing an x86-64 ABI), some facilities built around virtualization capabilities present in the Intel CPUs Apple is using and support for RAID, etc. capabilities built into the Intel chip sets.

It is already nice to see that Apple is leveraging GPT and enhancing/ed HFS+ to allow live disk partition resizing (as seen in Boot Camp and 10.4.6... likely more widely supported/exposed in 10.5).

In the 10.6 time frame we will likely see ELO for support of PowerPC and possibly more radical changes to xnu... but not in 10.5.