Apple's Intel Move... Early Details and Notes

[zakatov]

Why Photoshop loaded so slowly at the Keynote... The universal binaries do not take advantage of the EMT64 and are not going to.

Photoshop opened so slowly because it is (for now) a native PPC app and thus it was emulated thru Rosetta. I thought Steve made it clear enough during the keynote.

 

[shawnce]

My predicition for MacIntels...

(from an email I sent to Mac 0S X talk list)
------
Again it isn't about now but the near future.

Found a summary of the near future (may I say it looks bright)...

http://freespace.virgin.net/m.warner/Roadmap2006.htm

My Wiji boards says...

### iMac / Mac mini system of mid/late 2006 will use:

Intel Sossaman CPU is expected to be released in H1 2006. Although Sossaman is expected to be a desktop CPU, based on the E7520 chipset, it is expected to be a dual core processor based on the Yonah core. Sossaman is a low power processor, with the 2Ghz revision having a TDP of 31W, with the 1.67Ghz LV version having a TDP of just 15W.

### PowerBook / iBook system of mid/late 2006 will use:

Intel Pentium-M (Jonah / Yonah) mobile CPU is expected to be released in Q1. Jonah is part of the Napa platform and the successor to Dothan. It is expected to be built on a 65nm process and contain two CPU cores on a single die taking 151.6 million transistors (compared to around 140m in Dothan). Jonah is expected to feature a 667Mhz FSB speed and a 2Mb L2 cache shared between the two cores. In addition to the featureset of Dothan, Jonah will feature an updated SSE instruction set (SSE3 support and SSE/SSE2 support tied to MicroOps Fusion, and throughput improvement in the SSE Decoder) along with improved performance of the x86 FPU. These improvements to the floating point performance of the processor are dubbed "Digital Media Boost" by Intel. Jonah will also feature improvements to power management and thermal output, meaning that the power consumption of Jonah will be similar to that of Dothan, despite the increased core size and complexity. Intel's Advanced Thermal Manager offers finer grained thermal management and Yonah will allow for one core to be slowed down independently of the other if power consumption or heat dissipation rises too much. Jonah will also feature Intel's Vanderpool technology and LaGrande Security technology, but will not contain x64 instruction set support in it's first incarnation.

### With the following waiting in the wings for those systems in 2007:
Intel Merom Mobile processor, the successor to Jonah and part of the Santa Rosa platform, is expected to be released in Late 2006. Merom will be built on a 65nm process and is expected to feature a different architecture to previous Banias-based processors, providing a 20-30% performance improvement per clock. Merom is expected to be a dual-core CPU.

Intel Conroe desktop CPU is expected to be released in Late 2006. Conroe is a dual core CPU based around the Merom core (i.e. it is based on a non-Netburst architecture). In order to increase the performance of this CPU for the desktop market, some of the power constraints from Merom will be removed. Conroe will feature virtualization capabilities, LaGrande technology and 64-bit capability in addition to EDB, EIST and iAMT2.

### PowerMacs will use PPC into 2007, then something along the line of the following

Intel Conroe desktop CPU is expected to be released in Late 2006. Conroe is a dual core CPU based around the Merom core (i.e. it is based on a non-Netburst architecture). In order to increase the performance of this CPU for the desktop market, some of the power constraints from Merom will be removed. Conroe will feature virtualization capabilities, LaGrande technology and 64-bit capability in addition to EDB, EIST and iAMT2.

Intel Xeon DP (Woodcrest) processor is expected to be released in Q1. Woodcrest is the successor to Dempsey, but is based on the Conroe core.

(nice to have some actually generally public road map information)
----

Again these will be Macintosh systems running Mac OS X not Wintel [crap] systems.

-Shawn

 

[swingerofbirch]

great for those of us with windows needs

I work for a company that I am not supposed to talk about..but I can say I do customer service from home...and it requires a PC and PC hardware...won't even work under virtual PC..so I have a $600 pc that does nothing for me but one app.....i would love to consolidate my machines to one mac!!!!!

 

[GrannySmith_G5]

Anyone know an estimate of Apple's market share in the home market?

 

[~Shard~]

So let me state first that there is absolutely no proof provided. I asked you for proof and you quote some web article that gives me no details. Just some vague facts. I want to know what permanently connected to the internet means (do they exclude all office PC's?). I'd really like to know the makeup of these computers they looked at. I want to know what some real facts not some potentially made up statistics. I checked on the web site that is mentioned in the article and found nothing unless I pay.

Yep, feel free to read the entire comprehensive study and the answers to your quesitons - for a low, low price of only 23 quid! 😉

So, I did some more digging and it seems that some people don't really believe/like mi2g.

Of course there are people who disagree, especially Linux advocates - this is to be expected. I'm sure they have some valid points, but I am also sure that an organization such as mi2g has some valid points as well. There are always at least two stories and two opinions to every claim.

So, please provide me some details here to back up your claim. Surely if it is true, you can provide me with more than one highly suspect source.

This was essentially the world's most comprehensive study ever done on OSes by a very reputable company (calling them a highly suspect source isn't quite accurate, but again, everyone is entitled to their opinions). If that isn't enough proof then I don't know what is. 😉

Edit: Let me provide some facts and statistics here: I'm in charge of say 20 linux boxes, probably 5 Macs, maybe 30 Windows PCs and hell 6 Sun Sparc boxes. Several of the Linux boxes have gone months or years without updates as have the Sparc boxes and Macs.

My statistics are:

0 hacks with Linux/Mac/Sparc
~50 exploits on the Windows boxes in the last year

Fair enough - you have personal experience which I do not have, and that definitely counts for a lot. All I was doing was presenting you with some significant data on the subject, that's all. Take it how you like.

 

[~Shard~]

I have added the bold highlighting and here is my rebuttle:

http://www.google.com/search?hl=en&q=mi2g&btnG=Google+Search

(It looks to me like this company is anything but renowned and respected)

Again, people are entitled to their opinions. And I understand where you're coming from too - you personal experience has been contrary to this study's findings, and also you obviously are a fan of/enjoy using Linux. Nothing wrong with that, and nothing wrong with defending your likes - I would defend something that I believe in as well (especially when it comes to Apple/Mac! 🙂)

So all in all, I respect your thoughts and opinions. Once again, just presenting you with some facts. Take them how you like.

 

[DakotaGuy]

Intel Sossaman CPU is expected to be released in H1 2006. Although Sossaman is expected to be a desktop CPU, based on the E7520 chipset, it is expected to be a dual core processor based on the Yonah core. Sossaman is a low power processor, with the 2Ghz revision having a TDP of 31W, with the 1.67Ghz LV version having a TDP of just 15W.

As far as the iMac is concerned...how is this better then the G5 in the iMac right now? Power consumption doesn't really matter with the iMac. They can cool a G5 just fine in it and I would assume they could cool a 2.3Ghz G5 without any problems. I don't see the performance increase there especially when you start running programs on Rosetta. You are then taking Alti-Vec away from the system as well...plus the awesome bandwidth of the G5. Sure you may gain a dual core, but you loose a lot of the G5 benefits.

 

[tsk]

This was essentially the world's most comprehensive study ever done on OSes by a very reputable company (calling them a highly suspect source isn't quite accurate). If that isn't enough proof then I don't know what is. 😉

Again, I have to object to "very reputable" company. Just doing a google search, I can't find anything but negative comments about the company. That's not very reputable. I found some info that they only considered manual attacks (ie they ignore worms and viruses, but only look at exploits where the hacker sniffs a password and then hacks or something--I don't know if it's true, but they have no info to back their claims).

Hell, most anyone who runs a Linux box sees the effects of Windows security in their log files.

 

[jared_kipe]

Windows and Mac in one box... *salivates
Wait that sux... I'm gonna be spending half my time on the computer reformatting and clearing out the viruses, spyware, pornware, trojans and crap , might as well get 2 seperate boxes so i have a computer to use while reformatting windows.

You realize you can reformat a specific partition? You only have to erase windows, your mac files would still be there.

 

[isaacc7]

Yep, feel free to read the entire comprehensive study and the answers to your quesitons - for a low, low price of only 23 quid! 😉



Of course there are people who disagree, especially Linux advocates - this is to be expected. I'm sure they have some valid points, but I am also sure that an organization such as mi2g has some valid points as well. There are always at least two stories and two opinions to every claim.



This was essentially the world's most comprehensive study ever done on OSes by a very reputable company (calling them a highly suspect source isn't quite accurate, but again, everyone is entitled to their opinions). If that isn't enough proof then I don't know what is. 😉

Well, their methodology is pretty terrible. Raw numbers of total systems hacked tells you nothing about an OS's robustness. A better measure would be a percentage inside of each OS being hacked. An even more useful thing would be to show the types of exploits being used. Odds are that most of the Linux hacks were caused by sloppy administration and/or human engineering rather than actual vunrabilities in the OS. The same things can result in OS X/BSD compromised systems, but they will only be a blip on that company's radar since so few peopleare running that system. The fact remains that there are very few open vunrabilites in Linux systems (if any) and that Windows is terrible in this regard.

Isaac

 

[jiggie2g]

This Is going to be you're Next CPU.

1st I will say you need to worry , you will never see a P4 in a retail MacIntel PC
these are the most likey candidates. By 2006 the P4 will be EOD'd as even Celeron will be dual core.


Enjoy 😀

http://theinquirer.net/?article=23055

This Is what Jobs ment about Intel will have more units per watt , 70 for Intel vs. 15 for IBM.

 

[elo]

I just bought a Mac for the first time, and so far I love it! But what does this switch mean for me, having just bought a computer using the IBM two weeks ago? Are software companies going to stop making software for my version of the Mac, or will it matter? I'm I going to have to buy a new iMac a year from now because I can't find software for mine (I was hoping to get atleast 3 years out of this one)
-----
And if you ask me...
Apple might as well just buy a toilet to flush their company down the drain, it'd be much faster than what thier doing now.
After all of this I predict apple will loose the qualitly they were once known for. And people running Windows on a Mac, sounds like a bad dream.
And if I'm wrong about any of this, or I'm clearly misunderstanding, please let me know.

There's no reason why Apple would lose quality in this switch. They won't be giving up any control over their hardware designs. In fact, it's highly probable that they negotiated a good bit of control over what Intel will be sending them. Intel had been trying for years to secure Apple as a client, so Apple had very good bargaining leverage.

We actually know quite a lot about how software developers will respond, as they are being asked to do almost exactly the same thing that they were during the 68xxx to Power PC transition. After the switch was announced, development for the old chip continued in earnest for a couple of years, driven by the number of Macs still using the old hardware. While software can push new hardware development, it can't get far in front of the machines people actually have. If you buy a G5 iMac today, for example, you should expect about two years of fairly normal upgrades. The third year, you should notice that most new software either doesn't support you, or supports you in a limited fashion. That, of course, doesn't mean the computer becomes any less useable at that time. I'm planning to pick up a new iMac later this summer, and feel very comfortable with the investment.

For anyone who cares about the future of the Mac, the universal version of Microsoft Office is worth its weight in gold. For the first time ever, a typical business/professional user can switch without the added expense of a new version of Office. For the first time ever, a company that depends (as most do) on one or two Windows-only applications can seriously consider switching to Macs. This will boost the Mac at least as much as the swtich to standardized peripherals (USB, VGA) did with earlier Macs.

People concerned about the Mhz myth and Apple's benchmarks are somewhat confused. The Mhz myth is very real. Intel, too, acknowledged it. That's why Pentium M chips run at lower Mhz numbers than Pentium 4's of comparable speeds. Apple was also completely correct in saying that the G5 had the more promising roadmap when it was released. But IBM never delivered the goods. Sometimes you have to have a spine. Apple did.

 

[jared_kipe]

Listen, to the people that are worried about Rossetta not working well with much software. Consider the following, the ibook had a G3 processor in it up until late 2003. Basically it sounds like Rossetta will run anything that will run on a 2003 ibook G3, this was after the G5 powermacs were announced incase you forgot.

 

[intrepkid21]

Oh boy! Fresh thread lol.

It's going to be interesting to see what Apple has up it's sleeve for the first x86 Mac. Not so much the OS and software, I'm sure it will still "just work." Well, it will somehow...

What concerns me is the commodization of the Mac hardware platform absent PPC and Altivec. With an x86 Intel processor, what else is inside exactly to distinguish it from any other garden variety PC? Hard drives are the same. Interconnects and protocols are the same. Memory is the same. The form factors may look great, but if the only difference is the installed OS (and apps), I think it's a big problem.

Imagine Mr. Switcher visiting one of Apple's 200+ stores in the next 2 years and eyeing Apple's brand new hot box, the x86 PentiMac 5260. Mr Switcher tries it, and he's really impressed. He wants it. He asked the salesperson about the machines specs, and the conversation turns toward the guts of the machine. Mr Switcher is surprised to learn that virtually every component in the brand new PentiMac is identical to his Dell Fifth Dimension running LongHorn. Well, it's running when it's not defragging, and auto cleaning spyware, but thats another story.

Mr Switcher tells the salesperson, "Hey, I have the exact same equipment in my box right now, let me just buy Leopard and I'll be on my way." "Sorry" the salesman replies, "can't do that" "What?" Mr. Switcher stares at him dumbfounded, "Why not? Look, it's all the same stuff, the same specs. Why do I have to pay all of this extra money for equipment I already have at home?" What do you mean I can't just install the OS?"

"Well...." the Salesman begins, realizing it's going to be another long night.

Now, Apple says they are going to lock OS X down to Apple Macs, but if there isn't a performance benefit exclusive to the Mac platform in however they lock it down, or if whatever Intel that's inside isn't tuned in some special way to perform better with OS X, then I think they have a problem.

Right now Apple has defined the perception of the Mac with the OS, the advantages of the PPC architecture, and the unparalleled attractiveness and uniqueness of the form factors. Commodization of the hardware could seriously upset the balance of Mac persuasion power, at least regarding hardware.

Fair enough..but most of the switchers I have dealt with as a Student Rep know for a fact that they are switching to a machine that is slower and is more expensive than the PC's he could buy (laptops). They switch for Apple's amazing design and quality and for the operating system. Now Apple's systems will be on par (performance wise) with those of other companies. Im sorry, but I can't see how this will be anything but better for sales. Plus there is a good chance that macintels will be cheaper! My job is going to get a whole lot easier next year!! I do want Apple to lockdown OS X to macs and only let windows run in an emulator. This will give mac systems the value added of the best OS and will ensure that OS X development continues.

 

[shawnce]

As far as the iMac is concerned...how is this better then the G5 in the iMac right now? Power consumption doesn't really matter with the iMac. They can cool a G5 just fine in it and I would assume they could cool a 2.3Ghz G5 without any problems. I don't see the performance increase there especially when you start running programs on Rosetta. You are then taking Alti-Vec away from the system as well...plus the awesome bandwidth of the G5. Sure you may gain a dual core, but you loose a lot of the G5 benefits.

1) It will likely be dual core (or at least capable).
2) Power consumption does matter in the iMac and iMac mini.
3) It will likely outperform G5s while using less power (remains to be seen how much G5s continue to advance into 2006/2007).
4) It has decent bandwidth and a large L2 cache.
5) Rosetta won't likely get used much unless you don't want to update software (note developers can soon start shipping universal binaries even before Apple ships MacIntel systems).
6) ...2006 is only the beginning.

Is PPC a very good ISA, yes.
Is Altivec a great capability and better then SSE, yes.
Would I love to have a PPC based system that out competes Intel/AMD, yes.
Is the G5 currently relatively competitive, yes.
Has the G5 be stagnating for longer then expected, yes.
Is Apple concerned about the future, yes.

 

[jared_kipe]

People concerned about the Mhz myth and Apple's benchmarks are somewhat confused. The Mhz myth is very real. Intel, too, acknowledged it. That's why Pentium M chips run at lower Mhz numbers than Pentium 4's of comparable speeds. Apple was also completely correct in saying that the G5 had the more promising roadmap when it was released. But IBM never delivered the goods. Sometimes you have to have a spine. Apple did.

No kidding, remember all the rumors in 2003 about bringing out the Power 5 derivative to the mac, what happened to how good they looked?

 

[nagromme]

Look over this roadmap summary.

Thanks for the link! Time to get up to speed...

Maybe you should read my entire post, that is what i was saying..

I don't see that in your post, but I see we're on the same page 🙂

The form factors may look great, but if the only difference is the installed OS (and apps), I think it's a big problem.

Only if people find Windows acceptable and don't know about Tiger or Leopard. But the press has already woken up, and plenty of buyers are too. The OS is by FAR the biggest difference. And it's the most visible--it's what draws people to the Mac, not internal details.

And Apple doesn't need to conform to generic PC boxes. They can do what they like, same as always--ditching the floppy, adding WiFi ahead of the pack, etc. The CPU is but one (vital) component.

Also, Apple has always done well on specs comparisons beyond just the CPU--if you look at ALL the specs. Conveniently overlooking some features (FireWire 800 perhaps, or noise output, or fantastic software bundle) is a time-honored tradition in the field of pretending Macs cost more 🙂

 

[jiggie2g]

This Is going to be you're Next CPU.

1st I will say you need to worry , you will never see a P4 in a retail MacIntel PC
these are the most likey candidates. By 2006 the P4 will be EOD'd as even Celeron will be dual core.


Enjoy 😀

http://theinquirer.net/?article=23055

This Is what Jobs ment about Intel will have more units per watt , 70 for Intel vs. 15 for IBM.

 

[Furrybeagle]

Im in the same boat as EricNau, having bought my first Mac, and also loving it. 😀

What worries me though, is that Apple may start phasing out their own software from PowerPCs relatively soon. Would it be realistic to say that Leopard and all of Apple's other software will still be coming to PowerPCs until late 2007 or 2008? I mean I can't imagine them abonding all the people who own PowerPC hardware...

Unless they released a backwards compatible Rosetta, that converted Intel code to PowerPCs. But then we'd have the problem we have now with Altivec...

As long as my G4 stays in the loop until its time for a new machine. 😎

 

[nagromme]

Why no SSE3?

Speculation re Apple telling devs not to use SSE3 in their apps (if that's true).

Correct my misunderstandings please, but this sounds to me like either:

1. Apple doesn't see a need for SIMD in the future.

or

2. Apple knows something we don't about Intel's plans, and doesn't want to frustrate devs by optimizing for SSE3 when in fact Intel Macs will use something else--at which time optimizing for that WILL make sense, as a replacement for AltiVec.

#1 sounds unlikely, so I'm left with a mystery, but still hoping for the best.

 

[shawnce]

Im in the same boat as EricNau, having bought my first Mac, and also loving it. 😀

What worries me though, is that Apple may start phasing out their own software from PowerPCs relatively soon. Would it be realistic to say that Leopard and all of Apple's other software will still be coming to PowerPCs until late 2007 or 2008? I mean I can't imagine them abonding all the people who own PowerPC hardware...

I don't see Apple themselves winding much down on PPC software before well into 2008 or 2009. Of course new versions of software may require better horse power that only the MacIntel will be providing (PPC will not live much beyond 2007 in Macs) and hence set some limits not directly related to PPC/x86.

 

[shawnce]

Speculation re Apple telling devs not to use SSE3 in their apps (if that's true).

Not hearing those comments yet from folks I know... consider the fact that Darwin uses SSE when compiled for Intel already.

Of course Apple with things like the Accelerate framework and Core Image can buffer developers form changes.

 

[jiggie2g]

Memrom (Notebook 64-bit Intel CPU) , and Conroe (64-bit Desktop Intel CPU) are the Future for Intel.

Especially Conroe is said to be 12-13 stage CPU built from the ground up.

http://theinquirer.net/?article=23055

 

[iMeowbot]

I was just curious what is being done to the apple computers so that you can't install this os x on any regular pc.

There's a little chip embedded in the DVD that releases poisonous gases if you slide it into a non-Apple drive.

 

[nagromme]

Would it be realistic to say that Leopard and all of Apple's other software will still be coming to PowerPCs until late 2007 or 2008?

Very realistic! The work to put these apps on PPC is already done. Why abandon them? Apple worked in parallel for 5 years on OS X, and nobody even USED the second version. So working in parallel on OS X and apps for a while in the future is a no-brainer.

Consider:

* The number of PPC Macs--even recent ones--will exceed the number of Intel Macs for some time.

* Apple MAKES the tools for developing Mac apps--universal binaries that run on both PPC and Intel--so you can bet they'll use them.

* Apple needs to set that example for developers, not take a destructive shortcut in public.

* Apple needs to reassure users that it handles big transitions well.

* Leopard is coming in the MIDDLE of the Intel transition, not after it.

* Asking for Apple to support PPC Macs with new software updates into 2008 is asking them to continue only a year--or LESS--after they stop SELLING PPC Macs, and a PPC Mac OS (Leopard). That's not asking much. I'd say support will continue to some extent well into 2009 or 2010.

* This is a LOT of time for people to give feedback, and for Apple to see what users' needs really are.

Conclusion:

1. Current Macs have nothing to fear but fear itself 🙂

2. Intel Macs (not even sold for another 1-2 years) may have a challenge: is every little app you want going to be ported, or at least work OK with Rosetta?

It may sound like #2 will be a small issue by the time this all shakes out, and I agree. But choosing #1--a PowerPC Mac--actually sounds like the BETTER bet in some ways!

Not to mention... you CAN buy a PowerPC Mac, and Intel Macs don't exist yet...