Intel and PowerPC Mac OS X Up to Date?

[Lanbrown]

You are correct that the chips are not in production yet, but there are a number of samples floating around the industry right now. This allows Apple to handle most of the major design work now and simply wait for initial and full-scale production of the CPUs and chipsets. I fully believe that we will see an Intel-based Mac in the early to mid-April timeframe (essentially 2-3 months after Yonah enters production).

They are NOT going to wait until WWDC for this because the main focus of that conference next year will be Leopard (and how it will dominate Vista).

Hickman

Keep in mind Intel has a history of announcing chips and shipping them in low numbers just to meet a timeline. They are also not doing well on their Performance per Watt strategy either. Their current dual core is a power hog and a stopgap measure to try to compete with what AMD has. Then we have Intel planning to use FB-DIMM’s, which consumes more power over DDR or DDR2. Intel is already saying one thing and doing another in terms of power per watt. Their switch to 90-nm was also a colossal failure. IBM had issues as well, but there were others in the market place that didn't have those issues.

 

[wdlove]

Did Steve Jobs wait 3 years after OS X before he declared OS 9 dead? With that as an example of his timing we have less than 3 years before we have to depend on 3rd party only to support OS 9 & X on the PPC..

Bill the TaxMan

I'm optimistic that with so many PPC Mac's in use that Steve will continue support. At least 3 years if not longer.

 

[BenRoethig]

Keep in mind Intel has a history of announcing chips and shipping them in low numbers just to meet a timeline. They are also not doing well on their Performance per Watt strategy either. Their current dual core is a power hog and a stopgap measure to try to compete with what AMD has. Then we have Intel planning to use FB-DIMM’s, which consumes more power over DDR or DDR2. Intel is already saying one thing and doing another in terms of power per watt. Their switch to 90-nm was also a colossal failure. IBM had issues as well, but there were others in the market place that didn't have those issues.

I look at it this way, AMD may have the better processors at the moment, but intel has the resources to make the move much easier. Moving to AMD or another x86 is simple once everything is ported.

 

[PureMac]

The REAL thing

A friend of mine has just downloaded the latest Torrent MacOS x86 version 10.4.3 and cracked it in his Pentium 4 PC.
Check this out:
http://img496.imageshack.us/img496/4958/dsc93098ce3na.jpg
http://img496.imageshack.us/img496/7550/11201469981uj.jpg

He says it runs smooth and fast. Moreover, he was able to install PC apps too

I'll add some more pics later on

 

[BenRoethig]

Like I've said on many occasions, it's going to be next to impossible to keep OS X to themselves.

 

[cgc]

Like I've said on many occasions, it's going to be next to impossible to keep OS X to themselves.

I think Apple realizes this. I bet significantly more money could be earned by Apple allowing OS X to run on certain systems. The one thing Apple has that other PC manufacturers don't have is control over hardware and software. This would be my only concern with 3rd party PCs running OS X - would they be as integrated? To solve this issue, Apple needs to lay down specifications that, if followed, would yield an "Apple Certified" stamp of approval...or something. That way, Apple would still have sme control over the quality of machine OS X is installed (officially) on.

To think Apple expects OS X to remain Mac only is ludicrous.

 

[generik]

Can the currrent dev builds run current PPC binaries?

 

[Bulb]

There is a lot to be gained from the move to x86.. Sheer speed increases will be great. What i hope really comes from this is stability for the platform. There is too much fundamental change in Mac land and not enough steady increases.. I find the changes are a total mess to try and negotiate, stuff like rosetta to me just feels like a total hack.. This move needs to happen so that within a month of release 100% of all software is available as native x86 and all as a free or minimal cost upgrade from PPC versions, then you got yourself a transition.. I know the politics would make something like this hard to achieve but if they extend themselve enough and really make their best efforts then it will get everything off on the right foot. Nothing will dog the momentum more than stuff being incompatible or needing some half assed emulator to try and run.

The worst part about something like rosetta is it will never be great because its a makshift tool with a conflict of interest. Apple wont put real effort into it because from day 1 its never intended to last. The conflict of interest is that if does work well then its less motivation for people to consider upgrading.. They will just throw it together to look like theyve made an effort to all the poor bastards who have spent thousands of dollars on what is rapidly becoming obsolete..

 

[driftingaway]

Can the currrent dev builds run current PPC binaries?

OS X x86 runs them through a platform called Rosetta. This translates current PPC binaries to x86 with only a minor performance drop.
However it can't access things like OpenGL at the moment, which means most Pro Apps won't run.

 

[BenRoethig]

OS X x86 runs them through a platform called Rosetta. This translates current PPC binaries to x86 with only a minor performance drop.
However it can't access things like OpenGL at the moment, which means most Pro Apps won't run.

The problem will be that everything will emulated as a G3 class processor. So anything designed altivec optimization is not going to be able to use it.

 

[Marky_Mark]

I find the changes are a total mess to try and negotiate, stuff like rosetta to me just feels like a total hack...
This move needs to happen so that within a month of release 100% of all software is available as native x86...
The worst part about something like rosetta is it will never be great because its a makshift tool with a conflict of interest...
Apple won't put real effort into it because from day 1 its never intended to last...
The conflict of interest is that if does work well then its less motivation for people to consider upgrading...
They will just throw it together...

I think you need to give the guys at Apple a little more credit here. I remember the transition from 680x0 to Power PC and the emulation software was very stable, fast enough for all but the highest of high-end apps and made the whole exercise unique in the scale and smoothness of the transition.

A lot of people back then shared your doubts and reservations before the event, only to be full of praise at the skill and professionalism that was actually demonstrated. The emulation engine was so good that Apple were able to concentrate on making only the most mission-critical aspects of the OS native early on, and the final elements of 68K code persisted until System 8.6, five years after the introduction of the first PPC-based machines.

I see no difference here, and I would suggest that you reserve judgement and be prepared to show credit where it's due next summer.

 

[Bulb]

Ok, i will wait and see.. I would still like to see the need for these things to disappear altogether. I hope x86 becomes a stable architecture for many years to come. If you want a kickass system the last thing you need in the equation is emulation software.

 

[dguisinger]

Keep in mind Intel has a history of announcing chips and shipping them in low numbers just to meet a timeline. They are also not doing well on their Performance per Watt strategy either. Their current dual core is a power hog and a stopgap measure to try to compete with what AMD has. Then we have Intel planning to use FB-DIMM’s, which consumes more power over DDR or DDR2. Intel is already saying one thing and doing another in terms of power per watt. Their switch to 90-nm was also a colossal failure. IBM had issues as well, but there were others in the market place that didn't have those issues.

Keep in mind, you obviously don't follow Intel too well since you'd know the processors Apple are planning to use are based off their successor to the Pentium M, considered to be one of the most power efficient as well as highest-performance per MHz chip Intel has ever made. It runs circles around the Pentium 4 in both categories, and of course their current dual-core chips are based on the current generation of Pentium 4 and Xeon processors; which are of course, from the failed paradigm Intel had where faster MHz would move more sales.....designing for MHz only -- it eventually caught up to them.

The Pentium M has not suffered from the issues you refer to, and infact is a wonderful processor. Next time research before making your claims that Intel processors suck. Some have been good, some haven't been that awsome, but in reality, they've never made a horrible chip. Infact, even today, I have yet to find an Intel motherboard with an Intel CPU and chipset that is horrible when it comes to stability. On the other hand, experience in life has shown me that AMD Athlons, which typically have the AMD CPU, 3rd party chipset, and another 3rd party motherboard......quite often have stability problems. So while you may not agree with everything Intel does; there is a reason IT has stuck behind Intel instead of moving to AMD. Most high-end workstations, including some from Dell, come with genuine Intel motherboards...

Intel has built their business around stability, much like Apple has built theirs around it. Obviously Intel cannot guarentee the stability of the OS as they don't make it, but they do a damn good job on the parts of the system they design at making sure they work together. I'm typing this on a Pentium 4 which is 5 years old. May not be the best performing system in the world; but the only thing thats ever failed is the HD because I kick the box when I'm mad. Not bad, sure its running old RDRAM, but you know what? It performs well, I don't care what memory it uses. I've never had stability issues from heat; I've never had stability issues period. Why? Because even though its a Dell, the motherboard says "Intel", the chipsets say "Intel", and the CPU says "Intel".

So while I hate the Pentium 4 processor design, all the excess pipeline stages that are there just to move instructions from one side of the chip to the other, the extra power it uses, the extra heat it makes: I still trust it, I'd still buy it over AMD, and my experience with the Pentium M in my other machine says "Hold on to your butts" because you won't be upset with the performance you are about to see in future Macs with upcoming Intel CPUs.

 

[projectle]

What it all comes down to is OSX for x86 is still very new. It can not be judged off of the impressions of it on random unsupported hardware configurations.

What you are doing is comparing apples to rocks... sometimes they are simmilarly shaped.

Try using any of Microsoft's development operating systems and you will see it blow monkey turds every chance it gets. That being the case, Apple is doing pretty well. Sure you can not play UT2004 or whatever you want to on it, but as a reminder, you shouldn't even have it right now anyways.

Compare Vista to Tiger x86 and you will see something really spectacular...
Tiger reliably runs and is not a resource hog, demanding an ultra-bleeding edge system just to power up. (Sure, there is the dependance on SSE3, however the resources required to run and perform the same types of tasks is significantly less).

Now, to qualify that statement, I am a software developer designing database appliactions for telecommunications companies, and thus have access to both a legitimate Apple Transitional Developer Kit as well as the current test version of Microsoft's Vista.

Just sit at the desktop and see the CPU + Memory requirements to idle under a clean installation.

OSX 10.4.2 (x86)
Average CPU Utilization: 1%
Average Memory Utilization: 165 MB

Vista (5203)
Average CPU Utilization: 1%
Average Memory Utilization: 211 MB

While this is not a "major" difference, it is worth noting that even on things such as web browsing, vista will crash (Explorer.exe halts and needs to be manually restarted through the task manager) rather than prompting that one needs to install a plugin to view content, at the moment.

So, stability at the moment is a real plus for OSX.

 

[slooksterPSV]

While this is not a "major" difference, it is worth noting that even on things such as web browsing, vista will crash (Explorer.exe halts and needs to be manually restarted through the task manager) rather than prompting that one needs to install a plugin to view content, at the moment.

So, stability at the moment is a real plus for OSX.

How long have they been working on Vista? How long have they been working on OSX x86? - Seriously I'm curious. If Vista has been worked on longer, then I'm not surprised at stability. OSX just works! YAY!!!

 

[projectle]

Development of Vista officially started July 2001.
First development release September 2003.
Initially to be released Fourth Quarter 2003.
Actual release date Third Quarter 2006.

Development of OSX for x86 is unknown.
First development release June 2005.
Initially to be released, never.
Actual release date Second Quarter 2006.

All we really have is a recent assurance by Apple that it has been working to maintain concurrent releases for the PPC and x86 Architectures since the initial release of OSX. So, if they are being truthful, Marklar's development started on or about March 24, 2001.

 

[50548]

After reading some people's posts saying that the G4 is "crap" or "dead slow" in comparison to the Pentium or Centrino, I wonder why, in real life, my browsing experience, as well as PDF manipulation, are waaaaaay faster and more stable on my "old" iBook G3 than most modern crappy PCs, with their lousy multitasking and virtual memory implementation...it's amazing to see how much hard disk activity can be heard when a PC is working...in Macs such thing is a breeze...that's why I prefer to get REAL productivity and performance, not spec numbers...

PCs are for DirectX gaming, NOTHING else...period.

 

[driftingaway]

The problem will be that everything will emulated as a G3 class processor. So anything designed altivec optimization is not going to be able to use it.

Not neccesarily. Rosetta converts Altivec instructions to SSE2/SSE3 instructions on the fly.

 

[shamino]

Yonah and Merom (Apprently the first Intel chips) are they 64-bit?

We don't know what chips Apple will use. Yonah and Merom are speculation.

Right now, Intel ships both 32- and 64-bit x86 chips. 64-bit is only in the most expensive models (Xeons), but is expected to be much more pervasive in the near future.

I'm pretty sure (without having seen anything official) is that Apple will replace today's G4 systems with 32-bit Intel chips. They will not replace any G5 system until the appropriate 64-bit Intel chips are available.

Those who are afraid that this will be the end opf 64-bit computing on the Mac platform are getting scared for nothing.

I would be more happy with Apple going to AMD actually.

Won't happen? Did you see the announcements from Intel about low power chips? AMD doesn't have anything close. (And I'm sure they'd be making press releases if they had something coming soon, because this is going to impact their sales bit time once those low-power chips actually ship.)

AMD and Intel are almost equally matched in terms of raw computing power. The winner in this war is the one that can do it for less watts. (Price has little to do with it either, since volume purchasers always get huge discounts.)

As for actualy reason for the post - I dont think OSx86 can be at 10.4.3 - just a feeling. Its like as if there's enough people complaining about their PPC chips running after the upgrade, let alone the x86 people!

There are always a lot of complaints following every system update. What I'm reading about 10.4.3 is nothing unexpected. Comparing the number and type of complaints, it's a heck of a lot better than the 10.4, 10.4.1 and 10.4.2 releases.

 

[shamino]

By the end of 2007 there is projected to be a quad core single chip (no northbridge/southbridge) PowerPC available with low enough power to put into a powerbook. I wonder if some of us will be looking enviously at that chip at that time.

http://www.pasemi.com/processors/

From a completely unknown vendor? You'd want Apple to stake their future on that?

 

[shamino]

I respectfully disagree. Many mac users don't need their computers to do more than basic tasks that Tiger, iLife, and iWork will handle.

And these people don't have any immediate need to get an 8x6 box either.

I'm under the inpression that there will be no new PPC macs released.

And this impression is based on what?

There are many who won't buy new Macs until Intel, myself included.

That's your decision. If you seriously think the PPC Mac will become useless as soon as the x86 boxes ship, I think you're woefully misguided.

I can't afford to buy a new computer every 3 years, I have to make 'em last, and the only way to do that is to buy the "next great thing" right after they release it. My iMac's great, don't get me wrong. But at 800mhz, It's starting to show it's age. Personally, I'd like to make it last until the Intel swich.

I fail to see your point here.

The next great thing was recently released - new iMac/G5, new PowerMac, new PowerBook. They'll serve you very nicely for quite a long time.

I don't think you're going to find all the PPC software disappearing before you get your next upgrade after that (at least 4-5 years from now.)

My PowerMac at home (a dual 1GHz system) is approaching four years old. It still works great. I expect to get several more years out of it.

As for your existing iMac, you may find that a memory upgrade may give it a new lease on life. Especially if you've only got 256M of RAM in there. If your hard drive is close to full, deleting stuff (perhaps transferring it to DVD or an external hard drive) or a bigger drive will help enormously. (A friend of mine got a huge performance boost after deleting 20G of iMovie projects from his PowerBook's drive.)

 

[shamino]

Did Steve Jobs wait 3 years after OS X before he declared OS 9 dead?

Bit despite his declaration, every Mac sold today can still run the apps. Classic won't be gone until the PPC is fully gone. Which means at least two years from now - far beyond the date of that declaration.

With that as an example of his timing we have less than 3 years before we have to depend on 3rd party only to support OS 9 & X on the PPC..

It was four years from the introduction of PPC Macs until they shipped the first OS release incompatible with the 68K platform. And the apps remain compatible to this very day. (And won't break until the PPC goes away.)

So which precedent will you choose to believe?

 

[shamino]

Can the currrent dev builds run current PPC binaries?

Yes. OS X binaries (Carbon and Cocoa) will run through Rosetta (a PPC emulation layer.)

Pre-Carbon apps (stuff that runs in the Classic environment today) won't run on the Intel platform. Additionally, code that directly accesses the AltiVec PPC instructions will need to be modified. (But apps that use Apple's system frameworks to access AltiVec will not need modification - they'll just start using Intel SSE instructions.)

 

[shamino]

The problem will be that everything will emulated as a G3 class processor. So anything designed altivec optimization is not going to be able to use it.

You're only telling half the story.

Code that uses the Accelerate framework (introduced in Mac OS 10.3) for accessing AltiVec won't need any changes. The framework has been ported to x86, allowing applications to seamlessly switch over to SSE instructions.

 

[shamino]

Not neccesarily. Rosetta converts Altivec instructions to SSE2/SSE3 instructions on the fly.

No. It does not. Apple was very specific about this.

Code that directly makes Altivec instruction calls will need to be updated. Code that accesses AltiVec through a system framework (like Accelerate, or CoreImage, etc.) will not need updating, because Apple has developed x86 versions of all the system frameworks.