Why Apple Are In No Hurry To Build A New Mac Pro

[handsome pete]

If it is, why didn't Intel update their Xeon lineup?

Since when has Intel stopped updating the Xeons?

The F-150 analogy doesn't make any sense. Pick-up trucks aren't bought for the purpose of going fast.

I actually think that's an apt comparison. I wouldn't think of workstation processing power in terms of speed as much as I would heavy lifting. After all, the workstation chips have slower clock speeds for the most part but they offer more cores/threads. So I wouldn't think of it as a sports car getting someone faster from point A to B. I would think of it more as a heavy duty pickup that can carry a large amount of cargo from A to B instead.




Machines like the iMac and HP Z1 will eat into workstation sales simply because they'll be powerful enough for the work that certain users never really needed a full blown workstation for in the first place. So they'll definitely cross over in the low end. But I don't see it encroaching on the high end user market significantly. And personally, I still like the freedom of choosing my monitors, being able to easily service parts, and have everything in one box rather than multiple pieces strewn around my desk.

 

[thekev]

It gets more work done in the same amount of time.


Sure (at least for the next 18-24 months), but the iMac's performance is on par with a 3.33GHz Hexa-Core Mac Pro already, and that adds $500 to the calculation I made some posts ago based on the low-end Mac Pro. The iMac is perfectly capable of replacing single-processor personal workstations. Especially, as HP puts it, "on similar or shrinking budgets".

It depends whether all cores are stressed. The 6 core can still pull out ahead, but the current gap may not be worth the difference in dollars. Even refurbished at $2500 it seems expensive for the age of the hardware.

The multi-processor personal workstation market boils down to 3D animations and video processing.
Is that market large enough to maintain the Mac Pro as a stand-alone product? If it is, why didn't Intel update their Xeon lineup?

They did. It was called Sandy Bridge E/EP. It came out last year after multiple delays. Apple just didn't use it. The lack of updated gpus along with just reshuffling nehalem/westmere cpus suggested that they hadn't done the work necessary for an update. The Ivy Bridge equivalent comes out later this year. The max core count for Sandy is 8. Ivy goes as high as 12, so in dual configurations 16 and 24. I wouldn't fall for the something big must be coming mentality. Other products that have languished prior to a major upgrade didn't remain unchanged this long.

If you want to own a Lamborghini, that's what you should be buying then. As far as computers are concerned, they either shrink in size or get more powerful. The iMac primarily does the latter while the Mac Pro currently does the former. It's not my idea to call the product of that process iMac or Mac Mini (which, by the way, always happens to be as powerful as the fastest Mac Pro from three years ago at less than 1/3 the size and 1/3 the price).

Car analogies really don't follow here until someone adds a sway bar to the mac pro.

 

[deconstruct60]

Since when has Intel stopped updating the Xeons?

The whole Xeon line up ? They didn't. For what is now the Xeon E5 ( was 3500-3600 and 5000 series )? They pragmatically did in 2011. That was a year of no significant upgrades ( at least for workstations). For Xeon E7 Intel completely skipped Sandy Bridge micro-architecture to get to the process shrink.


Partially that was due to some hiccups on transitioning to PCI-e v3.0 (which evolved slower than projected launch roadmaps) and some bugs in chipset implementations. However, also partially because the market was weak (both in x86 core competition for Intel and on the demand side. )

The Xeon E5 and E7 upgrade cadence has been slowed. The Xeon E5 announced in March '12 (and shipped in systems in May '12 ). The Xeon E5 v2 isn't going to show unitil Aug '13 (and if same slide in systems Sept-Oct ). Xeon E5's are shipping v2 ("Ivy Bridge") after the mainstream line up is already on v3 ( "Haswell" )

I actually think that's an apt comparison. I wouldn't think of workstation processing power in terms of speed as much as I would heavy lifting.

It can be speed but it is speed under a bandwidth load. Attach a 6,000 lbs trailer to a F-150 and a Lamborghini and see who wins the drag race. The PCI-e bay and the F-150's cargo bay are a bitter analogical fit than the non cargo capacity of a Lamborghini. Likewise being hitched to SAN/NAS , A/V capture, external I/O on Mac Pro side and a trailer / hauling / pull out of a ditch capabilities of a F-150.

Now there are posers who buy both F-150 and Lamborngini that have no real need for the capabilities and far more so for crotch grabbing bragging rights. Either one of those is a match to the posers who buy Mac Pro for same reason.

Machines like the iMac and HP Z1 will eat into workstation sales simply because they'll be powerful enough for the work that certain users never really needed a full blown workstation for in the first place.

That isn't really true. While "640K ought to be enough for anybody" is famously used to promote the notion that no limits should be put on specs users needs, the reality is that increases in performance can outpace the increases in user workloads. The constant onslaught of performance improvements have pushed past many folks. Including some of those who formerly needed a higher end workstation to get their work done.

And iMac and HP Z1 are capable of doing better than a TFLOP in computations with the proper software. That outclasses some supercomputers from the turn of the century and many workstations from several years ago. To claim that many folks can't do significant work on machines with that level of performance is bogus. They did a decade ago and it is more than possible now.

So they'll definitely cross over in the low end. But I don't see it encroaching on the high end user market significantly.

Looking in the wrong place to see the encroachment. It isn't the systems, it is the workloads. Relatively slow growth (or non growth) workloads will be encroached on by these systems.

And personally, I still like the freedom of choosing my monitors, being able to easily service parts, and have everything in one box rather than multiple pieces strewn around my desk.

'Form over function' is in the same boat as '640 otta be enough for everybody'.

 

[sash]

You know, 5 years from now we'll still be discussing this topic. Apple and other manufacturers will keep building workstations until we'll need this technology -- or maybe better this power.

 

[deconstruct60]

. The max core count for Sandy is 8. Ivy goes as high as 12, so in dual configurations 16 and 24. I wouldn't fall for the something big must be coming mentality. Other products that have languished prior to a major upgrade didn't remain unchanged this long.

We'll see when Intel releases the E5 v2 product line, but it is very likely that either the 12 core count models won't initially be there ( will come later ) and/or will be priced in the "if have to ask how much it costs, you can't afford it" zone. Either way it is doubtful it would ever show up in a Mac Pro.

The initial projections out of Intel was for 10 cores for Ivy Bridge. I wouldn't be surprised if there was a E5 4600 with 12 cores, but a E5 2600 could be done if the process technology is mature enough for typical 2600 like pricing.

The 6 , 6 , 8 core line up Apple could have rolled out with Xeon E5 2600 ("Sandy Bridge") is still pretty likely the same core count for Xeon E5 2600 v2 ("Ivy Bridge") if Intel's pricing doesn't significantly move down for 2600 v2 class offerings. Same core count clocked higher but at the same price will likely make more folks happier than cranking core count at lower clock speeds and higher prices.

The Xeon E5 1600 series goes

1620 4 core 3.6 GHz
1650 6 core 3.2 GHz
1660 6 core 3.3 GHz

while

2620 6 core 2.0 GHz
2640 6 core 2.5 GHz
2665 8 core 2.4 GHz

so the 2600 is traiing by almost a whole 1 GHz. I think in a workstation context there will be more demand to close that gap than to crank up the core count. The v2 1600 series are all getting bumped about 0.2 GHz (and no core count change at all ) so that would mean either widening the gap or cranking up cores.

 

[fastlanephil]

Has this little tidbit been posted yet? It sounds plausible.

The rack mount form factor makes some sense but I think both an upright and rack mount capability makes more sense.

http://www.ear.net/products/new-apple-mac-pro/

 

[deconstruct60]

You know, 5 years from now we'll still be discussing this topic. Apple and other manufacturers will keep building workstations until we'll need this technology -- or maybe better this power.

All vendors being constant has never been true in the computer industry. Some vendors drop out over time. In this segment, Apple could be one of them. Apple doesn't "have to" stay in this market. If viable they don't "have to" quit either but 5 years projection it is smoke as to what state ( growing , shrinking, etc) that this market will be in .

Baring Intel, AMD, and Nvidia all hitting some unexpected huge barrier 5 years from now Mac mini's and iMacs will be very substantially faster machines. Storage technologies will likely be very different. Memory heirachies will likely be very different (e.g., the eRAM being added to Haswell offerings this year. That's a new trend that will be mainstream in 3-4 years. )

The "box with slots" form factor driven by the needs of technology demands of the 1980s may be as aligned with technology of 2018 as you think.

 

[deconstruct60]

Has this little tidbit been posted yet? It sounds plausible.

First it doesn't sound accurate. The current Mac Pro has 4 slots not 3. So the claim that there is only 2 is dubious since have demonstrated an inability to count. That it is actually 3 is more believable. Pragmatically required Thunderbolt infrastructure was likely to "consume and reallocate" at least one one slot worth of bandwidth. That could easily result in a whole physical slot (e.g., a fixed x16 slot assigned to embedded GPU).

The rack mount form factor makes some sense but I think both an upright and rack mount capability makes more sense.

Again "rack mount form factor" is a dubious assertion. If the "desktop" version is shrunk by 2/3 that is more than sufficient to make it non-rack hostile. It is really primarily only the gratuitous handle height that make the current rack hostile. No radical changes were needed.


A box 2/3 the size runs in opposition of this being a dual E5 2600 box though. This speculating that Thunderbolt gives some magical box size shrinkage is more than dubious. It doesn't and won't. 2/3 the size but still delivering 16 cores is likely wrong.


If Apple capped themselves solely to a E5 1600 solution (or dropped down to Xeon E3 ) then are repackaging and design adjustments that could remove 1/3 of the Mac Pro. Capped at 4 DIMMs slots and lower TDP they could carve space out of the current CPU thermal zone. More than likely the ODDs are also completely tossed. That would free up moving storage drive trays up into that space (in addition to flipping most of them to 2.5" ) and removing drive sleds from PCI-e thermal zone. Yet another shrink. One or two PCI-e slots and even more space out of the PCI-e thermal zone.

If Apple dropped down to an Xeon E3 (or mainstream desktop i7. i.e., iMac class part) then 2 total PCI-e slots would be believable since really only have about 20 PCI-e lanes to allocate anyway ( 16 + 4 and done. )

If Apple drops down to mainstream desktop Intel CPU package offerings ( either E3 or iMac class i7's ) , then I do think the will suffer huge backlash though as this not being Mac Pro upgrade. Comparatively speaking they would have gutted the I/O bandwidth and the performance range. They can try to sprinkle all the Thunderbolt kool-aid on top of they want to but will be lots of solutions they will be tossing under the bus.

 

[deconstruct60]

but I think both an upright and rack mount capability makes more sense.

For customers two uprights would make more sense.

1. this 2/3 smaller box. If there was a significant blocker for mixing E5 1600 and Thunderbolt, then use an E3 v3 solution. Fewer slots and more 2.5" drive sleds along with sub $2,000 entry (maybe mid-range) price points.

2. Same volume box with most current internal lay out , perhaps no Thunderbolt, and dual Xeon E5 2600 only. Shorter but deeper would get rid of the gratuitous handle height. three x16 and one x8 PCI-e v3.0 slots


Both would be rackable for the subset that need to place the box in a rack but there is very little need for a box that would be primarily oriented to being racked.


It is plausible though that Apple is fully guzzling Thunderbolt kool-aid and that there is faction who want to force to new desktop onto the actual desktop/tabletop so that "tail wags the dog" Thunderbolt Display will work with its fixed length cable. Reusing iMac logic board components and R&D probably has more than a few fans too ( since it helps goose profit margins ). Apple would deliver most of the xMac but not at expected xMac prices.

The new "Mac Box" would be capped under $3,000 in base configurations. So Apple wouldn't have to worry about long term viability of the > $3,000 PC market.

 

[gpzjock]

Supply and demand, chicken and egg.

I suppose you could look at it that way, or you could take another perspective in which the Mac Pro isn't selling well, because prospective buyers don't see the value in purchasing the overpriced, legacy hardware in the Mac Pro currently been offered. With Apple focusing all of their attention on iOS and it's respective hardware these past few years, there's no wonder that revenues from this Pro segment are down. I also can't blame, as they are facing some new competition from Samsung and Google in this market that they created.

I guarantee you that revenues for the Mac Pro would soar in this current and next quarters if Apple were to announce re-designed, cutting-edge Mac Pro hardware available for purchase today. But you can't grow a segment if you completely dismiss it for well over two to three years.

After 5 years of using a 2008 Mac Pro octo I looked at the options for upgrade/replacement, due to I needing to give my GF more time on the original MP for her art and design projects. I was faced with £3500 for a Hex core 2012 Mac Pro from Apple with the things I wanted in it or about £2200ish if I bought a base unit and filled it with stuff myself. Then Apple pulled the whole Mac Pro line in Europe because a some fool could possibly get their fingers caught in a running fan without need of a screwdriver.
So, Mac Pro availability is slim to none and slim just left town. I solved this problem before they pulled the range. £1200 of PC parts and a copy of Mountain Lion later I have my upgrade Hackintosh and it works just fine. Apple will have to try very hard in the future to convince me that their overpriced workstations are worth reinvesting in the future.
I sincerely hope they do release a new Mac Pro but I stopped holding my breath back in January 2013, hope you aren't holding yours beyond August.

 

[calaverasgrande]

That isn't really true. While "640K ought to be enough for anybody" is famously used to promote the notion that no limits should be put on specs users needs, the reality is that increases in performance can outpace the increases in user workloads. The constant onslaught of performance improvements have pushed past many folks. Including some of those who formerly needed a higher end workstation to get their work done.

And iMac and HP Z1 are capable of doing better than a TFLOP in computations with the proper software. That outclasses some supercomputers from the turn of the century and many workstations from several years ago. To claim that many folks can't do significant work on machines with that level of performance is bogus. They did a decade ago and it is more than possible now.

History has shown that no matter how much power/ram/storage we get under the hood, the developers of the applications we use (and the OS) take it away. They add in more eyecandy that eats up our cpu cycles. They add "features" we didn't need in the first place, and in shoehorning those features in they move the stuff we used to the side.
The trick is balancing the power of a newer machine, with applications that had lower expectations for hardware. Though of course the way Apple has been deprecating code lately it's hard to use any application more than two years old on ML.

 

[MyMac1976]

History has shown that no matter how much power/ram/storage we get under the hood, the developers of the applications we use (and the OS) take it away. They add in more eyecandy that eats up our cpu cycles. They add "features" we didn't need in the first place, and in shoehorning those features in they move the stuff we used to the side.
The trick is balancing the power of a newer machine, with applications that had lower expectations for hardware. Though of course the way Apple has been deprecating code lately it's hard to use any application more than two years old on ML.

IMHO (I apologize in advance) that coders today are either lazy or have no sense or art and finesse.Most programs today feel very hammer like no dancing or lightness.

 

[tomvos]

IMHO (I apologize in advance) that coders today are either lazy or have no sense or art and finesse.Most programs today feel very hammer like no dancing or lightness.

I beg to differ. Before Mac OS X 10.2 the display server used the CPU for compositing the image. Everybody who ever used 10.1.x will remember the pretty GUI which was just a tad slow.

Offloading most parts of the image composition on the GPU via OpenGL kept the elegance and made the GUI much more fluid than any other competing GUI on other platforms.
Some may call this using a hammer to get the desired effect — at least you need a little bit of craziness — to throw all the power of OpenGL just to do what mainly seems to be some simple 2D image compositing.

To be honest, in the recent releases of Mac OS X this kind of cleverness seems to be absent … but then, my impression at the moment is, that Mac OS X (and iOS) is not the platform that leads technical innovation. Apple seems to be content with making a sh..load of money and does not bother about technical leadership.

 

[goMac]

History has shown that no matter how much power/ram/storage we get under the hood, the developers of the applications we use (and the OS) take it away. They add in more eyecandy that eats up our cpu cycles. They add "features" we didn't need in the first place, and in shoehorning those features in they move the stuff we used to the side.
The trick is balancing the power of a newer machine, with applications that had lower expectations for hardware. Though of course the way Apple has been deprecating code lately it's hard to use any application more than two years old on ML.

I have to disagree with that as well. OS X has become more and more optimized (and in many cases the drawing subsystem now is faster than OS 9's.) My Mac Pro is also sitting under <%10 CPU usage total right now.

If the code is meant for mobile devices, that also changes things as well. An iPhone 4 has an 800 mhz processor. A developer isn't going to even be able to write code that could come anywhere near to maxing out a Mac Pro if they're sharing code between platforms.

 

[Anon Tobin]

I actually think that's an apt comparison. I wouldn't think of workstation processing power in terms of speed as much as I would heavy lifting. After all, the workstation chips have slower clock speeds for the most part but they offer more cores/threads. So I wouldn't think of it as a sports car getting someone faster from point A to B. I would think of it more as a heavy duty pickup that can carry a large amount of cargo from A to B instead.

Not all workstations have lower clock speeds.

A 3.4 i7 isn't going to have any dramatic performance boost over a 3.3 Xeon unless you overclock it.

I tend to look at rendering as a formula one race rather than hauling equipment. But I can see your point.

 

[KaraH]

And iMac and HP Z1 are capable of doing better than a TFLOP in computations with the proper software. That outclasses some supercomputers from the turn of the century and many workstations from several years ago. To claim that many folks can't do significant work on machines with that level of performance is bogus. They did a decade ago and it is more than possible now.

The iMac is great for a lot of applications but because it stresses form I claim the right to do so as well. I do not want nor need the form of a monitor glued to my CPU. I do not want nor need the form of only having one drive bay. I do not want nor need a desktop that is more an appliance than a computer. If sheer power was my sole criterion I would have bought a 2012 iMac and maxed it out, but I am getting a computer.

IMHO (I apologize in advance) that coders today are either lazy or have no sense or art and finesse.Most programs today feel very hammer like no dancing or lightness.

I have that reaction as well. When you have coded at how few bytes possible (using logic that broke quite a few rules) the application bloat we have recently is jarring.

 

[calaverasgrande]

I have to disagree with that as well. OS X has become more and more optimized (and in many cases the drawing subsystem now is faster than OS 9's.) My Mac Pro is also sitting under <%10 CPU usage total right now.

If the code is meant for mobile devices, that also changes things as well. An iPhone 4 has an 800 mhz processor. A developer isn't going to even be able to write code that could come anywhere near to maxing out a Mac Pro if they're sharing code between platforms.

The code in OS X is horrible.
Why does it hourglass on the wifi applet in the menu bar?
It does this on my 2010 MBP and my 2013 rMBP as well as pretty much every other recent Mac laptop I've glanced at.
That is such a simple elemental task to anyone using a Mac. It should be bulletproof, especially with desktop composition offloaded to a GPU!
That is just one minor example, but there has been a noticeable bugginess in OS X the last few revisions. All of the more astute Mac users I work with/know have been noticing this and mention it. Another one is the strange lagginess in the UI when the CPU (or GPU) is positively not under any load.
There are a lot of "internet folk remedies" for fixing these glitches. Such as changing Downloads into a folder, or turning off other features of finder.
Quite a few users I know think that when they replaced the old G4/G5 code in OS X they didn't do a very good job. As if they didn't really QA it at all.

 

[deconstruct60]

The iMac is great for a lot of applications but because it stresses form I claim the right to do so as well.

The iMac doesn't stress form so you are out there on that "form over function" island by yourself.

If the iMac were a 'slave' to form it would still either looking like Luxo Jr or the same rectangular box it has for most of its lifetime over the last 7-8 years. It doesn't.

When the iMac didn't need flat sides to deliver a DVD slot anymore, it dropped them. That is function over form.

I do not want nor need the form of a monitor glued to my CPU.

The HP Z1 isn't glued. This is an implementation issue. It is neither a form or function one.

I do not want nor need the form of only having one drive bay.

Given the iMac can take two drives this really shouldn't be a problem. Again fixated on a specific form rather than function.

I do not want nor need a desktop that is more an appliance than a computer. If sheer power was my sole criterion I would have bought a 2012 iMac and maxed it out, but I am getting a computer.

A computer is primary a tool. In that sense, it is an appliance. The Mac has always been oriented to a "computer that just worked" not as the foundation for an erector set. A "real computer" is not an erector set or tweaker lab , etc.

I'm sure there is a limited set of folks who like "big American car with chrome bumpers and vertical tail fins" but that isn't the definition of a car.











I have that reaction as well. When you have coded at how few bytes possible (using logic that broke quite a few rules) the application bloat we have recently is jarring.[/QUOTE]

 

[goMac]

The code in OS X is horrible.
Why does it hourglass on the wifi applet in the menu bar?
It does this on my 2010 MBP and my 2013 rMBP as well as pretty much every other recent Mac laptop I've glanced at.
That is such a simple elemental task to anyone using a Mac. It should be bulletproof, especially with desktop composition offloaded to a GPU!

You're assuming the beachball is drawing related. It's likely not. The operating system has to go talk to an external card that has to do a radio frequency burst to look for local wifi networks. Sure, they probably could have threaded that. But it's not burning GPU or CPU cycles.

That is just one minor example, but there has been a noticeable bugginess in OS X the last few revisions. All of the more astute Mac users I work with/know have been noticing this and mention it. Another one is the strange lagginess in the UI when the CPU (or GPU) is positively not under any load.

Again, lagginess in the UI does can imply a lot of things. Slow disks spinning up, or other slow hardware.

There are a lot of "internet folk remedies" for fixing these glitches. Such as changing Downloads into a folder, or turning off other features of finder.
Quite a few users I know think that when they replaced the old G4/G5 code in OS X they didn't do a very good job. As if they didn't really QA it at all.

Replace? The Intel code is older than the PowerPC code. The PowerPC code was the replacement code. OS X was originally Intel only. If you want to get technical, Intel, and especially Intel x64, has a much newer ABI that is much more efficient than PowerPC 32.

OS X has it's issues, but not a lot of it has to do with bloat. All the foundational APIs are way faster than anything that was available on OS 9.

 

[KaraH]

The iMac doesn't stress form so you are out there on that "form over function" island by yourself.

If the iMac were a 'slave' to form it would still either looking like Luxo Jr or the same rectangular box it has for most of its lifetime over the last 7-8 years. It doesn't.

How could I have been so blind? I guess I do not really NEED more than one 3.5" drives like I could on the 2011 model and before. Or the ability to work with/replace/upgrade the hardware like you can with actual computers.

The iMac is targeted at consumers. I have no idea why you would say on the mac pro forum that it is doing things purely to satisfy function. If the 2012 iMac was the optimal design for everyone why are we even waiting for a new MP?

 

[LorenK]

Mac Pro availability is slim to none and slim just left town. I solved this problem before they pulled the range. £1200 of PC parts and a copy of Mountain Lion later I have my upgrade Hackintosh and it works just fine. Apple will have to try very hard in the future to convince me that their overpriced workstations are worth reinvesting in the future.

I can understand your reticence to wait any longer and unwillingness to order a US machine, but I think one factor that Apple continues to have in its favor is the quality of its parts. That 2008 Mac Pro that your gf is using is now over four years old, how many Wintel boxes last that long, as well as continue to look good?

I have one of those 2008 Mac Pros, albeit a quad, and just spent $120 for another processor and $200 for an SSD, and my old Mac Pro has been rejuvenuated. I hope Apple recognizes the value in that flexibility for its redesign.

 

[goMac]

I have one of those 2008 Mac Pros, albeit a quad, and just spent $120 for another processor and $200 for an SSD, and my old Mac Pro has been rejuvenuated. I hope Apple recognizes the value in that flexibility for its redesign.

I'd assume that could also be a bad thing from Apple's perspective.

 

[calaverasgrande]

The iMac doesn't stress form so you are out there on that "form over function" island by yourself.

if the iMac isn't a perfect example of form over function, what is?
Logically, the headphone jack and SD card reader should be on the front. But that would mar the unblemished front fascia, so they are placed in a harder to reach place to preserve the "form".

 

[calaverasgrande]

You're assuming the beachball is drawing related. It's likely not. The operating system has to go talk to an external card that has to do a radio frequency burst to look for local wifi networks. Sure, they probably could have threaded that. But it's not burning GPU or CPU cycles.



Again, lagginess in the UI does can imply a lot of things. Slow disks spinning up, or other slow hardware.



Replace? The Intel code is older than the PowerPC code. The PowerPC code was the replacement code. OS X was originally Intel only. If you want to get technical, Intel, and especially Intel x64, has a much newer ABI that is much more efficient than PowerPC 32.

OS X has it's issues, but not a lot of it has to do with bloat. All the foundational APIs are way faster than anything that was available on OS 9.

I guess you didn't get the memo. There are a number of applications that no longer function in ML because the older powerpc based libraries that were deprecated are now no longer part of the OS. You know, when you upgrade a Snow Leopard Mac to Lion or Mountian Lion and some apps have a the circle-slash on top of their icon? Those apps had legacy code that should have been updated.
This is not a theory. It is a fact of using a Mac with ML. I am setting up a rMBP for one of our execs and we are having to go to a VM running Windows to run alternate versions of these applications until they catch up with Apple. It has nothing to do with which was first or which is the oldest API. It has to do with Apple getting rid of legacy code. Microsoft has done the same thing with it's OS family though it is not as big of a deal because the current Wintel platform never jumped processors in the same way OS X did.
The conclusion that me and my buddies came to was that they must have hired some cheap, overworked coders, or just plain didn't QA very well when they re-worked those parts of OS X that depended on teh deprecated code.
The wifi menu item, the lagginess in the UI. This is happening on systems that are the highest spec you can buy from Apple right now.
We have 6 Mac Pros in the Art Dept with max ram, SSD, 5870 GPU, and the biggest CPU Apple is selling. they all suffer from UI lag even with nothing open but finder. It isn't crippling, just a minor annoyance.
But it certainly pushes Apple closer to the "plug and pray" and farther from the "Just works".

FYI

• x86-64: versions 10.4.7 through 10.8
• IA-32: versions 10.4.4 through 10.6.8
• PowerPC: versions 10.0 through 10.5.8

While PowerPC stopped being supported as a processor as of 10.6, portions of the OS and bundled apps were still running emulated PowerPC code until very recently. Heck I wouldn't even say for certain that all of it is gone.

 

[goMac]

I guess you didn't get the memo. There are a number of applications that no longer function in ML because the older powerpc based libraries that were deprecated are now no longer part of the OS. You know, when you upgrade a Snow Leopard Mac to Lion or Mountian Lion and some apps have a the circle-slash on top of their icon? Those apps had legacy code that should have been updated.

Right?

This is not a theory. It is a fact of using a Mac with ML. I am setting up a rMBP for one of our execs and we are having to go to a VM running Windows to run alternate versions of these applications until they catch up with Apple. It has nothing to do with which was first or which is the oldest API. It has to do with Apple getting rid of legacy code. Microsoft has done the same thing with it's OS family though it is not as big of a deal because the current Wintel platform never jumped processors in the same way OS X did.

Ok? This happened with OS 9 apps too...

The conclusion that me and my buddies came to was that they must have hired some cheap, overworked coders, or just plain didn't QA very well when they re-worked those parts of OS X that depended on teh deprecated code.

No, they'd actually have to keep updating the PowerPC code for it to work. It's not just a matter of maintenance.

The wifi menu item, the lagginess in the UI. This is happening on systems that are the highest spec you can buy from Apple right now.
We have 6 Mac Pros in the Art Dept with max ram, SSD, 5870 GPU, and the biggest CPU Apple is selling. they all suffer from UI lag even with nothing open but finder. It isn't crippling, just a minor annoyance.
But it certainly pushes Apple closer to the "plug and pray" and farther from the "Just works".

I don't think the UI lag has to do with UI code as much as disk. Finder being the culprit would imply that I'm right. Hard disks haven't improved in speed much in the last decade.

FYI

• x86-64: versions 10.4.7 through 10.8
• IA-32: versions 10.4.4 through 10.6.8
• PowerPC: versions 10.0 through 10.5.8

Missing a bit of your history here...

NeXTStep was on Intel only. Followed by Rhapsody which was Intel, and then got a PowerPC port. That PowerPC port went on to become 10.0, while the Intel version stuck around in Apple's labs and reappeared at 10.4

So the Intel version was first, and is oldest. PowerPC was a port. The open source components of the Intel version have been on Apple's web site since before 10.0 came out.

http://en.wikipedia.org/wiki/NeXTSTEP
http://en.wikipedia.org/wiki/Rhapsody_(operating_system)

It consisted primarily of the OPENSTEP operating system ported to the Power Mac along with a new GUI to make it appear more Mac-like.

Keyword: ported to the Power Mac

I think by now the Intel version has also existed much much longer than the Power PC version as well.