By the time the Mac Pro actually ships it could have the next generation AMD GPU.
This isn't a discussion about mobile. Besides that your bench is useless and likely underwritten by NVidia. Look at some of the interesting numbers here: http://clbenchmark.com/, in simple terms NVidia sucks.
Actually I've found their drivers to be far more stable on Windows. On OSX AMD isn't bad, but I tend to look more at individual applications than benchmarks.
Quote:
PCI Express 3 is a huge improvement over 2. It means fewer lanes are needed to feed the demanding ports leaving more lanes free for TB and other technologies. Beyond that Xeon isn't locked into the Mac Pro and in fact I see it as a mistake. In fact I'd be rather surprised to find the next gen Mac Pro using conventional Xeon chips.
What's specifically objectionable about them? They do offer significantly more lanes than the mainstream lines. Xeon E/EP offers 40 lanes per cpu. The others are 16, with 20 on the E3s. The E3s are basically the same as what the imacs use minus integrated graphics. That seems like the wrong thing to lose.
Quote:
CUDA is dying on the vine like any vendor specific technology should.
How so? It supports things that aren't always possible or feasible in OpenCL. Developers used it due to stability.
So you raised a sleeping thead for a posting several months old. Nice!
A real leak or current information would be nice.
Sometimes I feel just a little information is enough to find out about a device. And software leaks are usually spot-on, that's how people were able to tell the iPhone 5 had a 1136x640 before it launched, it was leaked in the iOS Developer Suit. That's also how we found out the Macbooks and iPads were going to have retina displays (the 2x images were found in the Beta's of OS X and iOS)
There's also always an iTunes update for new devices because they need to add the drivers for them to sync with the devices as well. They don't add the drivers listed on the change logs or anything but you can tell because the iTunes size in MB increases by a bit. It;s not even bundled with Quicktime anymore because of that.
Actually I've found their drivers to be far more stable on Windows. On OSX AMD isn't bad, but I tend to look more at individual applications than benchmarks.
It does vary a bit. However AMD has taken ATI a long ways and in my opinion is doing a far better job with drivers.
What's specifically objectionable about them? They do offer significantly more lanes than the mainstream lines. Xeon E/EP offers 40 lanes per cpu. The others are 16, with 20 on the E3s. The E3s are basically the same as what the imacs use minus integrated graphics. That seems like the wrong thing to lose.
This is an interesting question which I don't have time to go into in depth. But a lot of it goes into user expectations and how the Mac Pro is marketed. To put it frankly Apples marketing of that machine as a high end workstation and then not aggressively keeping it up to date has really damaged the machines image in the community. One can't dismiss Intels role here with the slow Xeon roll outs either.
In any event dropping Xeon would allow Apple to deliver the type of machine that the majority of potential desktop customers want in a machine. That is really it in a nut shell. The drop doesn't have to be complete but then we get into discussions about XMAC and other solutions. The big problem is that there simply isn't much of a market for a desktop computer that starts at $2500 and offers little advantage at that starting point. So then the majority of sales go to even higher end models to the dwindling base of customers that can justify the price. In a nut shell Apple needs a rational price point for entry into the desktop market. That really starts somewhere around $1200 not $2500.
How so? It supports things that aren't always possible or feasible in OpenCL. Developers used it due to stability.
It is also vendor specific, which is the biggest problem with CUDA. Developers going with OpenCL can leverage a wide array of hardware and systems software. Lets put it this way, if I was a CEO piloting a new business I would stay far away from CUDA or other vendor specific solutions.
Sometimes I feel just a little information is enough to find out about a device. And software leaks are usually spot-on,
It is interesting how little has actually leaked about the new Mac Pro. However I don't see posting here as generating any leaks on its own.
that's how people were able to tell the iPhone 5 had a 1136x640 before it launched, it was leaked in the iOS Developer Suit. That's also how we found out the Macbooks and iPads were going to have retina displays (the 2x images were found in the Beta's of OS X and iOS)
If the Pro doesn't come at WWDC maybe the leaks will. I'm still hoping for a major game changer of a machine.
There's also always an iTunes update for new devices because they need to add the drivers for them to sync with the devices as well.
Is it me or is iTunes in need of a major overhaul?
They don't add the drivers listed on the change logs or anything but you can tell because the iTunes size in MB increases by a bit. It;s not even bundled with Quicktime anymore because of that.
It may very well be that the new Mac Pro is so radical that it will need an OS release of its own.
What is the big deal really? It is a machine that all it is.
It might be telling about where Apple is going with the entire Mac line up. Right now I don't need a Mac Pro in its current form, it is way to expensive for far too little. However that doesn't mean my interest in the platform doesn't exist, in fact I find the machine very interesting based on pass development. With respect to the future I seethe machine as an indicator of what Apple sees as viable technologies in the near future. For example does Apple stay with current RAM technologies or gamble a bit on the emerging high speed standards? For secondary storage does Apple via Anobit, break into the high performance PCI Express based storage market or stay with SATA for a bit longer. Some of these technologies could move into Apples laptops real fast so it is interesting to see where Apple goes.
Is it me or is iTunes in need of a major overhaul?
It may very well be that the new Mac Pro is so radical that it will need an OS release of its own.
I really doubt iTunes needs an even newer overhaul... the last update took them an extra month to release even though they announced it at last years WWDC. And many people complained about it's new design even though you can still add the sidebar on the left... the're isn't enough smart people in this world. It was actually faster than iTunes 10 so I can only complain about it not being able to play .AVI or .MKV video files but that's because of legal software bumps they could run into if they did that. I'm actually waiting for the new video format H.265 to come out already
And I don't really think you're being serious about anything, it's highly unlikely that they'll create a new OS for a Mac... hahahah. Maybe new drivers for a touchscreen keyboard with an extendable usb cord/port that allows it to be charged while in use but that's probably not going to happen. This next WWDC is focused on iOS and OS X and probably a new product. Who knows it could finally be an upgrade to the iPad Mini's processor or a new iPhone (not likely to be shown like last year til their ready to launch it).
That and AVI and MKV are terrible. It's much better being able to rely on just M4V and not have to worry about compatibility and what files are what type and difference in performance during playback…
…it's highly unlikely that they'll create a new OS for a Mac…
Oh? 30 years of keyboard and mouse isn't getting a little long in the tooth?
…keyboard with an extendable usb cord/port that allows it to be charged while in use…
Dear sweet heavens, I want a USB to MagSafe2 cable and a new lineup of Bluetooth devices.
Imagine an Apple Bluetooth Keyboard, Magic Mouse, and Magic Trackpad based on Li-ion batteries instead of AAs… Plug them in to charge or use, and unplug them when you want distance.
A touchscreen keyboard, however, wouldn't happen, you're right.
A touchscreen keyboard, however, wouldn't happen, you're right.
I'm not how much range of motion is necessary to maintain long term comfort typing. Fully touch based keys might be extremely uncomfortable for extended use. I don't like mice at all. The ergonomics are terrible.
I really doubt iTunes needs an even newer overhaul... the last update took them an extra month to release even though they announced it at last years WWDC. And many people complained about it's new design even though you can still add the sidebar on the left... the're isn't enough smart people in this world.
That supposed iTunes update was a joke and did almost nothing to improve the app.
It was actually faster than iTunes 10 so I can only complain about it not being able to play .AVI or .MKV video files but that's because of legal software bumps they could run into if they did that.I'm actually waiting for the new video format H.265 to come out alreadyAnd I don't really think you're being serious about anything, it's highly unlikely that they'll create a new OS for a Mac... hahahah.
Did you even read what I posted? I said the new Mac Pro may need its own OS release. Not an entirely new operating system.
Maybe new drivers for a touchscreen keyboard with an extendable usb cord/port that allows it to be charged while in use but that's probably not going to happen. This next WWDC is focused on iOS and OS X and probably a new product.
The focus of WWDC has always been developers. Product announcements allow for tie in to new software features and provide positive energy for the entire show. Make no mistake though WWDC is for developers
Who knows it could finally be an upgrade to the iPad Mini's processor or a new iPhone (not likely to be shown like last year til their ready to launch it).
Actually if there is significant technology in the new iPhone that developers need to know about it could launch soon after or with WWDC. At times WWDC lets the cat out of the bag so to speak.
It does vary a bit. However AMD has taken ATI a long ways and in my opinion is doing a far better job with drivers.
This is an interesting question which I don't have time to go into in depth. But a lot of it goes into user expectations and how the Mac Pro is marketed. To put it frankly Apples marketing of that machine as a high end workstation and then not aggressively keeping it up to date has really damaged the machines image in the community. One can't dismiss Intels role here with the slow Xeon roll outs either.
That actually makes sense. They have marketed it as a high end workstation while simultaneously displaying disinterest at times. They didn't even update gpus with the 2012 delay tactic. OpenGL performance hasn't been that great in recent versions of OSX. They have shown only moderate interest in OpenCL. It's not always supported on the developer software side either, but the lack of interest shown is pretty obvious. I get that other things have outpaced that market, although I disagree with certain individuals that the market for workstations would somehow implode by Apple leaving it. It could drop several vendors and still be a viable market for some even if it's not interesting to Apple. I think a lot of people will be disappointed with whatever is released later, as expectations are often built up due to the lack of an up to date product. The new thing must now be 10-100x as great as the old one to make up for the period in which it languished. Intel has been slow, but they were shipping in volume last June. Given the staggered nature of their updates, Apple could have played catch up and pushed something out late in the year or early this year with Ivy a year from there. If Ivy is shipping late Q3 to supercomputer vendors with everyone else another 3 months out like what happened with Sandy Bridge E, it could be either late this year or early next year before oem workstations based on Ivy are shipping in volume. Beyond that the E5-16xx models look like a rather conservative update. They are probably pushing a greater wedge between them and the E5-26xx dual models.
Quote:
In any event dropping Xeon would allow Apple to deliver the type of machine that the majority of potential desktop customers want in a machine. That is really it in a nut shell. The drop doesn't have to be complete but then we get into discussions about XMAC and other solutions. The big problem is that there simply isn't much of a market for a desktop computer that starts at $2500 and offers little advantage at that starting point. So then the majority of sales go to even higher end models to the dwindling base of customers that can justify the price. In a nut shell Apple needs a rational price point for entry into the desktop market. That really starts somewhere around $1200 not $2500.
It is also vendor specific, which is the biggest problem with CUDA. Developers going with OpenCL can leverage a wide array of hardware and systems software. Lets put it this way, if I was a CEO piloting a new business I would stay far away from CUDA or other vendor specific solutions.
Obviously you have to look a couple years out. I can't find the page I wanted at the moment. Here is a general comparison one. (Edit: note lack of noise reduction via OpenCL, not sure if it's a performance issue or due to different APIs) The Davinci Resolve example is just because of how hardware dependent that software could be in the past. It has OpenCL support due to how Mac heavy the non-Linux base tends to be, but some things still run only in CUDA. I suspect CUDA will remain viable as long as as GPGPU remains a bleeding edge feature in these programs, but even NVidia has been putting more research into OpenCL to ensure they aren't left behind there. If I was buying a new gpu today, I would still prefer NVidia regardless of OS for most applications. Pricing has been weird for the past several years. The early mac pros attempted to add value by using overbuilt base hardware. There was no single option. The base mac pro was a dual socket workstation starting as a quad then later 8 core. Now it starts with a quad hyperthreaded to 8 logical cores. They artificially limited performance growth and kept the starting price bracket. I also would not say it's a fraction of the costs. If the workstation was cheap, those users wouldn't go for the base model.
[...] Intel has been slow, but they were shipping in volume last June. Given the staggered nature of their updates, Apple could have played catch up and pushed something out late in the year or early this year with Ivy a year from there. If Ivy is shipping late Q3 to supercomputer vendors with everyone else another 3 months out like what happened with Sandy Bridge E, it could be either late this year or early next year before oem workstations based on Ivy are shipping in volume.
[...] The early mac pros attempted to add value by using overbuilt base hardware. There was no single option. The base mac pro was a dual socket workstation starting as a quad then later 8 core. Now it starts with a quad hyperthreaded to 8 logical cores. They artificially limited performance growth and kept the starting price bracket. I also would not say it's a fraction of the costs. If the workstation was cheap, those users wouldn't go for the base model.
Intel's production delays explain Apple's failure to put Sandy Bridge into the existing Mac Pro early last year. It also seems clear that a decision was made at that point to skip Sandy Bridge entirely and to not "play catch up" -- a decision that must have been made before Cook sent his post-WWDC 2012 email addressing "our Pro customers" indicating a 2013 Mac Pro redesign was in the works. It's hard to be sure about the basis for that decision, but it isn't hard to imagine that the production window for a Sandy Bridge Mac Pro was too small to be profitable before the introduction of the new Mac Pro planned for this year.
This decision can certainly be read as a "disinterest" in remaining competitive in the market for high-end workstations, but it could also mean the opposite. That is, it could mean that Apple is going to stop offering a Mac Pro that competes with high-end iMacs in the < $2999 range. Hmm's point about how the original Mac Pro has been increasingly limited in order to stay at the same entry price point (at $2499) is important. A lot of people seem to think the new Mac Pro needs to be be even lower-priced to survive. But I think Apple's actions indicate just the opposite is going to happen.
As others continually feel the need to point out in these Mac Pro threads, many low-end Mac Pro customers can easily handle their professional work with a high-end iMac or a high-end MacBook Pro with a Thunderbolt display. Indeed, Apple's allowing the Mac Pro to languish since 2010 has had the effect, whether intentional or not, of forcing a lot of these people to see they don't really need a Mac Pro.
Apple is well positioned now to offer a true, dual-socket Mac Pro that starts a step above the fully-loaded iMac range -- somewhere around $3299. Maybe as high as the current dual-socket entry point, $3799. I don't really see much reason for Apple to continue to have the Mac Pro compete with the iMac in the < $2999 range. Not if they really believe in Thunderbolt. Not to mention the MacBook Pro + Thunderbolt display combinations, which also fit mostly into that range.
As DED pointed out in his most recent editorial, the original 1984 Mac ($2495), adjusted for inflation, cost about $5600. The 1984 Mac 512K ($3195), arguably the first "Mac Pro," with a second drive ($495), cost about $8200 in today's dollars. On the low end, a 1984 Apple IIc sold for $1295 plus $199 for a monitor, or about $3300 today.
It's remarkable that the same basic price structure -- $1500 on the low end, $2500 for the high-end consumer, and $3500 as the starting point for professional customers -- is still in place today, more or less, after three decades of inflation.
That and AVI and MKV are terrible. It's much better being able to rely on just M4V and not have to worry about compatibility and what files are what type and difference in performance during playback…
Oh? 30 years of keyboard and mouse isn't getting a little long in the tooth?
Dear sweet heavens, I want a USB to MagSafe2 cable and a new lineup of Bluetooth devices.
Imagine an Apple Bluetooth Keyboard, Magic Mouse, and Magic Trackpad based on Li-ion batteries instead of AAs… Plug them in to charge or use, and unplug them when you want distance.
A touchscreen keyboard, however, wouldn't happen, you're right.
MKV and MV4 are just just containers of H.264, it's about the codecs you use to decode the video and sounds. I've personally noticed that MV4 is preferred because it's been in the industry longer but MKV is designed to be taken full advantage of in hardware so It's preference really. I've never been a fan of MV4 because it uses more resources than AVI or MKV and it makes the fans on my computer go crazy.
Well at this point I'll take whatever I can get... a magic mouse with a lithium battery instead of having to buy AA batteries all the time sounds awesome. It's a break through for the most part, not revolutionary, but evolutionary.
Intel's production delays explain Apple's failure to put Sandy Bridge into the existing Mac Pro early last year. It also seems clear that a decision was made at that point to skip Sandy Bridge entirely and to not "play catch up" -- a decision that must have been made before Cook sent his post-WWDC 2012 email addressing "our Pro customers" indicating a 2013 Mac Pro redesign was in the works. It's hard to be sure about the basis for that decision, but it isn't hard to imagine that the production window for a Sandy Bridge Mac Pro was too small to be profitable before the introduction of the new Mac Pro planned for this year.
This decision can certainly be read as a "disinterest" in remaining competitive in the market for high-end workstations, but it could also mean the opposite. That is, it could mean that Apple is going to stop offering a Mac Pro that competes with high-end iMacs in the < $2999 range. Hmm's point about how the original Mac Pro has been increasingly limited in order to stay at the same entry price point (at $2499) is important. A lot of people seem to think the new Mac Pro needs to be be even lower-priced to survive. But I think Apple's actions indicate just the opposite is going to happen.
I disagree with your assertion. It was already the same hardware for a very long time due to lack of updates from intel. Letting it go for 3-4 years with the same stuff is not going to help, as some people will take the opportunity migrate to other solutions. I suspect they estimated the numbers and determined it was acceptable. What do you mean increasingly limited to stay at the same price point? I suggested they increased their margins on the lower one to encourage people toward other hardware. The daugherboard design helps with the savings, as they were able to get away from the dual board with one socket populated design. That one required a more expensive board and cpu choice overall.
Quote:
As others continually feel the need to point out in these Mac Pro threads, many low-end Mac Pro customers can easily handle their professional work with a high-end iMac or a high-end MacBook Pro with a Thunderbolt display. Indeed, Apple's allowing the Mac Pro to languish since 2010 has had the effect, whether intentional or not, of forcing a lot of these people to see they don't really need a Mac Pro.
Apple is well positioned now to offer a true, dual-socket Mac Pro that starts a step above the fully-loaded iMac range -- somewhere around $3299. Maybe as high as the current dual-socket entry point, $3799. I don't really see much reason for Apple to continue to have the Mac Pro compete with the iMac in the < $2999 range. Not if they really believe in Thunderbolt. Not to mention the MacBook Pro + Thunderbolt display combinations, which also fit mostly into that range.
Having tried it both ways, you do run into certain disadvantages. Sometimes it was partly an issue of ram + hard drive bandwidth due to the amount of data generated by certain applications, as it couldn't be held in ram and ssds didn't exist. Part of it is that they cut back the hardware offered by the mac pro at the sub $3000 level. That gives the imac somewhat of an "unfair advantage". It seems like they've been trying to market people away from it over time.
Going dual package only would be a pretty big change in direction. Typically Apple avoids lower volume products. I suspect it might be too small that way to even maintain the line. The two don't compete right now anyway. Apple spaced it out by $500, but I think chopping down the hardware and leaving it without updates really lowered the level of interest in that model. If you recall correctly dual socket machines used to start at $2300 with the 1,1 and $2800 with the 3,1.
Quote:
As DED pointed out in his most recent editorial, the original 1984 Mac ($2495), adjusted for inflation, cost about $5600. The 1984 Mac 512K ($3195), arguably the first "Mac Pro," with a second drive ($495), cost about $8200 in today's dollars. On the low end, a 1984 Apple IIc sold for $1295 plus $199 for a monitor, or about $3300 today.
It's remarkable that the same basic price structure -- $1500 on the low end, $2500 for the high-end consumer, and $3500 as the starting point for professional customers -- is still in place today, more or less, after three decades of inflation.
Computers in general have become cheaper. Look at how many $30k turnkey workstation solutions have died out in favor of generic boxes. Computing devices have become cheaper than they were at their introduction.
That actually makes sense. They have marketed it as a high end workstation while simultaneously displaying disinterest at times. They didn't even update gpus with the 2012 delay tactic.
Of all the stupid things to do, delivering the 2012 machine without a GPU update has to be the craziest thing Apple has done in a long time. I can almost understand the modest CPU upgrade but to let the machine out the door without a GPU update just demonstrates a total loss of interest in the market.
OpenGL performance hasn't been that great in recent versions of OSX.
I get slammed for saying that but it is true. For all of its billions you would think the could higher a specialist engineer and get OpenGL up to date.
They have shown only moderate interest in OpenCL. It's not always supported on the developer software side either, but the lack of interest shown is pretty obvious.
I'm not sure if it is a lack of interest or something else. OpenCL seems to be in the same boat as OpenGL. That is slow, actually very slow adoption. As to OpenCL support on intel integrated GPUs I'm not sure who is to blame there. Intel just dropped another OpenCL package for Linux and there was no GPU support there either. Even if Intel is part of the problem, Apple has enough money to cause Intel to take a deeper interest.
I get that other things have outpaced that market, although I disagree with certain individuals that the market for workstations would somehow implode by Apple leaving it. It could drop several vendors and still be a viable market for some even if it's not interesting to Apple.
Many users would be better served by specialist dealers for their workstation needs. Unless Apple suddenly sees the light and firms up workstation hardware and software, there is little point in staying with Apple. In some cases a BSD box would do some power users just as well.
I think a lot of people will be disappointed with whatever is released later, as expectations are often built up due to the lack of an up to date product. The new thing must now be 10-100x as great as the old one to make up for the period in which it languished.
That is possible. Think about it the GPU is at least a couple of generations behind by now. CPU wise it wold be a bigger struggle but I still have this fantasy that Intel and Apple are working together on something Xeon Phi derived.
Intel has been slow, but they were shipping in volume last June. Given the staggered nature of their updates, Apple could have played catch up and pushed something out late in the year or early this year with Ivy a year from there. If Ivy is shipping late Q3 to supercomputer vendors with everyone else another 3 months out like what happened with Sandy Bridge E, it could be either late this year or early next year before oem workstations based on Ivy are shipping in volume.
Lots of complaints are directed at Apple but as you point out Intel doesn't make it easy for workstation vendors. This is one reason I'd like to see a desktop machine with a desktop processor. Call it Mac Pro Light if you want. This actually highlights that the workstation market isn't as big as some think it is. Intel probably ships more ATOM processors and that chip is a complete failure in the marketplace.
I don't keep track of which is which but one of those E series chips isn't much more than a desktop chip in a new package.
Obviously you have to look a couple years out. I can't find the page I wanted at the moment. Here is a general comparison one. (Edit: note lack of noise reduction via OpenCL, not sure if it's a performance issue or due to different APIs) The Davinci Resolve example is just because of how hardware dependent that software could be in the past. It has OpenCL support due to how Mac heavy the non-Linux base tends to be, but some things still run only in CUDA. I suspect CUDA will remain viable as long as as GPGPU remains a bleeding edge feature in these programs, but even NVidia has been putting more research into OpenCL to ensure they aren't left behind there.
Some vendors will prefer Open solutions no matter what.
If I was buying a new gpu today, I would still prefer NVidia regardless of OS for most applications.
I suppose it would depend upon what you use OpenCL for but NVidia sucks at double precision. If that is your game then AMD makes more sense.
Pricing has been weird for the past several years. The early mac pros attempted to add value by using overbuilt base hardware. There was no single option. The base mac pro was a dual socket workstation starting as a quad then later 8 core. Now it starts with a quad hyperthreaded to 8 logical cores.
That isn't exactly a hot performing Quad either. This is what bothers me as the so called enters level model is actually a pretty poor configuration considering the cost.
They artificially limited performance growth and kept the starting price bracket. I also would not say it's a fraction of the costs. If the workstation was cheap, those users wouldn't go for the base model.
The hardcore workstation user doesn't go for that model anyways. This is what is so frustrating. What should be their volume solution is a marketing mistake.
Intel's production delays explain Apple's failure to put Sandy Bridge into the existing Mac Pro early last year. It also seems clear that a decision was made at that point to skip Sandy Bridge entirely and to not "play catch up" -- a decision that must have been made before Cook sent his post-WWDC 2012 email addressing "our Pro customers" indicating a 2013 Mac Pro redesign was in the works. It's hard to be sure about the basis for that decision, but it isn't hard to imagine that the production window for a Sandy Bridge Mac Pro was too small to be profitable before the introduction of the new Mac Pro planned for this year.
Intel is a factor but they aren't completely responsible for Apple lack of interest in the desktop market
This decision can certainly be read as a "disinterest" in remaining competitive in the market for high-end workstations, but it could also mean the opposite. That is, it could mean that Apple is going to stop offering a Mac Pro that competes with high-end iMacs in the < $2999 range. Hmm's point about how the original Mac Pro has been increasingly limited in order to stay at the same entry price point (at $2499) is important. A lot of people seem to think the new Mac Pro needs to be be even lower-priced to survive. But I think Apple's actions indicate just the opposite is going to happen.
The Mac Pro doesn't compete with the iMac, it never has. I see this sort of comment all the time and each time it is as asinine as it was the last time I saw it. Look at the Mac Pros feature set (slots, drive bays, video card options, monitor options and the like). How that flexibility competes with iMac is beyond belief.
As others continually feel the need to point out in these Mac Pro threads, many low-end Mac Pro customers can easily handle their professional work with a high-end iMac or a high-end MacBook Pro with a Thunderbolt display. Indeed, Apple's allowing the Mac Pro to languish since 2010 has had the effect, whether intentional or not, of forcing a lot of these people to see they don't really need a Mac Pro.
In 2008 I felt pushed into a MBP because Apple didn't have a decent desktop at the right price. That sucked back then and it still sucks today even five year later. What is worst the same asinine lineup of hardware has been around for a decade now and what does Apple have to show for it. The vast majority of sales going to the laptop lineup, which stinks. The smell can best be describe as that smelt in the spring time when farmers clean out the pig barns.
Apple is well positioned now to offer a true, dual-socket Mac Pro that starts a step above the fully-loaded iMac range -- somewhere around $3299. Maybe as high as the current dual-socket entry point, $3799. I don't really see much reason for Apple to continue to have the Mac Pro compete with the iMac in the < $2999 range.
It doesn't compete and never has.
Not if they really believe in Thunderbolt. Not to mention the MacBook Pro + Thunderbolt display combinations, which also fit mostly into that range.
Whatever you think of TB is I'm pretty sure Apple has a different opinion. In a nut shell I suspect Apple sees TB as a docking cable with any other functionality as freebies.
As DED pointed out in his most recent editorial,
Referencing somebody with no credibility won't do your arguments any favors.
the original 1984 Mac ($2495), adjusted for inflation, cost about $5600. The 1984 Mac 512K ($3195), arguably the first "Mac Pro," with a second drive ($495), cost about $8200 in today's dollars. On the low end, a 1984 Apple IIc sold for $1295 plus $199 for a monitor, or about $3300 today.
The nature of the technology is such that hardware becomes cheaper each and every node change. You can't come to any rational conclusion about a machines price based on inflation. It makes no sense at all as what is being made is completely different from what was put into an Apple 11c. The value of Apple hardware should be judged solely on the value of the hardware that goes into it today, that is why it is so easy to bash Apple for Mac Pro pricing.
It's remarkable that the same basic price structure -- $1500 on the low end, $2500 for the high-end consumer, and $3500 as the starting point for professional customers -- is still in place today, more or less, after three decades of inflation.
Interestingly some of those machines where grossly overpriced for their time. After my Mac Plus ran its course I dropped out of the Apple rat race for a long time,until 2008 in fact. The problem was back then Apple hardware was even less of a bargain and the OS at that time was going nowhere fast. At that time I switched to a Windows machine, ran that for a year until totally frustrated and then turned to Linux (RedHat 4). Linux was tough back then but it did run well as a multitasking OS once you got hardware sorted.
During my leave of absence from the Apple world I keep a eye on things and longed for a Mac that was affordable and competitive with the generic hardware of the day. I actually thought Apple would die before it got its act together. Even the coming of Steve didn't impress me because the last thing we needed from Apple was some more high priced low speed hardware. The move to Intel was perhaps the most joyful for me as an Apple admirer. The move promised speedy machines and lower prices. Unfortunately the promises where not instant as it took a few years to move beyond the regressions but eventually performance was a bit better. I jumped on board with a 2008 MBP even though it was a bit more expensive than I'd like but not unreasonable so considering the competition. The problem then as it is now; Apple had no viable desktop machine. Viable in this context means decent performance, relative to the rest of the industry, at a decent price. The iMac isn't in the running here and never has been due to being a terrible machine.
In any event Apples pricing structure, on their desktop platforms, suck. The value equations just don't work out for most users thus the dwindling desktop sales which has been a problem at Apple well before the recent market crash. Defend Apple all you want but their desktop hardware has always been a bad deal which is really strange because their laptops aren't that bad. The fact that the lineup has been stagnate for a decade is another matter that hasn't helped sales any.
I disagree with your assertion. It was already the same hardware for a very long time due to lack of updates from intel. Letting it go for 3-4 years with the same stuff is not going to help, as some people will take the opportunity migrate to other solutions. I suspect they estimated the numbers and determined it was acceptable. What do you mean increasingly limited to stay at the same price point? I suggested they increased their margins on the lower one to encourage people toward other hardware.
This encouragement may very well be the case, I think many people in high places had a negative view of the Mac Pro and the market it served. The problem for Apple is that they encouraged a lot of people to leave the Mac world prod completely.
The daugherboard design helps with the savings, as they were able to get away from the dual board with one socket populated design. That one required a more expensive board and cpu choice overall.
I still see that design element as a mistake. Apple really should have a markedly lower cost desktop machine to either sit below the Mac Pro in the line up or to replace it entirely. Even with the daughter card arraignment it is still a far to costly machine in the intro models.
Having tried it both ways, you do run into certain disadvantages. Sometimes it was partly an issue of ram + hard drive bandwidth due to the amount of data generated by certain applications, as it couldn't be held in ram and ssds didn't exist. Part of it is that they cut back the hardware offered by the mac pro at the sub $3000 level. That gives the imac somewhat of an "unfair advantage". It seems like they've been trying to market people away from it over time.
As stated above I don't see any overlap at all in the markets. Even if the Mac Pro undercut the iMac there wouldn't be a defection from the iMac customer base to the rationally priced Mac Pro. Not to bring trucks and cars into it but it is pretty hard to sell a truck to somebody looking for a sedan.
Going dual package only would be a pretty big change in direction. Typically Apple avoids lower volume products. I suspect it might be too small that way to even maintain the line. The two don't compete right now anyway. Apple spaced it out by $500, but I think chopping down the hardware and leaving it without updates really lowered the level of interest in that model. If you recall correctly dual socket machines used to start at $2300 with the 1,1 and $2800 with the 3,1.
I can still see Apple and Intel baking something up that leverages Xeon Phi instead of another processor chip. The Phi coprocessor ought to work fairly well with GCD and OpenCL. This gives Apple the opportunity to rethink the marketing position of the machine. Mac Pro light could be a machine with a single chip processor of 4-6 cores and the Mac Pro heavy could be the same machine with Xeon Phi co processor installed.
The interesting thing here is that the main processor doesn't really matter much, it could be a Xeon or a more generic desktop processor. The expectation is that a GPU would also be there to support the machines video needs.
Computers in general have become cheaper. Look at how many $30k turnkey workstation solutions have died out in favor of generic boxes. Computing devices have become cheaper than they were at their introduction.
Comments
Quote:
Originally Posted by wizard69
By the time the Mac Pro actually ships it could have the next generation AMD GPU.
This isn't a discussion about mobile. Besides that your bench is useless and likely underwritten by NVidia. Look at some of the interesting numbers here: http://clbenchmark.com/, in simple terms NVidia sucks.
Actually I've found their drivers to be far more stable on Windows. On OSX AMD isn't bad, but I tend to look more at individual applications than benchmarks.
Quote:
PCI Express 3 is a huge improvement over 2. It means fewer lanes are needed to feed the demanding ports leaving more lanes free for TB and other technologies. Beyond that Xeon isn't locked into the Mac Pro and in fact I see it as a mistake. In fact I'd be rather surprised to find the next gen Mac Pro using conventional Xeon chips.
What's specifically objectionable about them? They do offer significantly more lanes than the mainstream lines. Xeon E/EP offers 40 lanes per cpu. The others are 16, with 20 on the E3s. The E3s are basically the same as what the imacs use minus integrated graphics. That seems like the wrong thing to lose.
Quote:
CUDA is dying on the vine like any vendor specific technology should.
How so? It supports things that aren't always possible or feasible in OpenCL. Developers used it due to stability.
Quote:
Originally Posted by wizard69
So you raised a sleeping thead for a posting several months old. Nice!
A real leak or current information would be nice.
Sometimes I feel just a little information is enough to find out about a device. And software leaks are usually spot-on, that's how people were able to tell the iPhone 5 had a 1136x640 before it launched, it was leaked in the iOS Developer Suit. That's also how we found out the Macbooks and iPads were going to have retina displays (the 2x images were found in the Beta's of OS X and iOS)
There's also always an iTunes update for new devices because they need to add the drivers for them to sync with the devices as well. They don't add the drivers listed on the change logs or anything but you can tell because the iTunes size in MB increases by a bit. It;s not even bundled with Quicktime anymore because of that.
No news if the new Mac Pro is finally coming out? Can't wait.
What is the big deal really? It is a machine that all it is.
No biggie, just an upgrade. Like every year. Some buy a new model every year, some need a MP now, and would be served if the new one was on sale now.
In any event dropping Xeon would allow Apple to deliver the type of machine that the majority of potential desktop customers want in a machine. That is really it in a nut shell. The drop doesn't have to be complete but then we get into discussions about XMAC and other solutions. The big problem is that there simply isn't much of a market for a desktop computer that starts at $2500 and offers little advantage at that starting point. So then the majority of sales go to even higher end models to the dwindling base of customers that can justify the price. In a nut shell Apple needs a rational price point for entry into the desktop market. That really starts somewhere around $1200 not $2500. It is also vendor specific, which is the biggest problem with CUDA. Developers going with OpenCL can leverage a wide array of hardware and systems software. Lets put it this way, if I was a CEO piloting a new business I would stay far away from CUDA or other vendor specific solutions.
It may very well be that the new Mac Pro is so radical that it will need an OS release of its own.
It might be telling about where Apple is going with the entire Mac line up. Right now I don't need a Mac Pro in its current form, it is way to expensive for far too little. However that doesn't mean my interest in the platform doesn't exist, in fact I find the machine very interesting based on pass development. With respect to the future I seethe machine as an indicator of what Apple sees as viable technologies in the near future. For example does Apple stay with current RAM technologies or gamble a bit on the emerging high speed standards? For secondary storage does Apple via Anobit, break into the high performance PCI Express based storage market or stay with SATA for a bit longer. Some of these technologies could move into Apples laptops real fast so it is interesting to see where Apple goes.
Is it me or is iTunes in need of a major overhaul?
It may very well be that the new Mac Pro is so radical that it will need an OS release of its own.
I really doubt iTunes needs an even newer overhaul... the last update took them an extra month to release even though they announced it at last years WWDC. And many people complained about it's new design even though you can still add the sidebar on the left... the're isn't enough smart people in this world. It was actually faster than iTunes 10 so I can only complain about it not being able to play .AVI or .MKV video files but that's because of legal software bumps they could run into if they did that.I'm actually waiting for the new video format H.265 to come out already
And I don't really think you're being serious about anything, it's highly unlikely that they'll create a new OS for a Mac... hahahah. Maybe new drivers for a touchscreen keyboard with an extendable usb cord/port that allows it to be charged while in use but that's probably not going to happen. This next WWDC is focused on iOS and OS X and probably a new product. Who knows it could finally be an upgrade to the iPad Mini's processor or a new iPhone (not likely to be shown like last year til their ready to launch it).
Originally Posted by dark defender
…not being able to play .AVI or .MKV video files…
That and AVI and MKV are terrible. It's much better being able to rely on just M4V and not have to worry about compatibility and what files are what type and difference in performance during playback…
…it's highly unlikely that they'll create a new OS for a Mac…
Oh? 30 years of keyboard and mouse isn't getting a little long in the tooth?
…keyboard with an extendable usb cord/port that allows it to be charged while in use…
Dear sweet heavens, I want a USB to MagSafe2 cable and a new lineup of Bluetooth devices.
Imagine an Apple Bluetooth Keyboard, Magic Mouse, and Magic Trackpad based on Li-ion batteries instead of AAs… Plug them in to charge or use, and unplug them when you want distance.
A touchscreen keyboard, however, wouldn't happen, you're right.
Quote:
Originally Posted by Tallest Skil
A touchscreen keyboard, however, wouldn't happen, you're right.
I'm not how much range of motion is necessary to maintain long term comfort typing. Fully touch based keys might be extremely uncomfortable for extended use. I don't like mice at all. The ergonomics are terrible.
Quote:
Originally Posted by wizard69
It does vary a bit. However AMD has taken ATI a long ways and in my opinion is doing a far better job with drivers.
This is an interesting question which I don't have time to go into in depth. But a lot of it goes into user expectations and how the Mac Pro is marketed. To put it frankly Apples marketing of that machine as a high end workstation and then not aggressively keeping it up to date has really damaged the machines image in the community. One can't dismiss Intels role here with the slow Xeon roll outs either.
That actually makes sense. They have marketed it as a high end workstation while simultaneously displaying disinterest at times. They didn't even update gpus with the 2012 delay tactic. OpenGL performance hasn't been that great in recent versions of OSX. They have shown only moderate interest in OpenCL. It's not always supported on the developer software side either, but the lack of interest shown is pretty obvious. I get that other things have outpaced that market, although I disagree with certain individuals that the market for workstations would somehow implode by Apple leaving it. It could drop several vendors and still be a viable market for some even if it's not interesting to Apple. I think a lot of people will be disappointed with whatever is released later, as expectations are often built up due to the lack of an up to date product. The new thing must now be 10-100x as great as the old one to make up for the period in which it languished. Intel has been slow, but they were shipping in volume last June. Given the staggered nature of their updates, Apple could have played catch up and pushed something out late in the year or early this year with Ivy a year from there. If Ivy is shipping late Q3 to supercomputer vendors with everyone else another 3 months out like what happened with Sandy Bridge E, it could be either late this year or early next year before oem workstations based on Ivy are shipping in volume. Beyond that the E5-16xx models look like a rather conservative update. They are probably pushing a greater wedge between them and the E5-26xx dual models.
Quote:
In any event dropping Xeon would allow Apple to deliver the type of machine that the majority of potential desktop customers want in a machine. That is really it in a nut shell. The drop doesn't have to be complete but then we get into discussions about XMAC and other solutions. The big problem is that there simply isn't much of a market for a desktop computer that starts at $2500 and offers little advantage at that starting point. So then the majority of sales go to even higher end models to the dwindling base of customers that can justify the price. In a nut shell Apple needs a rational price point for entry into the desktop market. That really starts somewhere around $1200 not $2500.
It is also vendor specific, which is the biggest problem with CUDA. Developers going with OpenCL can leverage a wide array of hardware and systems software. Lets put it this way, if I was a CEO piloting a new business I would stay far away from CUDA or other vendor specific solutions.
Obviously you have to look a couple years out. I can't find the page I wanted at the moment. Here is a general comparison one. (Edit: note lack of noise reduction via OpenCL, not sure if it's a performance issue or due to different APIs) The Davinci Resolve example is just because of how hardware dependent that software could be in the past. It has OpenCL support due to how Mac heavy the non-Linux base tends to be, but some things still run only in CUDA. I suspect CUDA will remain viable as long as as GPGPU remains a bleeding edge feature in these programs, but even NVidia has been putting more research into OpenCL to ensure they aren't left behind there. If I was buying a new gpu today, I would still prefer NVidia regardless of OS for most applications. Pricing has been weird for the past several years. The early mac pros attempted to add value by using overbuilt base hardware. There was no single option. The base mac pro was a dual socket workstation starting as a quad then later 8 core. Now it starts with a quad hyperthreaded to 8 logical cores. They artificially limited performance growth and kept the starting price bracket. I also would not say it's a fraction of the costs. If the workstation was cheap, those users wouldn't go for the base model.
You are right just an upgrade.
Originally Posted by hmm
[...] Intel has been slow, but they were shipping in volume last June. Given the staggered nature of their updates, Apple could have played catch up and pushed something out late in the year or early this year with Ivy a year from there. If Ivy is shipping late Q3 to supercomputer vendors with everyone else another 3 months out like what happened with Sandy Bridge E, it could be either late this year or early next year before oem workstations based on Ivy are shipping in volume.
[...] The early mac pros attempted to add value by using overbuilt base hardware. There was no single option. The base mac pro was a dual socket workstation starting as a quad then later 8 core. Now it starts with a quad hyperthreaded to 8 logical cores. They artificially limited performance growth and kept the starting price bracket. I also would not say it's a fraction of the costs. If the workstation was cheap, those users wouldn't go for the base model.
Intel's production delays explain Apple's failure to put Sandy Bridge into the existing Mac Pro early last year. It also seems clear that a decision was made at that point to skip Sandy Bridge entirely and to not "play catch up" -- a decision that must have been made before Cook sent his post-WWDC 2012 email addressing "our Pro customers" indicating a 2013 Mac Pro redesign was in the works. It's hard to be sure about the basis for that decision, but it isn't hard to imagine that the production window for a Sandy Bridge Mac Pro was too small to be profitable before the introduction of the new Mac Pro planned for this year.
This decision can certainly be read as a "disinterest" in remaining competitive in the market for high-end workstations, but it could also mean the opposite. That is, it could mean that Apple is going to stop offering a Mac Pro that competes with high-end iMacs in the < $2999 range. Hmm's point about how the original Mac Pro has been increasingly limited in order to stay at the same entry price point (at $2499) is important. A lot of people seem to think the new Mac Pro needs to be be even lower-priced to survive. But I think Apple's actions indicate just the opposite is going to happen.
As others continually feel the need to point out in these Mac Pro threads, many low-end Mac Pro customers can easily handle their professional work with a high-end iMac or a high-end MacBook Pro with a Thunderbolt display. Indeed, Apple's allowing the Mac Pro to languish since 2010 has had the effect, whether intentional or not, of forcing a lot of these people to see they don't really need a Mac Pro.
Apple is well positioned now to offer a true, dual-socket Mac Pro that starts a step above the fully-loaded iMac range -- somewhere around $3299. Maybe as high as the current dual-socket entry point, $3799. I don't really see much reason for Apple to continue to have the Mac Pro compete with the iMac in the < $2999 range. Not if they really believe in Thunderbolt. Not to mention the MacBook Pro + Thunderbolt display combinations, which also fit mostly into that range.
As DED pointed out in his most recent editorial, the original 1984 Mac ($2495), adjusted for inflation, cost about $5600. The 1984 Mac 512K ($3195), arguably the first "Mac Pro," with a second drive ($495), cost about $8200 in today's dollars. On the low end, a 1984 Apple IIc sold for $1295 plus $199 for a monitor, or about $3300 today.
It's remarkable that the same basic price structure -- $1500 on the low end, $2500 for the high-end consumer, and $3500 as the starting point for professional customers -- is still in place today, more or less, after three decades of inflation.
Quote:
Originally Posted by Tallest Skil
That and AVI and MKV are terrible. It's much better being able to rely on just M4V and not have to worry about compatibility and what files are what type and difference in performance during playback…
Oh? 30 years of keyboard and mouse isn't getting a little long in the tooth?
Dear sweet heavens, I want a USB to MagSafe2 cable and a new lineup of Bluetooth devices.
Imagine an Apple Bluetooth Keyboard, Magic Mouse, and Magic Trackpad based on Li-ion batteries instead of AAs… Plug them in to charge or use, and unplug them when you want distance.
A touchscreen keyboard, however, wouldn't happen, you're right.
MKV and MV4 are just just containers of H.264, it's about the codecs you use to decode the video and sounds. I've personally noticed that MV4 is preferred because it's been in the industry longer but MKV is designed to be taken full advantage of in hardware so It's preference really. I've never been a fan of MV4 because it uses more resources than AVI or MKV and it makes the fans on my computer go crazy.
Well at this point I'll take whatever I can get... a magic mouse with a lithium battery instead of having to buy AA batteries all the time sounds awesome. It's a break through for the most part, not revolutionary, but evolutionary.
Quote:
Originally Posted by TenThousandThings
Intel's production delays explain Apple's failure to put Sandy Bridge into the existing Mac Pro early last year. It also seems clear that a decision was made at that point to skip Sandy Bridge entirely and to not "play catch up" -- a decision that must have been made before Cook sent his post-WWDC 2012 email addressing "our Pro customers" indicating a 2013 Mac Pro redesign was in the works. It's hard to be sure about the basis for that decision, but it isn't hard to imagine that the production window for a Sandy Bridge Mac Pro was too small to be profitable before the introduction of the new Mac Pro planned for this year.
This decision can certainly be read as a "disinterest" in remaining competitive in the market for high-end workstations, but it could also mean the opposite. That is, it could mean that Apple is going to stop offering a Mac Pro that competes with high-end iMacs in the < $2999 range. Hmm's point about how the original Mac Pro has been increasingly limited in order to stay at the same entry price point (at $2499) is important. A lot of people seem to think the new Mac Pro needs to be be even lower-priced to survive. But I think Apple's actions indicate just the opposite is going to happen.
I disagree with your assertion. It was already the same hardware for a very long time due to lack of updates from intel. Letting it go for 3-4 years with the same stuff is not going to help, as some people will take the opportunity migrate to other solutions. I suspect they estimated the numbers and determined it was acceptable. What do you mean increasingly limited to stay at the same price point? I suggested they increased their margins on the lower one to encourage people toward other hardware. The daugherboard design helps with the savings, as they were able to get away from the dual board with one socket populated design. That one required a more expensive board and cpu choice overall.
Quote:
As others continually feel the need to point out in these Mac Pro threads, many low-end Mac Pro customers can easily handle their professional work with a high-end iMac or a high-end MacBook Pro with a Thunderbolt display. Indeed, Apple's allowing the Mac Pro to languish since 2010 has had the effect, whether intentional or not, of forcing a lot of these people to see they don't really need a Mac Pro.
Apple is well positioned now to offer a true, dual-socket Mac Pro that starts a step above the fully-loaded iMac range -- somewhere around $3299. Maybe as high as the current dual-socket entry point, $3799. I don't really see much reason for Apple to continue to have the Mac Pro compete with the iMac in the < $2999 range. Not if they really believe in Thunderbolt. Not to mention the MacBook Pro + Thunderbolt display combinations, which also fit mostly into that range.
Having tried it both ways, you do run into certain disadvantages. Sometimes it was partly an issue of ram + hard drive bandwidth due to the amount of data generated by certain applications, as it couldn't be held in ram and ssds didn't exist. Part of it is that they cut back the hardware offered by the mac pro at the sub $3000 level. That gives the imac somewhat of an "unfair advantage". It seems like they've been trying to market people away from it over time.
Going dual package only would be a pretty big change in direction. Typically Apple avoids lower volume products. I suspect it might be too small that way to even maintain the line. The two don't compete right now anyway. Apple spaced it out by $500, but I think chopping down the hardware and leaving it without updates really lowered the level of interest in that model. If you recall correctly dual socket machines used to start at $2300 with the 1,1 and $2800 with the 3,1.
Quote:
As DED pointed out in his most recent editorial, the original 1984 Mac ($2495), adjusted for inflation, cost about $5600. The 1984 Mac 512K ($3195), arguably the first "Mac Pro," with a second drive ($495), cost about $8200 in today's dollars. On the low end, a 1984 Apple IIc sold for $1295 plus $199 for a monitor, or about $3300 today.
It's remarkable that the same basic price structure -- $1500 on the low end, $2500 for the high-end consumer, and $3500 as the starting point for professional customers -- is still in place today, more or less, after three decades of inflation.
Computers in general have become cheaper. Look at how many $30k turnkey workstation solutions have died out in favor of generic boxes. Computing devices have become cheaper than they were at their introduction.
Of all the stupid things to do, delivering the 2012 machine without a GPU update has to be the craziest thing Apple has done in a long time. I can almost understand the modest CPU upgrade but to let the machine out the door without a GPU update just demonstrates a total loss of interest in the market. I get slammed for saying that but it is true. For all of its billions you would think the could higher a specialist engineer and get OpenGL up to date. I'm not sure if it is a lack of interest or something else. OpenCL seems to be in the same boat as OpenGL. That is slow, actually very slow adoption. As to OpenCL support on intel integrated GPUs I'm not sure who is to blame there. Intel just dropped another OpenCL package for Linux and there was no GPU support there either. Even if Intel is part of the problem, Apple has enough money to cause Intel to take a deeper interest. Many users would be better served by specialist dealers for their workstation needs. Unless Apple suddenly sees the light and firms up workstation hardware and software, there is little point in staying with Apple. In some cases a BSD box would do some power users just as well. That is possible. Think about it the GPU is at least a couple of generations behind by now. CPU wise it wold be a bigger struggle but I still have this fantasy that Intel and Apple are working together on something Xeon Phi derived. Lots of complaints are directed at Apple but as you point out Intel doesn't make it easy for workstation vendors. This is one reason I'd like to see a desktop machine with a desktop processor. Call it Mac Pro Light if you want. This actually highlights that the workstation market isn't as big as some think it is. Intel probably ships more ATOM processors and that chip is a complete failure in the marketplace. I don't keep track of which is which but one of those E series chips isn't much more than a desktop chip in a new package.
Some vendors will prefer Open solutions no matter what. I suppose it would depend upon what you use OpenCL for but NVidia sucks at double precision. If that is your game then AMD makes more sense. That isn't exactly a hot performing Quad either. This is what bothers me as the so called enters level model is actually a pretty poor configuration considering the cost. The hardcore workstation user doesn't go for that model anyways. This is what is so frustrating. What should be their volume solution is a marketing mistake.
During my leave of absence from the Apple world I keep a eye on things and longed for a Mac that was affordable and competitive with the generic hardware of the day. I actually thought Apple would die before it got its act together. Even the coming of Steve didn't impress me because the last thing we needed from Apple was some more high priced low speed hardware. The move to Intel was perhaps the most joyful for me as an Apple admirer. The move promised speedy machines and lower prices. Unfortunately the promises where not instant as it took a few years to move beyond the regressions but eventually performance was a bit better. I jumped on board with a 2008 MBP even though it was a bit more expensive than I'd like but not unreasonable so considering the competition. The problem then as it is now; Apple had no viable desktop machine. Viable in this context means decent performance, relative to the rest of the industry, at a decent price. The iMac isn't in the running here and never has been due to being a terrible machine.
In any event Apples pricing structure, on their desktop platforms, suck. The value equations just don't work out for most users thus the dwindling desktop sales which has been a problem at Apple well before the recent market crash. Defend Apple all you want but their desktop hardware has always been a bad deal which is really strange because their laptops aren't that bad. The fact that the lineup has been stagnate for a decade is another matter that hasn't helped sales any.
The interesting thing here is that the main processor doesn't really matter much, it could be a Xeon or a more generic desktop processor. The expectation is that a GPU would also be there to support the machines video needs.
That argument lost big time.