You have names to back that up. Names of millions of people that MUST have a Mac Pro to do their work, especially the kind of multitasking scenario given.
Oh, heavens no, which is why I highballed at tens of thousands for workplace Mac Pro users. I'm referencing the ripple effect from no longer having something on the high end. I realize I may still be highballing with that second number, but there are things the iMac will never do that people need done. And with no solution present, eventually the dominoes start to tip.
Not a completely accurate metaphor, to be sure. I suppose ice breaking off the coast of an Antarctic glacier is a better one, though still with its flaws. The glacier itself won't ever go away entirely, but it can be reduced by a significant portion.
Having said that, I don't think they'd get rid of it, and I certainly hope they don't. I'd actually love a redesigned solution as proposed by… whoever that was. The one guy on here who claimed to have met a guy that had seen the new "Apple Galaxy" system of Thunderbolt-connected distributed cube computers. That'd be something inspiring, you know?
Have a Mac Mini? Need more power for something? Buy a second one and plug 'er in. Upgraded your Mini to a new model and don't know what to do with the old one? Keep it around, plug it in, and have a little more processing oomph. Of course I mention the Mini because it's the easiest to envision a wider variety of uses in this scenario, as its form factor really lends to stacking, but a smaller, sleeker Mac Pro replacement/upgrade would be where the system truly shines.
it won't matter too much what I buy next as long as I have enough ports and options for storage connectivity
USB 3 on the new machines will help a lot. Getting one of those brushed metal Samsung 830 SSDs with just the raw SATA connector to USB 3 would be great for fast storage. They can be used as scratch disks for active projects.
The idea that the entertainment and advertising industries are some kind of minor niche market in decline that can and should be ignored is just laughable on its face.
It's not that they are minor nor in decline but clearly they aren't buying Mac Pros. You can't take an entire industry, stamp it with the professional label and declare that it requires a Mac Pro. The 2D design industry would find far more value in a 27" iMac than a Mac Pro because they don't do CPU intensive tasks.
A huge number of high-value jobs can be done in real-time now on basic core hardware with peripherals:
For the CPU intensive tasks like compositing and rendering, Mac Pros are great machines but, as I said before, that's only the higher-end models. This is only suitable if jobs can pay enough to cover the $3,500-6,000+ outlay on the hardware and it's not the case that jobs can't be done with the slower machines, you would just get e.g 1 minute per frame vs 20 seconds per frame.
Workflows these days are such that a lot of things are done in real-time though and you just need the raw processing right at the end so it's not that your entire workflow drops to 1/3rd of the speed using an iMac, only your final step, which is a fraction of the workflow.
I do think that it would be premature to kill the Mac Pro at this point but it was premature to kill the Macbook before SSDs dropped in price and I think now is the time they have to decide to commit to it for a long term or not.
There is a chance they can get away with using new CPUs/GPUs without even mentioning it (that would be quite funny actually) but I think the direction they are going points to a redesign of the enclosure. They won't do a redesign unless they are commiting to it long term and I don't think the Mac Pro has a long term ahead.
Not to mention potential growth areas for OS X like the hard sciences and medicine -- here the success of iOS devices is likely to drive growth back into the rest of the OS X ecosystem, including a new and improved Mac Pro. An Apple television would only add to the phenomenon.
It's a bit of a stretch to think that because someone owns an iPhone, they are going to consider spending upwards of $2500 on a 40lb workstation. It's good advertising for the brand but there's been no evidence that the current 250 million or more sales of iOS devices has impacted the Mac Pro line at all.
Your vision of an iMac solution for half of the current Mac Pro base reminds me of a tag used in a sports blog that I read, "i come up with an incredibly complicated solution to something that may not be a problem" -- there isn't a problem here -- the current solution, the Mac Pro, is elegant, flexible, and works beautifully.
The problem is people aren't buying them and people are managing to use iMacs just fine:
Oh, and your comment about how few of the "whiners" are actually going to buy ignores the fact that a fair segment of the existing Mac Pro base (stock 1,1 and 2,1 machines, like mine) is going to be left behind by Mountain Lion
Exactly, only upgarding when forced to due to software incompatibility. That's your typical Mac Pro crowd.
USB 3 on the new machines will help a lot. Getting one of those brushed metal Samsung 830 SSDs with just the raw SATA connector to USB 3 would be great for fast storage. They can be used as scratch disks for active projects.
I try not to rely too heavily on scratch disks these days. You must remember we were limited to 3-4 GB ram ceilings for many years. The capability to go higher was there long before the benefit really became noteworthy. These days I will max it when I buy the machine, or get as much as I can without the price going through the roof. Like 8GB dimms in the lower mac pros would have been pointless a couple years ago under most circumstances, just because they'd inflate your costs so much that you'd be better off going with a dual model and 8x4GB dimms rather than 4x8. It's getting to a point where much of this can be handled in ram, and it allows me to enable caching in many areas where it wouldn't have been practical in the past.
Quote:
Originally Posted by Marvin
There is a chance they can get away with using new CPUs/GPUs without even mentioning it (that would be quite funny actually) but I think the direction they are going points to a redesign of the enclosure. They won't do a redesign unless they are commiting to it long term and I don't think the Mac Pro has a long term ahead.
It's a bit of a stretch to think that because someone owns an iPhone, they are going to consider spending upwards of $2500 on a 40lb workstation. It's good advertising for the brand but there's been no evidence that the current 250 million or more sales of iOS devices has impacted the Mac Pro line at all.
The problem is people aren't buying them and people are managing to use iMacs just fine:
Exactly, only upgarding when forced to due to software incompatibility. That's your typical Mac Pro crowd.
That's your typical computer crowd these days. Laptop users sometimes replace frequently, but it's more frequently due to concerns over device failure than computing needs. I'd suggest the mac pro crowd is more the type to upgrade when their needs increase, and they can realize real gains from the upgrade. The current problem is to some degree an issue of segmentation. Look at the 1,1. It granted a very competitive quad core workstation for its time. In 2008 they went to 8 cores, but many software programs were still lagging greatly in core scaling. This included some of Apple's own applications like FCP. This is a problem. You need to be able to benefit from the upgrade. In 2009 they split up the line in a very bad way. Toward the sub $3k region, they've been almost running in place since 2008 on performance as core scaling and 64 bit applications have improved the performance of that model since its release. Geekbench and other testing comes out pretty close. To see a massive improvement from that point on, it's basically the 6 core if you need the highest possible clock speed or the 12 core to make such an upgrade worth it. There hasn't really been a way since 2008 to buy into higher performance around the same price point unless it's a case of heavy ram requirements. At that point the upgrade makes sense, but it's mostly to avoid the cost of FB dimms. As I mentioned, Snow Leopard on, you have a lot of applications that saw minimal benefit past 4GB of ram that can now use as much as you throw at them. In many cases computing times fall off a cliff once it can just cache everything to ram and finding contiguous chunks is no longer an issue.
Marvin, Many of the points you make are reasonable, although I'm not really sure where your data on Mac Pro sales comes from. I just think you're drawing the wrong conclusions. There is a place for a Mac Pro in the product line. Most of the problems that exist (apart from the current economic depression) can be addressed in a redesign of the Mac Pro. Plus, Cook is an operations guy. I'd look for innovations in how the Mac Pro is built and how it is "fulfilled," and not the simple bean-counter solution of killing it because it's not selling as well as other products.
My point about the iOS halo effect is that there are increasing numbers of specialized scientific and medical-industry apps for the iPad, and as the iOS continues to be best and most secure platform for this kind of thing we will see more adoption of the OS X platform. The same is true in the entertainment industry.
Quote:
Originally Posted by Marvin
Exactly, only upgarding when forced to due to software incompatibility. That's your typical Mac Pro crowd.
This is just ass-backwards. If iMacs are so powerful and capable, why wouldn't people hang onto them until they are forced to upgrade? How is the iMac any different? Certainly not price. Your main point is that the high-end iMac = the low-end Mac Pro, correct? The total cost isn't that much different. How could it be? As you and your ilk are so fond of pointing out, Apple isn't in the business of giving customers something for nothing:
Let's say $2300 for the iMac with the high-end i7 and 2GB of video memory. We'll forget about additional RAM, since nobody buys that from Apple. Add $1200 for a 4TB Thunderbolt RAID system (including cable). So that's about $3500 for what I need.
Now let's just, for the sake of argument, say the Sandy Bridge Xeons came out last year when they were supposed to and Apple did a minor refresh to the Mac Pro line (without Thunderbolt). I think it's safe to say whatever single processor was chosen for the low-end Mac Pro (which is basically just a server that Apple also sells as a workstation), it is comparable to the 3.4 GHz i7 -- most likely either the E5-1620 ($294) or the E5-1650 ($583). Let's also assume an upgrade in graphics, but maybe we still need to add $100 for 2GB of video memory. So that's $2600 plus $400 in hard drives to get to the same 5TB storage capacity. That leaves me $500 for a nice Samsung display, but since I'm not a child and I've got a job, I'll splurge and spend $1000 for the Apple display. Apple makes a profit not just on the computer, but also on the display. Same as it ever was.
The point is that it's great that the iMac and the MacBook Pro lines are now far more easily extensible than ever before. But those solutions aren't really any cheaper than the Mac Pro ever was, and there is a place in the world for a computer from Apple that doesn't come with a display built-in, preferably in a beautiful new, sleek, rack-mountable Jonathan Ive design.
I try not to rely too heavily on scratch disks these days. You must remember we were limited to 3-4 GB ram ceilings for many years.
I mean in terms of video editing to store the source footage rather than a VM location - source drives are called scratch drives as they are used for render files. It could stream high bitrate source clips from the SSD due to the fast read speeds e.g 4K ProRes to a MBA at 60MB/s per stream. Not much footage (70 mins or so) at that quality on 256GB of course but you get the idea.
Trying to upgrade or fix a 6 year old machine instead of buying a new model. The initial investment is high so upgrades to new models aren't cost-effective. It's also not that easy to migrate an internal RAID from one machine to another.
I'd look for innovations in how the Mac Pro is built and how it is "fulfilled," and not the simple bean-counter solution of killing it because it's not selling as well as other products.
I would too, I think it would be a bad move to leave it as is but where's the motivation? All the graphs show downward movement for desktops. Even if they made a stunning new small form factor Pro with the latest hardware, how much growth is there?
We always hear stories (I invented many of my own years ago) about how, if Apple just made a few better decisions with the Pro, it would change everything but they are really limited in what they can do with it and the sales are going to keep lowering over time.
There's only so long this will last. Even HP wanted to get out of the desktop business and they are one of the biggest desktop manufacturers. It's not just iMacs doing this in Apple's lineup but laptops. A Macbook Pro outperforms an entry-level MP now.
If iMacs are so powerful and capable, why wouldn't people hang onto them until they are forced to upgrade?
Because they run out of warranty and Apple often adds things to the iMac that are compelling upgrades like Thunderbolt, better displays, better backlight, SSD etc plus it is partly due to price. There is a fairly affordable upgrade path from a 3 year old iMac to a new one and a regular update cycle.
But those solutions aren't really any cheaper than the Mac Pro ever was, and there is a place in the world for a computer from Apple that doesn't come with a display built-in, preferably in a beautiful new, sleek, rack-mountable Jonathan Ive design.
The example you gave about the hard drives doesn't include hardware RAID. A Pegasus R4 has hardware RAID and to add this to a Mac Pro costs $700. You also get 4TB RAID Thunderbolt cheaper ($600):
If Sandy Bridge Xeons came out last year, they would match the top-end Sandy Bridge i7 but that didn't happen so now the Sandy Bridge Xeon would go up against the Ivy Bridge i7 and fall short.
Most things they do in an update are going to be really underwhelming. The top end chips that fit in the price bracket of the current 12-core are only 40% faster than the last models. If they put the following chip in a small form factor single CPU model for $2499, that would be good and reasonably faster than an iMac:
If Sandy Bridge Xeons came out last year, they would match the top-end Sandy Bridge i7 but that didn't happen so now the Sandy Bridge Xeon would go up against the Ivy Bridge i7 and fall short.
Are you suggesting that if a new Macpro is released, it will include a Sandy bridge and not Ivy bridge Xeon? Intel just released the Ivybridge Xeon for the E3 series which means the E5 and E7 can't be far behind or may have already been released. I don't follow Intel CPU updates closely enough to say for sure.
All the pieces are pretty much in place for Apple to update the Macpro and really all their computers. Although doubtful, I sure would like to see a Blu-ray writer included at least as an option. Prices have dropped on media and HD home movies look great on BR and nice to share with family or friends as well as a nice additional back up for files.
Those Xeons are not the type that Apple would want to use in the Mac Pro. In March Intel announced the Sandy Bridge Xeons, which are more cores and higher cache. According to earlier Intel roadmaps, we aren't expecting to see full power Xeons based on Ivy Bridge until late 2012 or early 2013. Intel is very out of sync between their consumer and server processors.
I certainly agree that the iMac has become a much more attractive platform (and I don't mean the way it looks) of late, but there are still many things for which a tower is simply a better arrangement. While the TB connectors are all fine and well, one must wait on their development and compatibility and there are performance hits involved if they are daisy chained, not to mention that, for the foreseeable future, it appears that there will be price premiums for TB external devices. Still, it is a useful technology. When TB eventually gets into fiber, a complete re-evaluation of the matter may be in order, but, for now anyway, there remains a place for towers. (The Mac Pro monicker is really misleading in many ways, but that is another matter.)
For example, in a tower, installation of a PCIe SSD (not PCIe/SATA) and a PCIe CF and/or SDHC card reader certainly will be desirable as image file sizes continue to grow...not to mention video files although they are probably better served by an HDMI input.
Don't forget the matter of RAM either. A proper tower can install more and keep it cooled better whether it is a Xeon system or not. Xeon systems, of course, can handle a lot more RAM if the logic board supports it.
Exactly how are Apple workstations more out of date than HP or Dell ones? The Xeons have only been updated just recently, and HP and Dell only updated their workstations in April. Before then they were using the same generation of processor as the current Mac Pro.
So give Apple a few months at least to release their new models too, and we can compare apples with apples.
Joe the Dragon covered this: "but HP and others did lower prices and bumped up video cards / ram size over the same time frame.
Apple same price same ram and same video cards for same 2 years." (Post #123).
It is about the system, not just the CPU. Apple was frozen in time selling dated components at today's premium prices which never sits well with people.
There is increasing doubt that there will be a Sandy Bridge Xeon in a timeframe that makes much sense. If the Haswell is on time, there really is no particular reason to release a Sandy Bridge Xeon a few months (at most) before Haswell. It would simply clutter up the market. Of course it remains to be seen if Haswell will be on time, but the major delay in Sandy Bridge was the new process which will be utilized in the (new architecture) Haswell.
Even if Apple does produce another Mac Pro, one has to suspect that it is only a matter of time before it goes the way of the X-Serve which may lead people to seek other solutions which will suit their needs in the long term rather than riding the ship down.
For example, in a tower, installation of a PCIe SSD (not PCIe/SATA) and a PCIe CF and/or SDHC card reader certainly will be desirable as image file sizes continue to grow...not to mention video files although they are probably better served by an HDMI input.
Don't forget the matter of RAM either. A proper tower can install more and keep it cooled better whether it is a Xeon system or not. Xeon systems, of course, can handle a lot more RAM if the logic board supports it.
USB 3 is going to cover a lot of what Thunderbolt is too expensive for and is plenty for 4K films being transferred off professional camera storage or digital images. Storage is still limited to the read/write speeds, which rarely top 1GB/s.
In terms of RAM, the dual processor Mac Pro can only manage double the iMac - 64GB vs 32GB. The single CPU one is the same 32GB.
Almost all of the benefits of the Mac Pro can largely be dismissed besides raw processing power.
With USB 3 and Thunderbolt, PCI slots just aren't that much use any more. While people cling to the specs, the reality is that the speed doesn't matter. 5Gbps for USB 3 is fast enough for almost any IO you'd ever need to do (1TB in 25 minutes) again partly because storage generally isn't that fast anyway. Data has to come from something and go to something else so the source and destination has to max the link speed.
You can have multiple GPUs but who uses multiple GPUs? Single cards can support 6 displays and you can even run displays over USB 3. You can't put more than a single high-end card in there due to the power limit so at best you get a GTX 670 in there and the iMac's 7970M will get 85% of the speed. The iMac doesn't have 6 display outputs of course but it's not a typical usage scenario.
The Mac Pro only makes sense for resource intensive tasks like rendering and even at this, it's only really the final render and 3x faster on the highest end but you can buy a cheap render box for this.
There are no reasons left that necessitates the existence of the Mac Pro. Thunderbolt was the last thing they needed to make to allow them to kill it off. Whether they kill it off now or later, they will kill it off.
USB 3 is going to cover a lot of what Thunderbolt is too expensive for and is plenty for 4K films being transferred off professional camera storage or digital images. Storage is still limited to the read/write speeds, which rarely top 1GB/s.
In terms of RAM, the dual processor Mac Pro can only manage double the iMac - 64GB vs 32GB. The single CPU one is the same 32GB.
Almost all of the benefits of the Mac Pro can largely be dismissed besides raw processing power.
With USB 3 and Thunderbolt, PCI slots just aren't that much use any more. While people cling to the specs, the reality is that the speed doesn't matter. 5Gbps for USB 3 is fast enough for almost any IO you'd ever need to do (1TB in 25 minutes) again partly because storage generally isn't that fast anyway. Data has to come from something and go to something else so the source and destination has to max the link speed.
You can have multiple GPUs but who uses multiple GPUs? Single cards can support 6 displays and you can even run displays over USB 3. You can't put more than a single high-end card in there due to the power limit so at best you get a GTX 670 in there and the iMac's 7970M will get 85% of the speed. The iMac doesn't have 6 display outputs of course but it's not a typical usage scenario.
The Mac Pro only makes sense for resource intensive tasks like rendering and even at this, it's only really the final render and 3x faster on the highest end but you can buy a cheap render box for this.
There are no reasons left that necessitates the existence of the Mac Pro. Thunderbolt was the last thing they needed to make to allow them to kill it off. Whether they kill it off now or later, they will kill it off.
Which model of Mac Pro are you planning on buying?
See bold, are you saying that a 7970M will only be 15% slower than a GTX670? Based on?
I happen to agree that Apple will kill it. And with its passing a number of former Mac users will find other solutions. I doubt Apple will miss them, but there are rather few applications now that are not available for other platforms which will work just fine.
I happen to agree that Apple will kill it. And with its passing a number of former Mac users will find other solutions. I doubt Apple will miss them…
Yeah, they sure won't miss the demographic that shares their core beliefs.
So Final Cut Pro will be discontinued, too. And Logic… and Aperture… No Mac OS X Server, so everyone will be forced to use Windows or another Unix flavor… No Apple Remote Desktop, since that won't work on those OS', Anything I missed?
Joe the Dragon covered this: "but HP and others did lower prices and bumped up video cards / ram size over the same time frame.
Apple same price same ram and same video cards for same 2 years." (Post #123).
It is about the system, not just the CPU. Apple was frozen in time selling dated components at today's premium prices which never sits well with people.
There is increasing doubt that there will be a Sandy Bridge Xeon in a timeframe that makes much sense. If the Haswell is on time, there really is no particular reason to release a Sandy Bridge Xeon a few months (at most) before Haswell. It would simply clutter up the market. Of course it remains to be seen if Haswell will be on time, but the major delay in Sandy Bridge was the new process which will be utilized in the (new architecture) Haswell.
Even if Apple does produce another Mac Pro, one has to suspect that it is only a matter of time before it goes the way of the X-Serve which may lead people to seek other solutions which will suit their needs in the long term rather than riding the ship down.
Apple always pulls this kind of crap. The problem is that they continue to make their mac pro configurations increasingly less competitive. Why would Haswell be on time in this socket? Ivy Bridge E is scheduled for Q2 2013. Sandy Bridge E has only been trickling out. What could possibly point to a Q2-Q3 release for Haswell? You'd basically need this to guarantee the ability to order a machine using it in 2013. Intel will still make their server money either way, but I wonder if they have any plans for catching up somewhere down the line.
Quote:
Originally Posted by Marvin
USB 3 is going to cover a lot of what Thunderbolt is too expensive for and is plenty for 4K films being transferred off professional camera storage or digital images. Storage is still limited to the read/write speeds, which rarely top 1GB/s.
In terms of RAM, the dual processor Mac Pro can only manage double the iMac - 64GB vs 32GB. The single CPU one is the same 32GB.
Almost all of the benefits of the Mac Pro can largely be dismissed besides raw processing power.
With USB 3 and Thunderbolt, PCI slots just aren't that much use any more. While people cling to the specs, the reality is that the speed doesn't matter. 5Gbps for USB 3 is fast enough for almost any IO you'd ever need to do (1TB in 25 minutes) again partly because storage generally isn't that fast anyway. Data has to come from something and go to something else so the source and destination has to max the link speed.
You can have multiple GPUs but who uses multiple GPUs? Single cards can support 6 displays and you can even run displays over USB 3. You can't put more than a single high-end card in there due to the power limit so at best you get a GTX 670 in there and the iMac's 7970M will get 85% of the speed. The iMac doesn't have 6 display outputs of course but it's not a typical usage scenario.
The Mac Pro only makes sense for resource intensive tasks like rendering and even at this, it's only really the final render and 3x faster on the highest end but you can buy a cheap render box for this.
There are no reasons left that necessitates the existence of the Mac Pro. Thunderbolt was the last thing they needed to make to allow them to kill it off. Whether they kill it off now or later, they will kill it off.
Which model of Mac Pro are you planning on buying?
Marvin you're getting into the preachy crap again. The mac pro can actually support 128GB of ram, but you top out at 96 supported by OSX. Apple only mentions 64 so that you don't find that limitation. I think the quad is limited to 32 though. On multiple gpus, that would be more useful PC side with tesla cards and things like that. It's not that no one uses them. They just offer less benefit under OSX. Regarding graphics cards, don't compare until you've seen them live. Drivers account for quite a lot there. It annoys me somewhat that everyone feels thunderbolt will solve every problem simply because it was marketed that way initially. Right now you don't have a very clean solution to everything via thunderbolt beyond your imagination even if you drop all preceding purchases as sunken costs and start over from nothing. If anything thunderbolt may address some of the desire for a mid range box with slots. As to if Apple will keep it, thunderbolt means very little in that equation.
Quote:
Originally Posted by dragit
Can someone please help me connect my $10,000 SAS Raid array to an iMac? Or should I just throw out any gear that isn't supported by "thunderbolt"?
There is serious money spent on peripherals that will never work with an iMac, the power of the machine is only really a small aspect of this problem…
Povilas obviously hasn't invested any money intoa Mac Pro studio and supporting hardware, what a spanker.
Something that seems to be missed is that thunderbolt only really covers those who were buying the low end of the mac pro line due to lack of a mid range tower from Apple. Even then it's not really a mature technology at this point.
Yeah, they sure won't miss the demographic that shares their core beliefs.
So Final Cut Pro will be discontinued, too. And Logic… and Aperture… No Mac OS X Server, so everyone will be forced to use Windows or another Unix flavor… No Apple Remote Desktop, since that won't work on those OS', Anything I missed?
I am not sure how much of the "core beliefs" comment is tongue in cheek.
I am also not sure whether Apple have any "core beliefs" left. Sure, the company is making money hand over fist, but there is increasingly little to distinguish it from any other computer company other than margins. Sir Jonny continues to create thermally challenged products and most of the public never notices.
The whole situation at Apple is increasingly discouraging...and don't get me started on iOS.
See bold, are you saying that a 7970M will only be 15% slower than a GTX670? Based on?
There are a couple of benchmarks but it shows in real games too. Here is a 7970M playing Skyrim at Ultra 1080p:
50-60fps
Here is a GTX 680 on Ultra:
It also gets 50-60 fps.
The GTX 680 will benchmark higher but in terms of gaming, small variations under even 50% difference are not really noticeable. 15% would mean 7970M = 50fps, GTX 680 = 57fps. Even the lower card in the new iMac (7770M or whatever), if it's as much a 50% difference, you just lower a few settings.
Yeah, they sure won't miss the demographic that shares their core beliefs.
So Final Cut Pro will be discontinued, too. And Logic… and Aperture… No Mac OS X Server, so everyone will be forced to use Windows or another Unix flavor… No Apple Remote Desktop, since that won't work on those OS', Anything I missed?
All those things run fine on all their other computers. OSX Server runs on a Mini Server and uses a fraction of the power of a Mac Pro Server. FCPX uses OpenCL so will use the large compute performance of the 7970M to at least double the performance of the quad-i7.
This makes it equivalent to an 8-core or maybe even 12-core Pro without GPU OpenCL. Obviously a Mac Pro would also have GPU OpenCL but that would be Apple competing with itself. Relative to Avid on a PC Xeon workstation, FCPX on an iMac will offer far better performance per dollar.
That's where Apple makes a statement. Outperforming the competition with a larger, more expensive workstation is expected. When they can do it with something cheaper in the form factor of a display, that's impressive.
The mac pro can actually support 128GB of ram, but you top out at 96 supported by OSX. Apple only mentions 64 so that you don't find that limitation. I think the quad is limited to 32 though.
Ah yeah, there are kits with 16GB dimms. They are quite expensive at $1850 for 128GB, $1400 for 96GB but the prices will drop over time. Again though, this is only meaningful for people who need more than 32GB of RAM. How many people are we talking about here? Plus, Ivy Bridge iMacs might support 16GB dimms.
Comments
Quote:
Originally Posted by charlituna
You have names to back that up. Names of millions of people that MUST have a Mac Pro to do their work, especially the kind of multitasking scenario given.
Oh, heavens no, which is why I highballed at tens of thousands for workplace Mac Pro users. I'm referencing the ripple effect from no longer having something on the high end. I realize I may still be highballing with that second number, but there are things the iMac will never do that people need done. And with no solution present, eventually the dominoes start to tip.
Not a completely accurate metaphor, to be sure. I suppose ice breaking off the coast of an Antarctic glacier is a better one, though still with its flaws. The glacier itself won't ever go away entirely, but it can be reduced by a significant portion.
Having said that, I don't think they'd get rid of it, and I certainly hope they don't. I'd actually love a redesigned solution as proposed by… whoever that was. The one guy on here who claimed to have met a guy that had seen the new "Apple Galaxy" system of Thunderbolt-connected distributed cube computers. That'd be something inspiring, you know?
Have a Mac Mini? Need more power for something? Buy a second one and plug 'er in. Upgraded your Mini to a new model and don't know what to do with the old one? Keep it around, plug it in, and have a little more processing oomph. Of course I mention the Mini because it's the easiest to envision a wider variety of uses in this scenario, as its form factor really lends to stacking, but a smaller, sleeker Mac Pro replacement/upgrade would be where the system truly shines.
USB 3 on the new machines will help a lot. Getting one of those brushed metal Samsung 830 SSDs with just the raw SATA connector to USB 3 would be great for fast storage. They can be used as scratch disks for active projects.
It's not that they are minor nor in decline but clearly they aren't buying Mac Pros. You can't take an entire industry, stamp it with the professional label and declare that it requires a Mac Pro. The 2D design industry would find far more value in a 27" iMac than a Mac Pro because they don't do CPU intensive tasks.
A huge number of high-value jobs can be done in real-time now on basic core hardware with peripherals:
http://goo.gl/wz5GC
For the CPU intensive tasks like compositing and rendering, Mac Pros are great machines but, as I said before, that's only the higher-end models. This is only suitable if jobs can pay enough to cover the $3,500-6,000+ outlay on the hardware and it's not the case that jobs can't be done with the slower machines, you would just get e.g 1 minute per frame vs 20 seconds per frame.
Workflows these days are such that a lot of things are done in real-time though and you just need the raw processing right at the end so it's not that your entire workflow drops to 1/3rd of the speed using an iMac, only your final step, which is a fraction of the workflow.
I do think that it would be premature to kill the Mac Pro at this point but it was premature to kill the Macbook before SSDs dropped in price and I think now is the time they have to decide to commit to it for a long term or not.
There is a chance they can get away with using new CPUs/GPUs without even mentioning it (that would be quite funny actually) but I think the direction they are going points to a redesign of the enclosure. They won't do a redesign unless they are commiting to it long term and I don't think the Mac Pro has a long term ahead.
It's a bit of a stretch to think that because someone owns an iPhone, they are going to consider spending upwards of $2500 on a 40lb workstation. It's good advertising for the brand but there's been no evidence that the current 250 million or more sales of iOS devices has impacted the Mac Pro line at all.
The problem is people aren't buying them and people are managing to use iMacs just fine:
http://www.apple.com/finalcutpro/in-action/
Exactly, only upgarding when forced to due to software incompatibility. That's your typical Mac Pro crowd.
Quote:
Originally Posted by Marvin
USB 3 on the new machines will help a lot. Getting one of those brushed metal Samsung 830 SSDs with just the raw SATA connector to USB 3 would be great for fast storage. They can be used as scratch disks for active projects.
I try not to rely too heavily on scratch disks these days. You must remember we were limited to 3-4 GB ram ceilings for many years. The capability to go higher was there long before the benefit really became noteworthy. These days I will max it when I buy the machine, or get as much as I can without the price going through the roof. Like 8GB dimms in the lower mac pros would have been pointless a couple years ago under most circumstances, just because they'd inflate your costs so much that you'd be better off going with a dual model and 8x4GB dimms rather than 4x8. It's getting to a point where much of this can be handled in ram, and it allows me to enable caching in many areas where it wouldn't have been practical in the past.
Quote:
Originally Posted by Marvin
There is a chance they can get away with using new CPUs/GPUs without even mentioning it (that would be quite funny actually) but I think the direction they are going points to a redesign of the enclosure. They won't do a redesign unless they are commiting to it long term and I don't think the Mac Pro has a long term ahead.
It's a bit of a stretch to think that because someone owns an iPhone, they are going to consider spending upwards of $2500 on a 40lb workstation. It's good advertising for the brand but there's been no evidence that the current 250 million or more sales of iOS devices has impacted the Mac Pro line at all.
The problem is people aren't buying them and people are managing to use iMacs just fine:
http://www.apple.com/finalcutpro/in-action/
Exactly, only upgarding when forced to due to software incompatibility. That's your typical Mac Pro crowd.
That's your typical computer crowd these days. Laptop users sometimes replace frequently, but it's more frequently due to concerns over device failure than computing needs. I'd suggest the mac pro crowd is more the type to upgrade when their needs increase, and they can realize real gains from the upgrade. The current problem is to some degree an issue of segmentation. Look at the 1,1. It granted a very competitive quad core workstation for its time. In 2008 they went to 8 cores, but many software programs were still lagging greatly in core scaling. This included some of Apple's own applications like FCP. This is a problem. You need to be able to benefit from the upgrade. In 2009 they split up the line in a very bad way. Toward the sub $3k region, they've been almost running in place since 2008 on performance as core scaling and 64 bit applications have improved the performance of that model since its release. Geekbench and other testing comes out pretty close. To see a massive improvement from that point on, it's basically the 6 core if you need the highest possible clock speed or the 12 core to make such an upgrade worth it. There hasn't really been a way since 2008 to buy into higher performance around the same price point unless it's a case of heavy ram requirements. At that point the upgrade makes sense, but it's mostly to avoid the cost of FB dimms. As I mentioned, Snow Leopard on, you have a lot of applications that saw minimal benefit past 4GB of ram that can now use as much as you throw at them. In many cases computing times fall off a cliff once it can just cache everything to ram and finding contiguous chunks is no longer an issue.
Marvin, Many of the points you make are reasonable, although I'm not really sure where your data on Mac Pro sales comes from. I just think you're drawing the wrong conclusions. There is a place for a Mac Pro in the product line. Most of the problems that exist (apart from the current economic depression) can be addressed in a redesign of the Mac Pro. Plus, Cook is an operations guy. I'd look for innovations in how the Mac Pro is built and how it is "fulfilled," and not the simple bean-counter solution of killing it because it's not selling as well as other products.
My point about the iOS halo effect is that there are increasing numbers of specialized scientific and medical-industry apps for the iPad, and as the iOS continues to be best and most secure platform for this kind of thing we will see more adoption of the OS X platform. The same is true in the entertainment industry.
Quote:
Originally Posted by Marvin
Exactly, only upgarding when forced to due to software incompatibility. That's your typical Mac Pro crowd.
This is just ass-backwards. If iMacs are so powerful and capable, why wouldn't people hang onto them until they are forced to upgrade? How is the iMac any different? Certainly not price. Your main point is that the high-end iMac = the low-end Mac Pro, correct? The total cost isn't that much different. How could it be? As you and your ilk are so fond of pointing out, Apple isn't in the business of giving customers something for nothing:
Let's say $2300 for the iMac with the high-end i7 and 2GB of video memory. We'll forget about additional RAM, since nobody buys that from Apple. Add $1200 for a 4TB Thunderbolt RAID system (including cable). So that's about $3500 for what I need.
Now let's just, for the sake of argument, say the Sandy Bridge Xeons came out last year when they were supposed to and Apple did a minor refresh to the Mac Pro line (without Thunderbolt). I think it's safe to say whatever single processor was chosen for the low-end Mac Pro (which is basically just a server that Apple also sells as a workstation), it is comparable to the 3.4 GHz i7 -- most likely either the E5-1620 ($294) or the E5-1650 ($583). Let's also assume an upgrade in graphics, but maybe we still need to add $100 for 2GB of video memory. So that's $2600 plus $400 in hard drives to get to the same 5TB storage capacity. That leaves me $500 for a nice Samsung display, but since I'm not a child and I've got a job, I'll splurge and spend $1000 for the Apple display. Apple makes a profit not just on the computer, but also on the display. Same as it ever was.
The point is that it's great that the iMac and the MacBook Pro lines are now far more easily extensible than ever before. But those solutions aren't really any cheaper than the Mac Pro ever was, and there is a place in the world for a computer from Apple that doesn't come with a display built-in, preferably in a beautiful new, sleek, rack-mountable Jonathan Ive design.
I mean in terms of video editing to store the source footage rather than a VM location - source drives are called scratch drives as they are used for render files. It could stream high bitrate source clips from the SSD due to the fast read speeds e.g 4K ProRes to a MBA at 60MB/s per stream. Not much footage (70 mins or so) at that quality on 256GB of course but you get the idea.
A certain segment of them sure but there's quite a lot of the following happens:
http://forums.macrumors.com/showthread.php?t=1306098
Trying to upgrade or fix a 6 year old machine instead of buying a new model. The initial investment is high so upgrades to new models aren't cost-effective. It's also not that easy to migrate an internal RAID from one machine to another.
I would too, I think it would be a bad move to leave it as is but where's the motivation? All the graphs show downward movement for desktops. Even if they made a stunning new small form factor Pro with the latest hardware, how much growth is there?
We always hear stories (I invented many of my own years ago) about how, if Apple just made a few better decisions with the Pro, it would change everything but they are really limited in what they can do with it and the sales are going to keep lowering over time.
There's only so long this will last. Even HP wanted to get out of the desktop business and they are one of the biggest desktop manufacturers. It's not just iMacs doing this in Apple's lineup but laptops. A Macbook Pro outperforms an entry-level MP now.
Because they run out of warranty and Apple often adds things to the iMac that are compelling upgrades like Thunderbolt, better displays, better backlight, SSD etc plus it is partly due to price. There is a fairly affordable upgrade path from a 3 year old iMac to a new one and a regular update cycle.
The example you gave about the hard drives doesn't include hardware RAID. A Pegasus R4 has hardware RAID and to add this to a Mac Pro costs $700. You also get 4TB RAID Thunderbolt cheaper ($600):
http://www.lacie.com/us/products/product.htm?id=10573
and a little less for USB 3 ($500):
http://www.lacie.com/us/products/product.htm?id=10310
If Sandy Bridge Xeons came out last year, they would match the top-end Sandy Bridge i7 but that didn't happen so now the Sandy Bridge Xeon would go up against the Ivy Bridge i7 and fall short.
Most things they do in an update are going to be really underwhelming. The top end chips that fit in the price bracket of the current 12-core are only 40% faster than the last models. If they put the following chip in a small form factor single CPU model for $2499, that would be good and reasonably faster than an iMac:
http://ark.intel.com/products/64601/Intel-Xeon-Processor-E5-1650-(10M-Cache-3_20-GHz-0_0-GTs-Intel-QPI)
and maybe this one at $3499:
http://ark.intel.com/products/64597/Intel-Xeon-Processor-E5-2665-(20M-Cache-2_40-GHz-8_00-GTs-Intel-QPI)
Quote:
Originally Posted by Marvin
If Sandy Bridge Xeons came out last year, they would match the top-end Sandy Bridge i7 but that didn't happen so now the Sandy Bridge Xeon would go up against the Ivy Bridge i7 and fall short.
Are you suggesting that if a new Macpro is released, it will include a Sandy bridge and not Ivy bridge Xeon? Intel just released the Ivybridge Xeon for the E3 series which means the E5 and E7 can't be far behind or may have already been released. I don't follow Intel CPU updates closely enough to say for sure.
All the pieces are pretty much in place for Apple to update the Macpro and really all their computers. Although doubtful, I sure would like to see a Blu-ray writer included at least as an option. Prices have dropped on media and HD home movies look great on BR and nice to share with family or friends as well as a nice additional back up for files.
http://www.engadget.com/2012/05/15/intel-launches-new-ivy-bridge-xeons-targets-microservers/?utm_source=twitterfeed&utm_medium=twitter
Those Xeons are not the type that Apple would want to use in the Mac Pro. In March Intel announced the Sandy Bridge Xeons, which are more cores and higher cache. According to earlier Intel roadmaps, we aren't expecting to see full power Xeons based on Ivy Bridge until late 2012 or early 2013. Intel is very out of sync between their consumer and server processors.
I certainly agree that the iMac has become a much more attractive platform (and I don't mean the way it looks) of late, but there are still many things for which a tower is simply a better arrangement. While the TB connectors are all fine and well, one must wait on their development and compatibility and there are performance hits involved if they are daisy chained, not to mention that, for the foreseeable future, it appears that there will be price premiums for TB external devices. Still, it is a useful technology. When TB eventually gets into fiber, a complete re-evaluation of the matter may be in order, but, for now anyway, there remains a place for towers. (The Mac Pro monicker is really misleading in many ways, but that is another matter.)
For example, in a tower, installation of a PCIe SSD (not PCIe/SATA) and a PCIe CF and/or SDHC card reader certainly will be desirable as image file sizes continue to grow...not to mention video files although they are probably better served by an HDMI input.
Don't forget the matter of RAM either. A proper tower can install more and keep it cooled better whether it is a Xeon system or not. Xeon systems, of course, can handle a lot more RAM if the logic board supports it.
Cheers
Quote:
Originally Posted by ascii
Exactly how are Apple workstations more out of date than HP or Dell ones? The Xeons have only been updated just recently, and HP and Dell only updated their workstations in April. Before then they were using the same generation of processor as the current Mac Pro.
So give Apple a few months at least to release their new models too, and we can compare apples with apples.
Joe the Dragon covered this: "but HP and others did lower prices and bumped up video cards / ram size over the same time frame.
Apple same price same ram and same video cards for same 2 years." (Post #123).
It is about the system, not just the CPU. Apple was frozen in time selling dated components at today's premium prices which never sits well with people.
There is increasing doubt that there will be a Sandy Bridge Xeon in a timeframe that makes much sense. If the Haswell is on time, there really is no particular reason to release a Sandy Bridge Xeon a few months (at most) before Haswell. It would simply clutter up the market. Of course it remains to be seen if Haswell will be on time, but the major delay in Sandy Bridge was the new process which will be utilized in the (new architecture) Haswell.
Even if Apple does produce another Mac Pro, one has to suspect that it is only a matter of time before it goes the way of the X-Serve which may lead people to seek other solutions which will suit their needs in the long term rather than riding the ship down.
I just wish Apple would let this segment of computer users know if they should switch to PC now or wait for a new Mac Pro.
For me, there is no other options...
Quote:
Originally Posted by Mike Fix
I just wish Apple would let this segment of computer users know if they should switch to PC now or wait for a new Mac Pro.
For me, there is no other options...
Exactly!!!
(Plus a whole bunch.)
USB 3 is going to cover a lot of what Thunderbolt is too expensive for and is plenty for 4K films being transferred off professional camera storage or digital images. Storage is still limited to the read/write speeds, which rarely top 1GB/s.
In terms of RAM, the dual processor Mac Pro can only manage double the iMac - 64GB vs 32GB. The single CPU one is the same 32GB.
Almost all of the benefits of the Mac Pro can largely be dismissed besides raw processing power.
With USB 3 and Thunderbolt, PCI slots just aren't that much use any more. While people cling to the specs, the reality is that the speed doesn't matter. 5Gbps for USB 3 is fast enough for almost any IO you'd ever need to do (1TB in 25 minutes) again partly because storage generally isn't that fast anyway. Data has to come from something and go to something else so the source and destination has to max the link speed.
You can have multiple GPUs but who uses multiple GPUs? Single cards can support 6 displays and you can even run displays over USB 3. You can't put more than a single high-end card in there due to the power limit so at best you get a GTX 670 in there and the iMac's 7970M will get 85% of the speed. The iMac doesn't have 6 display outputs of course but it's not a typical usage scenario.
The Mac Pro only makes sense for resource intensive tasks like rendering and even at this, it's only really the final render and 3x faster on the highest end but you can buy a cheap render box for this.
There are no reasons left that necessitates the existence of the Mac Pro. Thunderbolt was the last thing they needed to make to allow them to kill it off. Whether they kill it off now or later, they will kill it off.
Which model of Mac Pro are you planning on buying?
Quote:
Originally Posted by Marvin
USB 3 is going to cover a lot of what Thunderbolt is too expensive for and is plenty for 4K films being transferred off professional camera storage or digital images. Storage is still limited to the read/write speeds, which rarely top 1GB/s.
In terms of RAM, the dual processor Mac Pro can only manage double the iMac - 64GB vs 32GB. The single CPU one is the same 32GB.
Almost all of the benefits of the Mac Pro can largely be dismissed besides raw processing power.
With USB 3 and Thunderbolt, PCI slots just aren't that much use any more. While people cling to the specs, the reality is that the speed doesn't matter. 5Gbps for USB 3 is fast enough for almost any IO you'd ever need to do (1TB in 25 minutes) again partly because storage generally isn't that fast anyway. Data has to come from something and go to something else so the source and destination has to max the link speed.
You can have multiple GPUs but who uses multiple GPUs? Single cards can support 6 displays and you can even run displays over USB 3. You can't put more than a single high-end card in there due to the power limit so at best you get a GTX 670 in there and the iMac's 7970M will get 85% of the speed. The iMac doesn't have 6 display outputs of course but it's not a typical usage scenario.
The Mac Pro only makes sense for resource intensive tasks like rendering and even at this, it's only really the final render and 3x faster on the highest end but you can buy a cheap render box for this.
There are no reasons left that necessitates the existence of the Mac Pro. Thunderbolt was the last thing they needed to make to allow them to kill it off. Whether they kill it off now or later, they will kill it off.
Which model of Mac Pro are you planning on buying?
See bold, are you saying that a 7970M will only be 15% slower than a GTX670? Based on?
Quote:
Originally Posted by AandcMedia
See bold, are you saying that a 7970M will only be 15% slower than a GTX670? Based on?
Actually it manages to get 86% of the benchmark number. http://www.videocardbenchmark.net/high_end_gpus.html ; Yes this is for one specific software, but it has the numbers Marvin suggested.
More benchmarks for the 7970M are here (ugly page, but it has numbers) http://www.notebookcheck.net/AMD-Radeon-HD-7970M.72675.0.html
Quote:
Originally Posted by Marvin
...they will kill it off.
I happen to agree that Apple will kill it. And with its passing a number of former Mac users will find other solutions. I doubt Apple will miss them, but there are rather few applications now that are not available for other platforms which will work just fine.
Cheers
Quote:
Originally Posted by RBR
I happen to agree that Apple will kill it. And with its passing a number of former Mac users will find other solutions. I doubt Apple will miss them…
Yeah, they sure won't miss the demographic that shares their core beliefs.
So Final Cut Pro will be discontinued, too. And Logic… and Aperture… No Mac OS X Server, so everyone will be forced to use Windows or another Unix flavor… No Apple Remote Desktop, since that won't work on those OS', Anything I missed?
Can someone please help me connect my $10,000 SAS Raid array to an iMac? Or should I just throw out any gear that isn't supported by "thunderbolt"?
There is serious money spent on peripherals that will never work with an iMac, the power of the machine is only really a small aspect of this problem…
Povilas obviously hasn't invested any money into a Mac Pro studio and supporting hardware, what a spanker.
Quote:
Originally Posted by RBR
Joe the Dragon covered this: "but HP and others did lower prices and bumped up video cards / ram size over the same time frame.
Apple same price same ram and same video cards for same 2 years." (Post #123).
It is about the system, not just the CPU. Apple was frozen in time selling dated components at today's premium prices which never sits well with people.
There is increasing doubt that there will be a Sandy Bridge Xeon in a timeframe that makes much sense. If the Haswell is on time, there really is no particular reason to release a Sandy Bridge Xeon a few months (at most) before Haswell. It would simply clutter up the market. Of course it remains to be seen if Haswell will be on time, but the major delay in Sandy Bridge was the new process which will be utilized in the (new architecture) Haswell.
Even if Apple does produce another Mac Pro, one has to suspect that it is only a matter of time before it goes the way of the X-Serve which may lead people to seek other solutions which will suit their needs in the long term rather than riding the ship down.
Apple always pulls this kind of crap. The problem is that they continue to make their mac pro configurations increasingly less competitive. Why would Haswell be on time in this socket? Ivy Bridge E is scheduled for Q2 2013. Sandy Bridge E has only been trickling out. What could possibly point to a Q2-Q3 release for Haswell? You'd basically need this to guarantee the ability to order a machine using it in 2013. Intel will still make their server money either way, but I wonder if they have any plans for catching up somewhere down the line.
Quote:
Originally Posted by Marvin
USB 3 is going to cover a lot of what Thunderbolt is too expensive for and is plenty for 4K films being transferred off professional camera storage or digital images. Storage is still limited to the read/write speeds, which rarely top 1GB/s.
In terms of RAM, the dual processor Mac Pro can only manage double the iMac - 64GB vs 32GB. The single CPU one is the same 32GB.
Almost all of the benefits of the Mac Pro can largely be dismissed besides raw processing power.
With USB 3 and Thunderbolt, PCI slots just aren't that much use any more. While people cling to the specs, the reality is that the speed doesn't matter. 5Gbps for USB 3 is fast enough for almost any IO you'd ever need to do (1TB in 25 minutes) again partly because storage generally isn't that fast anyway. Data has to come from something and go to something else so the source and destination has to max the link speed.
You can have multiple GPUs but who uses multiple GPUs? Single cards can support 6 displays and you can even run displays over USB 3. You can't put more than a single high-end card in there due to the power limit so at best you get a GTX 670 in there and the iMac's 7970M will get 85% of the speed. The iMac doesn't have 6 display outputs of course but it's not a typical usage scenario.
The Mac Pro only makes sense for resource intensive tasks like rendering and even at this, it's only really the final render and 3x faster on the highest end but you can buy a cheap render box for this.
There are no reasons left that necessitates the existence of the Mac Pro. Thunderbolt was the last thing they needed to make to allow them to kill it off. Whether they kill it off now or later, they will kill it off.
Which model of Mac Pro are you planning on buying?
Marvin you're getting into the preachy crap again. The mac pro can actually support 128GB of ram, but you top out at 96 supported by OSX. Apple only mentions 64 so that you don't find that limitation. I think the quad is limited to 32 though. On multiple gpus, that would be more useful PC side with tesla cards and things like that. It's not that no one uses them. They just offer less benefit under OSX. Regarding graphics cards, don't compare until you've seen them live. Drivers account for quite a lot there. It annoys me somewhat that everyone feels thunderbolt will solve every problem simply because it was marketed that way initially. Right now you don't have a very clean solution to everything via thunderbolt beyond your imagination even if you drop all preceding purchases as sunken costs and start over from nothing. If anything thunderbolt may address some of the desire for a mid range box with slots. As to if Apple will keep it, thunderbolt means very little in that equation.
Quote:
Originally Posted by dragit
Can someone please help me connect my $10,000 SAS Raid array to an iMac? Or should I just throw out any gear that isn't supported by "thunderbolt"?
There is serious money spent on peripherals that will never work with an iMac, the power of the machine is only really a small aspect of this problem…
Povilas obviously hasn't invested any money into a Mac Pro studio and supporting hardware, what a spanker.
Something that seems to be missed is that thunderbolt only really covers those who were buying the low end of the mac pro line due to lack of a mid range tower from Apple. Even then it's not really a mature technology at this point.
Quote:
Originally Posted by Tallest Skil
Yeah, they sure won't miss the demographic that shares their core beliefs.
So Final Cut Pro will be discontinued, too. And Logic… and Aperture… No Mac OS X Server, so everyone will be forced to use Windows or another Unix flavor… No Apple Remote Desktop, since that won't work on those OS', Anything I missed?
I am not sure how much of the "core beliefs" comment is tongue in cheek.
I am also not sure whether Apple have any "core beliefs" left. Sure, the company is making money hand over fist, but there is increasingly little to distinguish it from any other computer company other than margins. Sir Jonny continues to create thermally challenged products and most of the public never notices.
The whole situation at Apple is increasingly discouraging...and don't get me started on iOS.
There are a couple of benchmarks but it shows in real games too. Here is a 7970M playing Skyrim at Ultra 1080p:
50-60fps
Here is a GTX 680 on Ultra:
It also gets 50-60 fps.
The GTX 680 will benchmark higher but in terms of gaming, small variations under even 50% difference are not really noticeable. 15% would mean 7970M = 50fps, GTX 680 = 57fps. Even the lower card in the new iMac (7770M or whatever), if it's as much a 50% difference, you just lower a few settings.
All those things run fine on all their other computers. OSX Server runs on a Mini Server and uses a fraction of the power of a Mac Pro Server. FCPX uses OpenCL so will use the large compute performance of the 7970M to at least double the performance of the quad-i7.
This makes it equivalent to an 8-core or maybe even 12-core Pro without GPU OpenCL. Obviously a Mac Pro would also have GPU OpenCL but that would be Apple competing with itself. Relative to Avid on a PC Xeon workstation, FCPX on an iMac will offer far better performance per dollar.
That's where Apple makes a statement. Outperforming the competition with a larger, more expensive workstation is expected. When they can do it with something cheaper in the form factor of a display, that's impressive.
You can buy a Thunderbolt to fibre-channel adaptor.
Ah yeah, there are kits with 16GB dimms. They are quite expensive at $1850 for 128GB, $1400 for 96GB but the prices will drop over time. Again though, this is only meaningful for people who need more than 32GB of RAM. How many people are we talking about here? Plus, Ivy Bridge iMacs might support 16GB dimms.