New Mac Pro - Page 4

They use the same basic protocols according to intel. For some reason everyone on here is convinced that they can only be routed through integrated graphics connections when intel has stated they support both displayport and PCIe protocols. A couple of them also claimed that it is impossible to route through PCIe or any machine with such a standard in place.

I knew this wouldn't be the case simply because Intel wants to encourage widespread adoption on the windows and linux side.

Intel has been pretty good about evolving hardware in a rational manner.

As to Linux I'm surprised that hardware and drivers haven't surfaced yet. I guess Apple must have gotten an exclusive here.

That makes me sad. I've always had a soft spot for Linux. Aside from that I'm not sure why so many people claimed it could be routed solely through intel's integrated graphics. Then there are the responses I get that the mac pro doesn't need it because it has PCI slots. To me it provides a cheap standard for attaching a secondary hard drive enclosure (a four drive limit is nothing ). I imagine there will soon be a way to attach it to a proper NAS, but regardless of that as more thunderbolt devices emerge we'll want the ability to share devices between a macbook pro and mac pro. It's really clunky when your laptop can run something that your desktop cannot.

I haven't seen (or perhaps noticed) these claims that the next Mac Pro doesn't need or won't get Thunderbolt. I wouldn't take them at all seriously.

One of Linux's great weakness is the slow way hardware gets supported. Frankly I' m leaning towards AMD these days as a stronger supporter of Linux. At least from the standpoint of GPU support.

Then again if you are like Intel with crap GPUs your Linux support will likely be weak.

Some of the discussions I've seen clearly indicate that people think Thunderbolt replaces slots. This is contrary to what I think. Seriously if it is Apples intention to replace slots with TB ports they are in serious denial. Right now I'm hoping that they see TB as no more of a slot replacement than USB was.

I believe TB has a greater potential than that. However I can see TBs primary usage being disk array interfacing.

The thing here is there are lots things commonly used externally that can use the speed of TB. TB might even enable new uses or concepts, it seems like TB is destined for instrumentation interfacing. I suspect what will make TB interesting is the new or novel uses to come.

I think this reflects upon the Mac Pros seriously bad sales figures. Apple most likely figures there is little to loose. It is the right machine for far to few people.

Thunderbolt is an imperfect substitute for slots. From Apple's perspective, Thunderbolt is an improvement over slots.


Few customers are willing to screw around with slots.

They've just been forced to during their lives as PC owners. Apple removes that possibility from their realm and picks up people who couldn't care less how it works as long as it works.

Don't sue me, Matrix franchise.

We can think of customers as falling into three categories.
1. Customers who take their PeeCee to someone else because they absolutely won't screw around with slots.
2. Customers who will screw around with slots if they have to, but would rather be doing something else.
3. Customers who love screwing around with slots.

I'm in category 2, but I suspect well over 90% of customers are in category 1. I know very, very few people in category 3.

I believe that the overwhelming majority of current Mac Pro customers are in either category 1 or category 2 and would be happy with a slot-free Mac Pro. I believe that the number of category 1 customers that Apple would pick up by eliminating slots from a future Mac Pro would overwhelming exceed the number of category 3 customers they would lose as a result of such a move.

The above is mostly non sense. The people that care about slots are the ones that use them. That should be pretty obvious and just as obvious is the idea that if you don't use them your input isn't needed as Apple has plenty of hardware for you.

As for category #3 it doesn't exist, people don't use slots to screw around. Rather they are used to configure a machine to their needs. What is sad about this debate is that people who have never touched a slot can't see what they have to offer a more advanced user.

I think your trifecta is a bit of nonsense. If you need those slots their use is not screwing around.

Again garbage as one of the big reasons people buy Mac Pros is the slots, with the number one motivator being a high performance video card.

That also makes no sense at all. Apple already has plenty of slot less machines to service the people that need that sort of hardware.

At least one of us (and perhaps both of us) has a screwed perspective.

I know three people who buy a new Mac Pro every time Apple release a new model. None of the three ever make use of the slots. One of them certainly knows about the slots, but I rather doubt the other two do. They buy these machines because they need the fastest machines they can get. Right now, all they have 12-core Mac Pro models with loads of RAM. These three people may or may not be representative of Mac Pro buyers, but I suspect they are. I haven't heard of anyone putting a card (not counting DIMM cards) into a Mac in this century, though obviously some people have.

I have a great deal of difficulty believing that most Mac Pro buyers would not be entirely satisfied with screaming fast graphics directly on the motherboard rather than on a replaceable card. I think nearly all Mac Pro buyers care about the machine's performance, not about opportunities to tweak that performance. My guess, which could be significantly wrong, is that about 1% of Mac Pro buyers care about slots.

The argument that there are plenty of Mac models without slots to satisfy people who don't care about slots is based on an implicit false premise: that people who don't need slots don't need a fast machine. I think the overwhelming reason why Mac Pro buyers buy a Mac Pro is for the performance. Anyway, I'm prepared to wait to see what Apple will do next.

Though it could be possible, I don't think that many of these Pros are being installed without Video cards. Right now the only way to a high performance GPU is via a slot. Now I'd be the first to admit that the actual users may not care about where that GPU is, but never the less slots are required for the card.

Right now slots are the only way to high performance GPU technology. I'd also be the first to admit that at some point in the future this might not be the case. In a couple of generations or so the best performance will likely come from highly integrated GPUs and CPUs on System on Chips. That isn't today though.

In a nut shell every Mac Pro is employing it's slots to some extent.

Actually if you have read some of my XMac posts in the past you would see that we agree somewhat. But there are still qualifications. An XMac is not a Mac Pro, not by a long shot. Second I'd still want a couple of slots in the XMac. Ideally the slots could support an additional high performance video card.

To repeat my point made previously the Mac Pro is a good machine for those that need it. It is basically a workstation / server like machine. The problem is it is way to expensive for what many of us want. That is to put it simply a low cost machine with easy expansion.

When you use the word tweak it tells me you dont have a grip on the issues involved. It isn't tweaking at all, rather it is configuring a machine to make it usable for a purpose. In other words it is about being able to use Apple hardware or being forced to use other hardware. If one can't build functional systems with Apple hardware the alternative is Linux machines on generic hardware.

Generally people that need solid or fast performance these days need slots. There is no way around it as that is where the fast GPUs are. Apple could certainly address this in a new machine, but they have repeatedly failed to do so. Even the new Mini with the discreet GPU is a half hearted effort, that leaves you feeling that they simply don't care about performance at any level. In the end if Apple did integrate a GPU on the motherboard of a suitable machine, they would likely castrate it to the point that a card in a slot would be required.

Yes it is a long wait. I smell a major overhaul to the Mac Pro simply to save it from XServs fate. What Apple will deliver is a mystery. However I suspect it will distill some of the ideas we have discussed here.


As a side note from the land of dreams I kinda wish that Apple would implement AMDs coming Bulldozer base Fusion products in an XMac. That would allow for low cost and very good overall performance. It would give us a machine where the GPU is good enough for most and with slots capable of a quick boost in video or OpenCL performance.

You may note my focus on GPU performance which I consider to be very important and like to become even more important. There are a couple of reasons. The first is Retina class displays on laptops and desktops which could quadruple GPU demands. The second is the continued expansion of the use of GPUs for OpenCL and special purpose functions. The iPads highlight just how important GPUs are these days in delivering a good user experience.

Apple can't get rid of slots, period. Graphics cards, NICs, HBAs... too much backwards compatibility is needed for the Pro market

I'd like to believe that is the case. However Apple could go in a direction that leaves the Pros sitting high and dry. The problem with the current Mac Pro is getting enough business out of the platform to justify its existence.

True, but completely irrelevant to the question of what Apple may do with a future Mac Pro.


There is absolutely no reason why Apple couldn't take the very same chips that are now found on high-end graphics cards and use them directly on the motherboard of a future Mac Pro. Slots are absolutely not required for high performance GPU technology.


I understand that is what you and some other people want. I don't believe it makes any sense from Apple's perspective.


Yes, exactly! However, the xMac you propose would also be unnecessarily expensive. Drop the slots and you bring the cost down.


If you want to hang your hat on some perceived nuance in meaning between "tweak" and "configure" to make a personal insult, then be my guest. It doesn't strengthen your argument.


That's true, but you still haven't given a single example of a solution that is possible with slots but not possible with Thunderbolt.


Exactly! Apple could address this in a new machine, you propose a new machine, but your xMac proposal doesn't address this. I believe the next Mac Pro (or whatever replaces it) will address this by dropping the slots and putting a fast GPU directly on the motherboard.


The Mini is not a high-end, high-performance machine. It's a low-end, low-cost machine. Your complaint that the Mini doesn't include high-performance graphics does not even begin to suggest that a future Mac Pro won't include high-performance graphics.


You have provided no evidence, no facts, no logic, no argument to support this wild assertion.


Yes, it would be a surprise if we're both wrong on every point in this thread.


I'm impressed by your remarkable ability to make your pitch for AMD over Intel in every thread regardless of the topic at hand.


I agree with you that GPU performance is increasingly important. Where I disagree is with the idea that it needs to be implemented on a discrete card rather than directly on the motherboard -- the possibility of which even you have conceded above, while denying it in other paragraphs.


All the items you list can be implemented with Thunderbolt.

No, they can't. Educate yourself.


I don't believe they make that mistake again (see Final Cut Pro X release). Apple re-learned a valuable lesson, the Pro market needs backwards compatibility features. I don't have any actual numbers but I believe the Pro market is justified, for now.

Just because you don't how to implement graphics, Ethernet, and HBAs over Thunderbolt doesn't mean it can't be done. It can. Ethernet over Thunderbolt is trivially easy to demonstrate because Apple have already done it with the Thunderbolt Display. Implementing HBAs over Thunderbolt is a very similar problem. As for graphics, there is already work being done on HiDPI monitors with the GPU built-in to the monitor and driven by Thunderbolt.

I never said that slots are required in the future what I've said is that right now it is the avenue to high performance GPUs. Even if integrated on the motherboard the chips interface will still be PCI Express.

Nothing about Apples desktop line up makes any sense to me and stands in sharp contrast to their laptop line up. Think about it, Apples laptops are now easier to service and support than most of the desktop line.

Simply put the desktop line up needs to make sense from the customers perspective. In the end that is what grows sales. Apples approach to the desktop seems to be to throw their hands in the air and whine about the beauty of the iMac.

BS. The whole point of this machine is to have the slots available. The machine would not be anymore expensive than any other desktop platform from any number of makers. XMac is by no means a Mac Pro replacement.

Where is the insult???? Seriously you call people tweakers because they need slots to accomplish something important to them and you don't expect a response.

I gave you a single example already, that being video cards. I'm convinced though that you don't understand the difference between PCI - Express slots and TB. TB has three issues working against it, it is slow, external and it is serial.

You may ask what does external have to do with it to which I have to say look at the Mini Display Port connector, cabling requirements and external chassis.

I have nothing against a fast GPU integrated on the motherboard but for the fact that Apples history here sucks royally. Even if they did implement a decent video subsystem that still doesn't eliminate the need for slots.

My fear here is that Apples motivation here is to make the video hardware easy to integrate with TB not to deliver good GPU performance.

Why the need to twist things up here. I never said the Mini was a high performance machine, I said that the Radeon machine had a need for more video RAM. 512 MB of video RAM is not a high performance implementation these days.

In any event the current Mini in combination with Apples history dies indeed suggest that Apple would under deliver in such a Mac Pro. It is almost like it is in Apples DNA to minimize a given GPUs capability.

All I need is Apples history here.


Sadly I'm hopeful but not convinced that anything compelling will come out this year nor even next.

It takes talent!!!

On the serious side I think it is very important for AMD to remain viable as a competitor to Intel. We have seen how slow intel can become when not under pressure from rivals. Even more important is the idea that the GPU will just become more and more important over time which puts AMD at an advantage.

Frankly I wish that Apple would market at least one line of AMD based hardware, that is hardware with an AMD CPU in it. It would give people a choice and frankly keep Intel on it's toes.

As a side not it is really good to see Intel getting hit from different sides these days. AMD does what it can do from i86 space and the ARM camp is shooting from below. I have no doubt at all this is pushing things along at Intel like Haswell.

Again I've never said that, I said we don't have a choice right now. Today, as in right now, if you want high performance graphics or GPU facilities you need to buy a card. I've also have pointed out that Apple doesn't have a history here worth admiration as they have constantly under powered the GPUs in their machines. They have only moved to respectable in the last couple of years on the iMac.

As to video facilities in general if you have followed my other posts you will see that I expect the need for discreet GPUs to rapidly decrease. This change over is actually happening right now but will be acceptable to a wider array of people with the next generation of chips. In effect you have your on motherboard video right there.

The question then becomes this how do you deal with the limited need from people that need or want a discreet GPU? The answer is that we are back to slots again. Other wise we end up with a Mini like solution. That is a basic Mini with SoC graphics and a discreet GPU in a half a$$ implementation.

You can not do anything with Thunderbolt that requires more than 4x PCI - Express speeds and you can't even do that with all devices due to the serial bus nature of TB.

Your arguments here are non sense. Right now you can do Ethernet and video over USB, that doesn't make it acceptable for every use.

As to those HiDPI displays, do you really think that TB has the bandwidth to do that well for every application? Especially when that TB port may have other devices dangling off it. Consider this those USB to video adapters do function but who amongst us wants his primary display ran through one? Go to a Hi DPI display via TB and the situation is much the same, your bandwidth isn't there. Even if you manage to compress the data and commands down you are still trying to push to much through a tiny hose.

Oh by the way do you really want to go to the trouble of these adapters when things like Ethernet can be built into a computer trivially? This whole adapter mentality just drives me nuts.

Belkin has a FW800, ethernet and USB 2 dock connecting over Thunderbolt. Sonnet has FW800 and ExpressCard (which itself has USB 3). ViDock and Sonnet have Thunderbolt solutions for running an external GPU.

So yeah, there are products to do pretty much anything and everything internal PCI expansion slots can do. At the moment, the bandwidth is just lower, which will impact a small minority of internal slot usage.

Looking inside desktop PCs, it's clear to see even at a glance how much space it saves removing internal slots. Smaller power supply, at least 1/3rd of the internal volume reduced, no requirements set for the thickness of the machine to allow for card height, no giant holes at the back of the machine, fewer fans required, smaller enclosure (reducing weight).

The discussion is about the next Mac Pro. That the current Mac Pro implements graphics on cards is irrelevant.


It makes perfect sense to me.


Products need to make sense from the mainstream buyer's perspective, in this case mainstream Pro buyers. Mainstream Pro buyers just want performance.


Apple's approach is making billions of dollars.


That's why Apple will never make the xMac you want. Sorry to be the bearer of bad news. Your xMac is a nice fantasy, but it makes zero sense from a business perspective.


Apple's history here is putting very fast graphics in the Mac Pro. I have no reason to think Apple would change that if they integrate the graphics onto the motherboard in a future Mac Pro. You have given no reason think Apple would suddenly start putting "castrated" graphics into the Mac Pro, which they've never done before. The fact that low-end Macs like the Mini ship with "castrated" graphics is irrelevant.


I agree that competition is important.


Again, the design of the current Mac Pro is not a constraint on the design of the next Mac Pro.


Again, you're comparing grapes to watermelons. A history of weak graphics at the low end and very fast graphics at the high end does not suggest that the future high end will have weak graphics.


Another grapes to watermelons comparison. There is no reason why graphics on the motherboard cannot be just as fast as graphics on a card in a slot. There is no reason to expect that Apple would cripple the graphics of a future Mac Pro.


What are you on about? First, Ethernet over Thunderbolt is fast enough for everyday use. Second, no one has suggested that a future Mac Pro won't come with a gigabit Ethernet port. The only reason we're discussing Ethernet on a card in a slot versus Ethernet over Thunderbolt is for the applications that require a second Ethernet port.


First of all, you demonstrate that you're out of arguments when you call Thunderbolt "a tiny hose".

More interestingly, no Mac Pro can be optimal for every application. I've already stipulated that Thunderbolt is an imperfect substitute for slots. Some customers would be better off with slots. Most customers would be better off without slots. Again, we're back to giving up a few customers with arcane needs in order to pick up far more customers with more mainstream needs. That sucks for the customers with arcane needs that can't be satisfied by Thunderbolt, but it's good business for a company the size of Apple.

It gives a benchmark to judge against. No matter what Apple does with the mac Pro it will be judged against previous versions.

Well maybe that is why we disagree.

Yep! Seriously we agree here, the question is will a Mac Pro with a built in video card give pros the performance they want. It certainly could be done, but what i'm saying is that Apples history sucks here. There is no way we can be certain that Apple would implement the type of GPU performance Pros want.

Off laptops but little if any off the desktop lineups. I wouldn't be surprised to find out they are in the red with the Mac Pro.

How is what I said an example of why Apple would never produce a XMac? it makes all the sense in the world from a business perspective as they can pick up a whole class of users they currently can't reach. So really I don't get what you are saying, is it that Apple isn't capable of designing a desktop machine.

Every implementation of the Mac Pro up until now has come with a marginal GPU card. I can't even imagine how you came up with the above statements.

I like that intel is feeling the heat from multiple directions, it makes for more innovation and novel thinking.

Are you being dense on purpose here?

OK lets just say you are the only one here that believes the Mac Pros come with high end graphics.

Actually there are technical issues to consider. A lot of engineering goes into making those graphics cards work. Thermal engineering being a big part of the effort. While Apple can certainly manage this if they really wanted too, I'm simply not convinced they would want to.

Lets face it the phrase "Ethernet over Thunderbolt is fast enough for everyday use" could be and has been replaced with: "Ethernet over (USB, SD, Firewire) is fast enough for everyday use" in the past. The problem is it is less than optimal solution and there is no reason to believe it would be any better on Thunderbolt.

Nope it is entirely fitting because that is exactly what it is compared to 16X PCI, especially if that is next ten PCI-Express. Thunderbolt is a serial channel or hose if you will that has a limited capacity. You can't really argue the point here.

This is where I disagree, especially in the context of a Mac Pro. Slots make the machine a pro level tool. If you remove the slots you end up with a fat Mini with all the draw that device has for pro usage.

That is total rubbish. Are you actually saying that Apple can manage to have one machine in its entire line up of Macs that comes with slots? Seriously, the removal of such a device would be a very significant blow to Apples credibility with respect to the Pro market.

So it is good business to eliminate the one option that customers have to use Apple hardware in anything more than a trivial manner? This I just can't agree with.

You can't be any more certain, not one iota more certain, that a hypothetical updated Mac Pro with slots and a discrete graphics card "would implement the type of GPU performance Pros want." I do concede, however, that in the slightly far-fetched scenario that Apple would fail to "implement the type of GPU performance Pros want" that slots would, in the very near term (until perhaps about 2013), simplify working around that limitation. In the longer term, when high-end monitors have their own built-in GPU (from perhaps about 2013), Thunderbolt would be the simpler, easier, and less expensive way to work around such a limitation.


You're kidding, right? I would eat my hat if Apple were not making money on the Mac Pro. I would bet that their margins on the Mac Pro are above 30%.


I believe that Apple could reach far more new customers by dropping the slots. We're probably going to have to agree to disagree on this.


!?!?! I'm the one who is confident that Apple can design and build a fast machine with fast graphics without slots. You're the one who seems convinced that Apple cannot do so, that "it is in Apples DNA to minimize a given GPUs capability."


I thought we were better than the five year olds in the sandbox calling each other "meanie".


Now we're getting somewhere. Thank you for conceding the major point of the discussion: that Apple can build a Mac Pro without slots that would have all of the graphics performance of a Mac Pro with slots. I agree that whether or not Apple want to build such a machine is an open question.


The question was not whether or not Ethernet over Thunderbolt is optimal. The question was whether or not it can be done. I have already stipulated that Thunderbolt is an imperfect substitute for slots.


Every interface has limited capacity. There is no point against with I could argue.


Let's imagine that Apple assign two engineering teams to independently design the next Mac Pro. Both teams are given the same list of features including which Ivy Bridge CPUs they must support, which graphics chipset, ports, etc. The only difference is that Team A is told to use the existing Mac Pro case and Team B is told to design a smaller case without expansion slots. The two teams present their designs and, no surprise, the benchmarks show all performance parameters within 1% of each other, one insignificantly faster in some tests and the other insignificantly faster in other tests. Team A's design will need to be priced at $2499 to $4999. Team B's machine will be half the weight, half the size, consume half the power, produce half the noise, and be priced at $1999 to $4499. Which design should the management team choose to produce?

I understand it's your belief that Apple could sell more of Team A's machine. It is my belief that Apple could sell more of Team B's machine. I've already stipulated that there are some customers who would buy Team A's machine but not Team B's machine, however, I believe there are far more customers who would buy Team B's machine but not Team A's machine. We may have to just agree to disagree.


Now there's a strong argument.


I have not written that, but it's obviously true since Apple currently "have one machine in its entire line up of Macs that comes with slots." I don't see how it's at all relevant to the argument at hand.


I think Apple would gain credibility with the Pro market by giving them the performance they want at a lower price point by dropping legacy cruft that very few Pros either use or want.


That's a straw man argument.

When NeXT replaced the NeXTcube with the NeXTstation, they did exactly what I'm suggesting that Apple do with the Mac Pro. In that case, graphics didn't suffer at all. Graphics improved dramatically when NeXT dropped the slots and switched to a more compact machine. NeXT were able to cut the price by half from $9999 to $4999. Of course, there were a few people who complained that NeXT were abandoning the Pro market, that Pros needed slots, etc. Sales increased many-fold.

Here's something to at least advance the discussion, how do you think quality graphics would be implemented in a slotless computer, and has intel mentioned anything about their roadmap for thunderbolt/copperpeak bandwidth? The mac pro line has inflated considerably over time, so yeah it's higher than it would need to be, especially if it was selling in higher volume.

Regarding thunderbolt bandwidth, I don't see it being much of a multi device standard with the current bandwidth. Currently it can handle the equivalent of 1 PCIe 3.0 lane, so even if you internalized the graphics, discreet graphics will be running off PCI rather than from a thunderbolt chip regardless.

Last thing, the slots aren't what keep the price high. Apple charged less for more on that line in previous years and the slots were there that entire time. If they take out the slots and reduce the price, it's simply an excuse. I wouldn't care so much if thunderbolt grew closer to what PCI is today in terms of raw throughput.

In the context of the next Mac Pro, my guess for the most likely implementation of the graphics directly on the motherboard would be the ATI Radeon 7xxx "Southern Islands" GPU (with at least 1GB of memory) using the PCIe 3.0 protocol to communicate with the CPU.

Intel's roadmap for Thunderbolt includes plans for a 10x increase in bandwidth. After that, who knows?


Of course. I don't think anyone has suggested using Thunderbolt for internal interconnects.


See Marvin's excellent post above.

If they stick with high-end cards, they can have a single PCIe slot but it would have to hold a GPU. They could go the iMac route (which I hope they do) and use an MXM slot with a mobile GPU like the Radeon 6990M. It's half the performance of a high-end PCI card but they only have a 100W power draw and still support double-precision computing.


There were examples of Thunderbolt products at IDF:

http://www.engadget.com/2011/09/14/t...ss-dock-seaga/

It's best to think of it like a better version of ExpressCard, which has been in a large number of laptops over the years.


I'd expect them to remove the 5.25" bay too that holds the dual optical drives. If you look at the small box on this page:

http://www.boxxtech.com/products/ren...o_overview.asp

it has up to 12-core Xeon processors and 192GB RAM. That looks like it's about 1/5th the size of a Mac Pro. Add 4 storage drives, an MXM GPU and you have a very small and very powerful machine. You should be able to buy multiple models and chain them together with Thunderbolt for a very compact personal render farm.

You are compromising on the GPU but it's not really going to affect a large amount of customers and it drives them into a more regular upgrade cycle because it's easier to sell the machine on again.

Or Apple could go the Mac Mini route (but with a high-end GPU, of course) and put the GPU directly on the motherboard. Like this:
http://www.techrepublic.com/photos/c...6265433?seq=52

Especially when the entire industry is moving the GPU as close as possible to the CPU? I just can't see a rational argument for stuffing a GPU into a monitor. Especially when process shrinks allow for ever increasing GPU capability right in the SoC. At best you end up with a short term fix embedded in a long term device.

Or to put it another way your monitors often outlast the hardware driving them. At least that is the way I see it. Even a monitor for a laptop would quickly be eclipsed by the laptops GPU in two years or so. Hey maybe those monitors will come with the GPUs in slots.

The boxx you linked may have compromised on the gpu, but it's because you linked me a render farm node . Regarding gpu, why would you want to gimp it? If anything they're becoming more important/used in a lot of areas outside of gaming and 3d. By the way even that has a spare PCI slot but you couldn't run discreet graphics over it due to the power draw. You'd have to be careful on hard drives as well. From what I can tell it looks like the power supply is only 300 watts due to heat concerns. It's designed as a compact server regardless so it has some limitations as a workstation. I do like the form though. Apple could learn something from those guys . They are a bit pricey but they custom build in Texas, so it tends to be expensive.





Higher end displays have a ton of electronics packed into them, but given the increased reliance on the GPU, I don't see it being moved out of the box. Extra heat is also bad for the lcd seeing as most of them rely on passive cooling.

I'm new here, but I've been reading this discussion and there are a lot of interesting predictions/ideas here.

I don't think Apple will be exiting the pro market. I think at this point they will treat the Mac Pro as an evolutionary product rather than a revolutionary one... with few frills in order to satisfy professional users who need raw processing and graphic processing power.

I'd guess they'll integrate thunderbolt into the new model with significant TB expansion (say, 3 or 4 TB ports)... but will also leave the PCI Express expandability.

I see one of two things happening:

1. The machine sees integration of thunderbolt with only a minor case redesign to reflect this and upgraded specs. They might get an early in on Sandy Bridge-E.

2. They redesign the case significantly to integrate TB, possibly add USB 3.0, and reduce, but still keep PCI-E slots for more limited, but available expandability. They might do away with the optical drives in order to shrink the case but keep room for 4 HDD bays.

I expect the price point will remain the same.

They will focus on PCI-E for graphics with TB for everything else.

I don't think the changes will be very drastic.. other than increased CPU and GPU performance for high end users that need it. They will continue to push prosumers and lower end professionals to the Mac Mini or the MacBook Pro with TB. The MBA and iMac and dual-core Mac Minis will remain the choices for consumers.

So, given the information on Intel's release dates, when are we expecting/estimating to see a refreshed Mac Pro? I really want to buy before the year is up.

When it's out. We're really not expecting anything.

Processors appropriate to the mac pro are supposedly coming out November 15th. Apple has announced machines using new processors early, and they've announced them months later. It's a safe assumption that you won't see these before mid november, but it could be next year.

I would think that we would start to hear leaks and other bits of evidence that would indicate that the new machines are about to come. At this point we have heard nothing. So it would be best to describe them as months off.

That is if they upgrade the machine at all. I suspect that we will see Mini and iMac updates before anything Mac Pro related happens.

Apple isn't going to just give up on the Mac Pro market. While nobody was buying the XServe (according to Steve Jobs), I find it hard to believe that Mac Pros are selling so poorly that they should just be phased out.

Obviously they are not priority number one for the company, which is dealing with the transition to Tim Cook, and perhaps we won't hear anything for a couple months... but the Mac Mini and iMac do not cover all of the market.

I think the biggest recent bit of news that Apple isn't giving up on these people was when they made the last generation of Final Cut Pro available and sought to address the many complaints of Final Cut Pro X in their most recent updates (of which there will likely be more).

I think the best and most convincing reason not to abandon the Mac Pro is that Apple has to make SOME kind of machine to develop all those wonderful iOS apps on!

I can't imagine it would go over well if Tim Cook/Steve Jobs came out and said "sorry.. but if you want to develop for the iPhone or iPad, you'll have to do it on a Mac Mini... or a PC". Not with the cut that Apple takes from every App sale... you bet they're gonna keep the Mac Pro around.

it isn't a question of the Mac Pros having a market, they do have one. However it is a high end market. The problem with the Mac Pro is all the missed opportunity. That is the market for an expandable machine in the $1200 to $1500 dollar range.

The transition happened long ago it was just made official recently.

You are right, the Mini and iMac don't cover all of the market but adding the Mac Pro to the mix covers little more. It is just to expensive for the desktop.

What Apple is giving up is the market that wants an affordable but expandable Mac. Simply put not everybody needs a Mac Pro nor the massive box.

The vast majority of that development is done on Mini, iMacs or laptops.

Can't be serious? I have a rather old MBP that still does a credible job developing code.

How is Apples cut in anyway involved in a developers selection of developmental hardware? I don't follow your logic at all.

How I wish someone at Apple was reading this and paying attention. This is the only Apple product I would be crazy enough to camp outside an Apple store to buy.

You can buy one on Craigslist or EBay. I bought one for under $1000. Now I'm looking for another. If Apple won't build a lower cost tower, someone else is getting my money. I too hope the right people are reading this forum, but will not hold my breath. How long ago was it that someone started the xMac discussion? I remember those who said, "It will never happen." They have been right so far, but I hope it changes.

I think Apple has claimed to serve business too. However, I've not seen any Macs in retail stores or reception desks. Concerning education, a local institute has put 60 Mac Pros on Craigslist in the Portland, OR area. These three-year-old Mac Pros are being replaced by Windows PCs.

It's kind of amusing to see the pundits here claim Apple is missing out on a great opportunity by not making an xMac when the company just sold 4.63 million computers and 12 million iPads. Laptops account for the vast majority of their computer sales, too.

And we're not even talking about iPhone sales.

I bet that if the xMac was a viable business proposal, Apple would already have one out. Fact is, they know the market better than most and their track record/sales figures prove it.

Face it, they know better. In fact, I use iDevices for pretty much everything and use the MacPro only for audio. My iPhone takes care of 80% of my internet usage, and I'm typing this on an iPad.
I'll go to the studio later and fire up the MacPro to work. That's when I use desktop machines. I have an Asus laptop that dual boots into Vista64 and Ubuntu 11. It sits in the living room as an internet access machine, but I'm usually too lazy to bother and just whip out my iPhone.

And I love my iMac, it is like a friend, but since I got the MacPro for work and the iPhone for everything else, I only switch it on maybe two or three times a week. And that is mostly out of a sense of guilt, too.

I stopped trying to get through to them long, long ago.

Take it from me, they won't quit whining. Ever.

And don't even think about suggesting they shut up and buy a PC. Because even though it's exactly what they claim to want, they don't want it.

Apple is making money hand over fist. Fine. But the fact remains that this Mac user of 18 years isn't buying anything new from Apple because what I want isn't being made. So Apple in its infinite wisdom is successful but it is also very close to losing a long time buyer and user of Apple computers.

Can I make Apple build what I want? No. But I think they should be wondering what would happen if more of the long term repeat buyers of Apple products start feeling the same way I do.