Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
That's actually the driver's job.

I guess it depends on your definition of "the driver". If IOKit itself (an Apple framework/kernel extension) can't handle it, then there's no chance for the underlying drivers (i.e. AMD or NVIDIA) to do anything about it.

We'll see I guess, the new design does suggest that Apple is moving in the direction of external Thunderbolt expansion only, so one would hope that this kind of thing (external PCIe chassis) is better supported than it is today. As I understand it, you can't just plug your Thunderbolt-equipped laptop into an external chassis with a GPU in it and have it just work right now.
 
Last edited:
I think the new Mac Pro is a solid step forward.

It's not for everyone right now but Apple is a forward thinking company if nothing else and I firmly suspect most of these criticisms will be moot within 2 years. Just like the iMac G3 dropping every piece of Apple legacy I/O in favour of USB and later Firewire, followed by the PowerMac G3 doing the same a year later. Followed by the PowerBook G4 in 2000. Then recently the MacBook Air, followed by the Retina MacBook Pro.

This is what Apple does, if you were expecting anything else then you've ignored 16 years of Apple's history. They remove a load of legacy, or soon to be legacy hardware, people moan and whinge because not all the wrinkles are ironed out on day one and it requires some new monetary investment, then 2 years later the PC industry follows suit.

When looking at the bigger picture this machine makes complete sense. Thunderbolt 2 isn't quite as fast as an internal PCI-E card? Thunderbolt 3 probably will be, and it will probably be cheaper as well. I know that doesn't help anyone now but I'm merely illustrating my point that whilst the bleeding edge may be hard to swallow at first everything changes in this industry within a matter of a year.

My personal opinion is that this side of the business is just not as lucrative enough as it was many years ago for Apple to spend as much time on anymore. Certainly not as much as iToys anyway. They'll keep a sizeable portion of their pro base, some will be forced to move to Windows or Unix to meet their hardware needs, especially the audio/video boys, but as Apple has grown in popularity they've gained a whole new pro base who need computers- software development. There has never been so much software on the platform!

At the end of the day I appreciate some of the concerns expressed but quite frankly you've had since 1998 to see that this is how Apple rolls and you've had since 1998 to invest in a different hardware/software platform if you don't like it.
 
Macvidcards said that it's unlikely that you need all that x16 bandwidth for a card running CUDA/OpenCL. So, what's the problem if you want to use NVIDIA cards and you put one or two in an external cage via Thunderbolt? Performance should be the same. Right?

Or I am missing something.

It's a workable solution, but there's caveats:

1. A Thunderbolt cage, drivers, etc. for said external cards needs to come into existence. This one is, I think, a good prospect, but given how long its taken things like just Thunderbolt docks to go from "announced" to "you can buy it" I'd be nervous.

2. You're now adding the cost of the cage to the cost of the graphics card. So far Thunderbolt peripherals haven't been cheap.

3. You're still paying for (worst case) two high-end FirePros you don't need.

So add the premium of those to the overall cost of the card, and the "Well, you can just slap it in an external Thunderbolt enclosure" is considerably more expensive than "Buy a Titan, stick it in the case."
 
Criticism is OK

I guess I don't like uninformed statements any more than the next guy, but I fail to understand why this board should be "supportive" of the Mac Pro. Obviously, we are all interested in the Mac Pro, but I don't want to read a bunch of fan boys any more than i want to read a bunch of haters. (OK, slightly more.) I am very interested in people's reactions, positive or negative, as long as they are factually accurate.

From a music composition perspective, I have mixed feelings about the new Mac Pro specs, but they are maybe 70 to 75 percent positive. This is by no means a push. The big deal to me is the apparent large amount of high speed connectivity accommodating SSD storage devices. That is crucial! A lot of serious computer musicians use vast sample libraries and load time is a very serious issue. The new Mac Pro, if it works as advertised, takes dead aim at that problem and that is a huge relief. Right now, a Mac Pro user has to use a PCIe slot to get the most out of SSD storage, and the competition for available slots is usually intense.

The design is largely irrelevant, although the small footprint is welcome and if the fan is really quieter, that is super important for music types.

Not so exciting: several posts above have noted that it is disappointing to see a single processor and only 12 cores. Sixteen would have been easy and 24 is certainly possible these days. I wasn't expecting a multiprocessor set up, and the Thunderbolt 2 capability raises the possibility of moving a lot of processing power off the CPU, but I have to admit it would have been exciting to see true multiprocessor support on a Mac. Similarly, the 60 GB memory limit is probably OK but it doesn't really blow you away. For most practical projects, it will be fine.

There are some design elements that suggest more improvements to come. But that's where I come up against the reality that (1) it took Apple forever to turn its attention back to the Mac Pro and (2) Apple these days is really a phone and tablet company -- computers and especially desktops are just not significant sources of revenue and profit for the company. So even if improvements are possible, you have to ask -- will they come?

As far as pricing goes, who knows? But my sense is that they will try to price these units slightly below where the comparable Mac Pros of today sit. That, to me, would be unbelievable value. I was just watching an old report on the Next Cube from 1988. That was $6,500, which is equivalent to $13,000 today (the general price level has roughly doubled since then). That thing had a 25 MHz processor, no color graphics and 256 MB of extremely slow storage. The performance-price ratio of computers has just gone off the charts in the last few decades. It will be very hard to complain about any price in the $5,000 to $7,000 range for a nicely spec'd machine.

Just my thoughts. I'm getting one for sure. And I don't even like Apple as a company - I don't have an iPhone and I think Apple stores suck -- but the Mac is a damn good product and I think the new one still will be.
 
All these critics and opinions...yet no one has hands on experience with the new Mac Pro.
 
Regarding internal PCIe slots and audio processing, we have reached a point in which "accelerator" DSP cards aren´t really needed, being effectively a dongle to ensure that you can't pirate the audio processing plugins.

Hmmm that depends. My projects, once the track count gets reasonably high, can bring the current Mac Pro to its knees, so I'm quite happy that I'm able to run UAD plugs offline. Thrilled, in fact. I know many people do native, so I can't say its impossible, but the question is can you run the plugs PLUS all the other stuff you have to run? For me, with current processor power, I still see offline processing as very value added.
 
All these critics and opinions...yet no one has hands on experience with the new Mac Pro.
There's a Foundry guy on the Modo forum who has used it, but his comments are rather general. He is impressed, though.
 
As far as pricing goes, who knows? But my sense is that they will try to price these units slightly below where the comparable Mac Pros of today sit. That, to me, would be unbelievable value. I was just watching an old report on the Next Cube from 1988. That was $6,500, which is equivalent to $13,000 today (the general price level has roughly doubled since then). That thing had a 25 MHz processor, no color graphics and 256 MB of extremely slow storage. The performance-price ratio of computers has just gone off the charts in the last few decades. It will be very hard to complain about any price in the $5,000 to $7,000 range for a nicely spec'd machine.
And it was only sold to Academics, and it floundered.
 
Welcome to the disposable computer age. Thats how the new iMac's, iPhone's and iPads are so why not follow suit for the Pro computer? Sad.... :confused: :apple:

Pretty much. I see this trend too. Everything going toward more integrated solutions, with planned obsolescence in 2 years. Even if this allows for more interesting design changes, and thinner products, I am pretty sure companies also use this a way too big excuse for non-upgradable products.
 
I posted this on another thread .... I'm wondering if Apple may indeed be "blowing us away" with a later announcement ....

but maybe wishful thinking ...


I'm not technically competent to discuss this intelligently .... but ...

What if ........

What if ....... the new Mac Pro is the basis for cluster or parallel computing

i.e. .... we could connect another Mac Pro 'module' and have a dual 12 core computer ....
 
I guess it depends on your definition of "the driver". If IOKit itself (an Apple framework/kernel extension) can't handle it, then there's no chance for the underlying drivers (i.e. AMD or NVIDIA) to do anything about it.

We'll see I guess, the new design does suggest that Apple is moving in the direction of external Thunderbolt expansion only, so one would hope that this kind of thing (external PCIe chassis) is better supported than it is today. As I understand it, you can just plug your Thunderbolt-equipped laptop into an external chassis with a GPU in it and have it just work right now.

That's not necessarily true. None of those solutions technically meet certification, although proof of concepts do exist. I also think you're mistaken on expansion. What this means is that the use of specialized components on Macs probably has a diminished future. You could see more historically pci devices built as breakout boxes instead, but that is different. I think you'll still see a lot of storage, but APU solutions have been growing in prevalence for years. Thunderbolt gpus are going to cost too much to interest most people when you factor in cost + kludge factor for the performance granted. If they do take off, it will be due to Windows notebooks. Mac users have complained about gpu selection for decades at this point, and we rarely get much in the way of third party support. Look at the EVGA 680 at $200 more than the normal ones. Add another $150-250 to that price for something that can use thunderbolt + $50 for the cable. When confronted with that, I think most of them will often put it off and later upgrade machines rather than rely on external solutions.
 
That's not necessarily true. None of those solutions technically meet certification, although proof of concepts do exist. I also think you're mistaken on expansion. What this means is that the use of specialized components on Macs probably has a diminished future. You could see more historically pci devices built as breakout boxes instead, but that is different. I think you'll still see a lot of storage, but APU solutions have been growing in prevalence for years. Thunderbolt gpus are going to cost too much to interest most people when you factor in cost + kludge factor for the performance granted. If they do take off, it will be due to Windows notebooks. Mac users have complained about gpu selection for decades at this point, and we rarely get much in the way of third party support. Look at the EVGA 680 at $200 more than the normal ones. Add another $150-250 to that price for something that can use thunderbolt + $50 for the cable. When confronted with that, I think most of them will often put it off and later upgrade machines rather than rely on external solutions.

Basically, Apple has just made a VERY good case for switching to Windows. Their arrogance has reached a new high.
 
That's not necessarily true. None of those solutions technically meet certification, although proof of concepts do exist. I also think you're mistaken on expansion. What this means is that the use of specialized components on Macs probably has a diminished future. You could see more historically pci devices built as breakout boxes instead, but that is different. I think you'll still see a lot of storage, but APU solutions have been growing in prevalence for years. Thunderbolt gpus are going to cost too much to interest most people when you factor in cost + kludge factor for the performance granted. If they do take off, it will be due to Windows notebooks. Mac users have complained about gpu selection for decades at this point, and we rarely get much in the way of third party support. Look at the EVGA 680 at $200 more than the normal ones. Add another $150-250 to that price for something that can use thunderbolt + $50 for the cable. When confronted with that, I think most of them will often put it off and later upgrade machines rather than rely on external solutions.

Which is why Apple did it..
 
Basically, Apple has just made a VERY good case for switching to Windows. Their arrogance has reached a new high.

They're always arrogant. I've tried to point out that it's generally a balance. It could always come in cheaper than I expect. If it's another round of cut features, raise price while trying to smooth it over with a new design, I don't think that will go over well. I expected things like dual 5.25 bays would be gone with a redesign, but using dual socket parts when they're only going to populate one strikes me as excessively expensive. Looking at Sandy Bridge EP, I can't see where a single 8 core 2400 cpu at roughly $2000 would make more sense than 2 6-core types at less than half the cost. I don't think they'll chase high core counts on the 1600s cpus. I guess you might see 8, as 6 is clocked fairly high with the current generation.

As soon as they unveiled it, I knew people like the OP of this thread would appear to tell the rest of us how we don't deserve Apple.

Which is why Apple did it..

I know that. I'm trying to point out why thunderbolt gpus are not something that should be counted on.
 
They're always arrogant. I've tried to point out that it's generally a balance. It could always come in cheaper than I expect. If it's another round of cut features, raise price while trying to smooth it over with a new design, I don't think that will go over well. I expected things like dual 5.25 bays would be gone with a redesign, but using dual socket parts when they're only going to populate one strikes me as excessively expensive. Looking at Sandy Bridge EP, I can't see where a single 8 core 2400 cpu at roughly $2000 would make more sense than 2 6-core types at less than half the cost. I don't think they'll chase high core counts on the 1600s cpus. I guess you might see 8, as 6 is clocked fairly high with the current generation.

As soon as they unveiled it, I knew people like the OP of this thread would appear to tell the rest of us how we don't deserve Apple.



I know that. I'm trying to point out why thunderbolt gpus are not something that should be counted on.

'twas not a criticism just pointing out the obvious and adding my person criticism, this stuff is quite polarizing.
 
They never said anything about fall. They said later this year. It could even be December.
I can only assume they want it on the shelves by Black Friday.
The Mac Pro hasn't been a huge seller for almost a decade. I think one of the motives behind this Mac Pro "reboot" is to get more mainsteam/prosumer buyers. I fully expect a price drop, heavy rotation ad spots and availability in time for the holiday season. Sure they could launch in December, but historically if a product hits the shelves in December it has already missed the holiday season.
I expect an early fall launch with product on shelves mid to late fall.
Fingers crossed for a new matching keyboard/pointing device.
 
Technically, Fall runs all the way to last day before Winter Solstice ( Dec 21). Fall is in December too. :) Colloquially no. However, Apple has never said Fall or Q3 2013. Those are always the echo chamber quoting back "telephone game " transformations of what Cook said back in 2012.
the 2012 27" iMac had display manufacturing issues. I can't think of anything comparable with the 2013 Mac Pro except possible stepping differences between manufacturing sample CPU and GPU and later production versions. Or of course there could be some teething issues with the stateside manufacturing getting on line too.
Either way, this is bigger than the iMac losing an optical and getting thinner. The Mac Pro is kind of a flagship product (vanity product?). I would think Apple is moving heaven and earth to stay on schedule to get this out in time for the holiday season. If it debuts in late Dec or January initial sales will still be good but not stellar, making recoupment of R&D costs harder.
And I am also hoping to buy one for my birthday in Dec, so hella wishful thinking on my part.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.