Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
nMP/new Mac Pro/Mac Pro 6,1 Reviews [MERGED]
- Thread starter VirtualRain
- Start date
- Sort by reaction score
Heaven is single GPU, right?
I'd like to see multiGPU + CPU benchmarks at the same time and see how loud/what the performance is.
So you can contemplate those results?
None game and very few apps uses it
Rather have one that works than two just to say I have two
I thought it said they stressed out the whole system, but on looking at it again it's not clear. However
I've had zero doubts it would actually be that quiet, with that design and Apple sensibilities. And this is a major reason I'm getting one, I hate noise and my 2009 MP makes enough that it only doesn't bother me playing games.
There was a German review which called it noisy.
The way the "thermal core" is designed, it seems to me it's not that much more efficient in terms of cooling ability, just that it consolidates 3 heat sinks into one larger one. It can cool one GPU or the CPU well, but 2 or 3 would necessitate higher fan RPM.
That's why I'd like to see a whole system stress test, especially given the reported 450Watt limit on power.
Yeah and that didn't make any sense. They used "Sones" instead of dB like everybody else, and when converted I got over 100 dB which is getting into jet engine territory.
Possibly. My degrees are in physics and all I can say is we don't know unless we model or physically test it. The design is as efficient as it can be made, with the vertical design. The interesting thing to me is how warm the case gets, indicating that it is part of the thermal design. The black makes sense now too as that will give black body radiative effects too. Not much but it probably helps.
Yeah me too.
Which German review was that? Every review I've read so far basically says you'd have to put you ear to the fan to hear it and even then its not loud. The NY Times review says the machine is "virtually silent".
http://www.nytimes.com/2013/12/26/t...ac-pro-computer.html?hpw&rref=technology&_r=0
To lift from a post I just made over in the Ars Technica forums....
That NY Times review is the latest in a string of terrible reviews from people who have no idea how to actually test a machine like this. The "4K and 5K" footage they tested with is almost certainly R3D, which is a distinction that really needs to be made. The bottleneck there is not 4K image processing, but de-bayering and wavelet decompression, which are not presently GPU-accellerated (anywhere except beta versions of Red's own software) but will be soon. 4K ProRes or DPX, or even other raw formats (Cinema DNG, Sony Raw and Arri Raw, to name a few) are not nearly as CPU-intensive to decode, and this machine, by the numbers, should go through those like butter.
Of course those formats aren't natively supported in FCP X... but that merely points to another hilarious flaw in the methodology all of these publications are employing in testing "pro video" workflows. They all tend to naively dump camera-original footage into an NLE. While this sort of workflow is not entirely unheard of, the standard workflow for projects shot on cinema cameras is to edit with offline transcodes and then conform back to original footage in an app like Smoke or Resolve. The performance of the offline editing environment with the camera-original footage is entirely irrelevant. Avid Media Composer, supposedly the most 'professional' NLE in the industry, doesn't even bother to support resolutions over 1920x1080 precisely because the further up-market you go, the less likely it is anyone will want to 'finish' a project out of an NLE. None of the reviews I've seen so far seem to indicate their authors are even slightly aware of this. The analogy isn't exact, but what's going on here is roughly analogous to reviewers who have never heard of Photoshop reviewing a machine designed for professional image editing by poking at iPhoto a bit.
Apple should have included a copy of Resolve with every one of these review units pre-loaded with projects containing lots of graded footage in various formats/resolutions, along with some basic instructions for loading the app, running playback, and toggling grades on and off. I'm sure Blackmagic would have done all the heavy lifting here for the promotional value it would bring. It would have shown the machine doing impressive stuff with 4K footage in a real-world context.
That NY Times review is the latest in a string of terrible reviews from people who have no idea how to actually test a machine like this. The "4K and 5K" footage they tested with is almost certainly R3D, which is a distinction that really needs to be made. The bottleneck there is not 4K image processing, but de-bayering and wavelet decompression, which are not presently GPU-accellerated (anywhere except beta versions of Red's own software) but will be soon. 4K ProRes or DPX, or even other raw formats (Cinema DNG, Sony Raw and Arri Raw, to name a few) are not nearly as CPU-intensive to decode, and this machine, by the numbers, should go through those like butter.
Of course those formats aren't natively supported in FCP X... but that merely points to another hilarious flaw in the methodology all of these publications are employing in testing "pro video" workflows. They all tend to naively dump camera-original footage into an NLE. While this sort of workflow is not entirely unheard of, the standard workflow for projects shot on cinema cameras is to edit with offline transcodes and then conform back to original footage in an app like Smoke or Resolve. The performance of the offline editing environment with the camera-original footage is entirely irrelevant. Avid Media Composer, supposedly the most 'professional' NLE in the industry, doesn't even bother to support resolutions over 1920x1080 precisely because the further up-market you go, the less likely it is anyone will want to 'finish' a project out of an NLE. None of the reviews I've seen so far seem to indicate their authors are even slightly aware of this. The analogy isn't exact, but what's going on here is roughly analogous to reviewers who have never heard of Photoshop reviewing a machine designed for professional image editing by poking at iPhoto a bit.
Apple should have included a copy of Resolve with every one of these review units pre-loaded with projects containing lots of graded footage in various formats/resolutions, along with some basic instructions for loading the app, running playback, and toggling grades on and off. I'm sure Blackmagic would have done all the heavy lifting here for the promotional value it would bring. It would have shown the machine doing impressive stuff with 4K footage in a real-world context.
It's the exact same test over various devices, therefore it's an outstandingly accurate comparison.
I believe that was a typo. Apple specifically in fact says that users cannot replace anything but the SSD and RAM. There are several threads speculating on 1) how easy it will be to swap GPUs and 2) the availability of aftermarket GPUs.
Great review. You surely must realize why they didn't send you a unit after writing such a critical review
I have a feeling Apple's being very careful about who they send units to and I'm wondering if there is some material that is off-limits--certain benchmarks, Bootcamp capabilities, etc.
There are some very important questions that still need to be answered.
Yes they're clearly replaceable, for repair if nothing else. The question is; with what? I doubt Apple will offer after market upgrades the the CEO of OWC already soft panned the idea.
Stay glued to their OWC Blog Home. They got it on the 24th. Give them a little time. Plus the already gave memory upgrades so give them time to start testing.
Last edited:
It was from an audio interview, I don't have the source. The interviewer asked him about video card upgrades and whether he'd do one but he was soft on the idea. In CEO-speak that basically means they're not planning on it, but he doesn't want to say no in case there's a change for some unforeseen reason. He also mentioned the large amount of engineering that would go into such an effort. It's one thing to make a PCIe SSD or 2.5" SSD, another to make a custom graphics card.
I think the only way they'd do it was if the Mac Pro was so enormously popular the sales would support the development costs. Remember they only capture a small part of the overall market that wants to upgrade.
yeah.. gpu sales would have to be enormously popular for owc to make/sell them.. just the custom machinery required to build them is a million+ dollars.
i imagine if 3rd party gpus show up one day, apple and/or amd will be involved in the manufacturing aspects to some extent. (3rd parties buying 1/2 made blanks from apple (or smthng))
I will enjoy seeing your honest review of the next one.
When is your ship date? Are you getting the D700?
Not going to be music to your ears if the Mac Pro is just simply falling behind at that stage if hold the computation level for an extended period of time. The fan can keep up if overtime ramp the innards down to fit the limitation.
Constant noise levels and sinking computation performance probably isn't going to make most folks happy. [ I suspect hacks will eventually show up that raise the ceiling on the noise levels a bit. ]
----------
OWC doesn't need to buy the machines any more than Apple does (which they very probably didn't. As Apple doesn't actually make pretty much anything themselves. )
The problem isn't so much the manufacturing infrastructure price but far more that 'enormously poplar' would primarily be driven by "much cheaper" than by performance. The likely large disconnect between costs to make with a reasonable margin and what folks will pay.