Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
They don't need technical help from Apple. The layer they interface with doesn't care what kind of hardware is on the other end.
If you were responsible for writing the drivers now shipping, I might believe you.

Having been there, you find that sometimes in drivers subtle things make a big difference. Maxwell had 40-bit addressing. Pascal has more. Since Apple doesn't support any system that needs more than 36-bit addressing, perhaps Pascal's additional addressing width ran into a problem with Apple's frameworks.

Or perhaps Pascal's more recent OpenGL support has issues with Apple's neglected version?

I think that you're promoting a conspiracy theory, not making a statement based on the engineering involved.
 
  • Like
Reactions: tuxon86
They could also just go down to one beefy GPU, but that may be hard to fit into the thermal core type design, where heat is relatively evenly contributed from the 3 sides of the triangle.
that's not an issue, and actually the nMP' s thermal core uses to operate unbalanced, since its unlikely both gpu to work in tandem on a same problem, the nMP gpus one is for rendering the 2nd for compute, so unless you do a concurrent render plus compute operation you'll have the thermal core unbalanced, a I said this is not an issue to put all the heath at one side and leave a core unpopulated (or populated bu the SSDs).
 
that's not an issue, and actually the nMP' s thermal core uses to operate unbalanced, since its unlikely both gpu to work in tandem on a same problem, the nMP gpus one is for rendering the 2nd for compute, so unless you do a concurrent render plus compute operation you'll have the thermal core unbalanced, a I said this is not an issue to put all the heath at one side and leave a core unpopulated (or populated bu the SSDs).

First, both GPUs can be used for compute. I don't know why that myth seems to still circulate around here. For example barefeats shows a benchmark with 1 and 2 GPUs enabled on the mac pro.

Second, in the current mac pro configuration, you have 3 chips, each contributing roughly 125 W from each side of the thermal core triangle. That works because in the highest load case the heat sources are spread out and can dissipate evenly. In cases where only 1 GPU is active then there is excess thermal dissipation capacity and the fan can stay quiet.

Now if you put a 250+ W GPU on one of those sides, its no longer balanced and its much more difficult to effectively cool, given that all that heat is coming from one side and the heat flux is much higher. Its not an impossible problem, maybe they could use heat pipes or something to spread out the heat, but its certainly not what the current chassis was designed for.
 
Overclocking coolers have mastered much more challenging thermal envelopes for years.
There is simply no excuse for a lack of cooling in small form factors anymore. Look at the countless examples in the HTPC and SFF market. Heatpipes, LCS, you name it.
 
Shouldn't that face be a bit red instead of orange to reflect the level of embarrassment that Timmy should be feeling about now?!:confused:

Embarrassed2.png
 
Overclocking coolers have mastered much more challenging thermal envelopes for years.
There is simply no excuse for a lack of cooling in small form factors anymore. Look at the countless examples in the HTPC and SFF market. Heatpipes, LCS, you name it.

Show me another computer that dissipates 450 W of heat in roughly a 7 L system. You won't find it. As someone who has built a SFF (if you can call a 20 L case SFF), its certainly not a solved problem. CPU loads are silent due to the giant heatsink that is strapped on the top of it. GPU loads are not silent and are certainly audible and this is with a GTX 1080 with a non-reference cooler that only puts out something like 180 W.
 
  • Like
Reactions: koyoot
I can deffinitely confirm what Stacc have wrote, because my company builds power efficient, powerful, small form factor computers. And the best way to dissipate large amounts of heat from small case is liquid cooling, possibly in custom loop. Its all about density of thermall heat output that makes difference here.

I know that concept of efficiency and heat output may seem illogical at first glance, but think about for example a small form factor case called SG13 and putting into it a CPU, mobo, 180-250w GPU, and the rest of stuff, that your customer may want.
 
I can deffinitely confirm what Stacc have wrote, because my company builds power efficient, powerful, small form factor computers. And the best way to dissipate large amounts of heat from small case is liquid cooling, possibly in custom loop. Its all about density of thermall heat output that makes difference here.

I know that concept of efficiency and heat output may seem illogical at first glance, but think about for example a small form factor case called SG13 and putting into it a CPU, mobo, 180-250w GPU, and the rest of stuff, that your customer may want.

The biggest issue to cooling a PC is you are limited in how you design the cooling system because you have to use standard parts. For example the ATX motherboard was designed decades ago when a CPU used < 50 W and GPUs didn't exist. Now in modern times CPUs can use 140 W and GPUs can use up to 300 W. Take a Nvidia Titan video card, does it make sense to cool that using a crappy blower? No, but PC makers have no choice.

For all the criticisms of the trash can form factor, it has superior cooling design due to the fact that it ditched standard components.
 
  • Like
Reactions: koyoot
The biggest issue to cooling a PC is you are limited in how you design the cooling system because you have to use standard parts. For example the ATX motherboard was designed decades ago when a CPU used < 50 W and GPUs didn't exist. Now in modern times CPUs can use 140 W and GPUs can use up to 300 W. Take a Nvidia Titan video card, does it make sense to cool that using a crappy blower? No, but PC makers have no choice.

For all the criticisms of the trash can form factor, it has superior cooling design due to the fact that it ditched standard components.
You nailed it perfectly here.
 
There are a few good examples:

https://www.zotac.com/us/product/mini_pcs/magnus-en1080-10-year-anniversary-edition

http://www.anandtech.com/show/11041/horizontal-cpu-coolers-3way-roundup

http://www.anandtech.com/show/11052...x-using-microstx-motherboard-with-mxm-support

Mostly standard components that leave headroom for overclocking and consequently even higher thermal loads. They may not be smaller or quieter than the nMP, then again not a single professional ever wanted a smaller Mac Pro. Apple's restrictions are 120% home made.

Last but not least there's the Razer Blade Pro that manages to stick a four core cpu and 1080 in a case no thicker than a 2012 mbp. Meanwhile Apple is still struggling fo provide adequate cooling for systems less than half as powerful. Again, entirely by choice and not necessity.
 
Last edited:
There are a few good examples:

https://www.zotac.com/us/product/mini_pcs/magnus-en1080-10-year-anniversary-edition

http://www.anandtech.com/show/11041/horizontal-cpu-coolers-3way-roundup

http://www.anandtech.com/show/11052...x-using-microstx-motherboard-with-mxm-support

Mostly standard components that leave headroom for overclocking and consequently even higher thermal loads. They may not be smaller or quiter than the nMP, then again not a single professional ever wanted a smaller Mac Pro. Apple's restrictions are 120% home made.

Last but not least there's the Razer Blade Pro that manages to stick a four core cpu and 1080 in a case no thicker than a 2012 mbp. Meanwhile Apple is still struggling fo provide adequate cooling for systems less than half as powerful. Again, entirely by choice and not necessity.
Are you sure about that? Zotac ZBOX, with 65W CPU and GTX 1060. https://www.computerbase.de/2016-10/zotac-zbox-magnus-en1060-test/3/

Look how this performance compares to desktop GPUs. GTX 1060 in this small form factor, because of thermal throttling is only slightly faster than GTX 960 and R9 380.

You cannot have power efficiency and performance at the same time, and small form factor without compromising something. And the first thing that is compromised is cooling efficiency.

Mac Pro was a great achievement from engineering perspective, because it did not throttled other way than by power Virus, that consumed way too much power.

Also, the Razer Blade that packs GTX 1080 is Razer Blade Pro. It has 17.3 inches, and weighs 3.7 KG's. Razer Blade is 14 inch computer with GTX 1060, and 180W PSU. So its not efficient, compared to MBP.
 
Last edited:
What? You’re serious? Some of the components are down clocked, ie throttled.
Was it throttling from the point of design? Was it throttling from 850 MHz, lower in thermally constrained situation, eg. load?
 
Was it throttling from the point of design? Was it throttling from 850 MHz, lower in thermally constrained situation, eg. load?
It was in effect pre-throttled. Like putting a less powerful engine in a car so that the traction control isn’t triggered due to a lack of torque at the wheels.
 
It was in effect pre-throttled. Like putting a less powerful engine in a car so that the traction control isn’t triggered due to a lack of torque at the wheels.
Or that Brabus took V8 block from Mercedes, tuned it down from 6.2 litres to 5.0, to fit specific thermal envelope.

You can call it downclocking. I will call it designing. Who will be right?
 
Or that Brabus took V8 block from Mercedes, tuned it down from 6.2 litres to 5.0, to fit specific thermal envelope.

You can call it downclocking. I will call it designing. Who will be right?

You change v8 block that worked perfectly fine, downclock it to fit some 'design' thermal requirements. How is that not throttling ? Why throttle ? Why let form dictate function ?
 
Or that Brabus took V8 block from Mercedes, tuned it down from 6.2 litres to 5.0, to fit specific thermal envelope.

You can call it downclocking. I will call it designing. Who will be right?
Designing,? Hilarious. It is STILL downclocking.
 
  • Like
Reactions: tuxon86
You change v8 block that worked perfectly fine, downclock it to fit some 'design' thermal requirements. How is that not throttling ? Why throttle ? Why let form dictate function ?
Perfect - since engines have "throttles" and one can easily "detune" an engine by limiting how wide the throttle can open.

Or that Brabus took V8 block from Mercedes, tuned it down from 6.2 litres to 5.0, to fit specific thermal envelope.
One of the worst car analogies ever to appear in MacRumours. The displacement of an engine is not a "tunable" parameter. It's fixed by the bore and stroke and number of cylinders. If you want a different displacement, you need to build a different engine.
 
  • Like
Reactions: SimeoneSergio
Perfect - since engines have "throttles" and one can easily "detune" an engine by limiting how wide the throttle can open.


One of the worst car analogies ever to appear in MacRumours. The displacement of an engine is not a "tunable" parameter. It's fixed by the bore and stroke and number of cylinders. If you want a different displacement, you need to build a different engine.

No you can bore it and add more stroke or destroke it.
 
Having been there, you find that sometimes in drivers subtle things make a big difference. Maxwell had 40-bit addressing. Pascal has more. Since Apple doesn't support any system that needs more than 36-bit addressing, perhaps Pascal's additional addressing width ran into a problem with Apple's frameworks.

That's not really a problem? As you've pointed out, Maxwell works just fine. That's because there isn't a requirement you need to use the full 40 bits, or whatever Pascal uses.

Much like how when you run a 32 bit binary on a 64 bit version you're ok because you don't have to use the full 64 bits.

I think that you're promoting a conspiracy theory, not making a statement based on the engineering involved.

I'm just pointing out that the OS X graphics ABI doesn't care what kind of hardware is on the other end. It's a private ABI, but you can still poke at it.

The bit depth thing is easily dismissed, and that's the only reason I've seen as to why Nvidia couldn't port the drivers if they wanted. An easier explanation is it's just not profitable for them to do it on their own.

I mean, do you think Apple actually made modifications to OS X to support Maxwell? Puh-lease.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.