[POLL] What are the Odds to have PCIe Slots in the new Mac Pro

[barmann]

Given that the iMac Pro runs silently 99.6% of the time, how can you say that they’re at their maximum heat dissipation already?

You seem to suggest the iMP runs silently 99.6% of the time, while using the full performance of its components, without any throttling , and for an extended period of time .

Is that indeed a fact ?

 

[-hh]

Given that the iMac Pro runs silently 99.6% of the time, how can you say that they’re at their maximum heat dissipation already?

Because it is known that they de-rated components (to make them run cooler).

All they would have to do is set the fan to normal i7 iMac noise levels in order to shift more heat than they are currently.

Could...but didn't (and de-rated components instead).

Plus running a fan "faster" has a cost with diminishing returns: because Kinetic Energy (KE) has a "velocity squared" term, a doubling of the air mass's velocity requires ~4x the power. Similarly, one also has to walk through the entire design to determine where/when an increased flow rate results in less net heat transfer, such as by invoking a change from laminar to turbulent which can actually _decrease_ the net flow (despite higher fan power input) such as through choking.

(the engineering of thermal design management isn't as trivial as many people often try to assume)

That also ignores the fact that they do have full control over the iMac’s case design, the back could be made deeper, and so there’s nothing preventing them from modifying it with bigger fans or even a liquid cooling system.

True, but the point here was to prevent the cost of retooling by having a "good" design that is able to accommodate internal updates (changes) without having to throw it away and start over.

When you’e not constrained to fitting everything into a tube of a specific size, all sorts of incremental changes are possible.

Precisely, but this observation isn't merely for a tube (trash can), but all objects: the ability to accommodate future incremental changes depends greatly on how good the base design was done in planning ahead for them, and avoiding pitfalls of introducing unnecessary (= artificial) engineering design constraints.

Overall, think of this like choosing a suitcase to go on vacation with: do you really want the one that absolutely won't fit even an ounce more than what you're leaving home with?

-hh

 

[AidenShaw]

Overall, think of this like choosing a suitcase to go on vacation with: do you really want the one that absolutely won't fit even an ounce more than what you're leaving home with?

Only if I'm leaving home with a lightweight duffle bag in the suitcase to deal with "vacation expansion".

I guess that Apple thought that T-Bolt would be a duffle bag.

 

[ThatSandWyrm]

You seem to suggest the iMP runs silently 99.6% of the time, while using the full performance of its components, without any throttling , and for an extended period of time .

Is that indeed a fact ?

You seem to suggest that every component in the iMac Pro will/must be used at full power at all times.

I CAN run rendering computations where the CPU and GPU are both maxed out to the point where the fans become quite audible. But it’s much more computationally efficient to use just the CPU or the GPU, depending on the nature of the computation. In which case you never hear the fans. Apple throttles to eliminate noise, not because they necessarily have to.

Yes, I would like finer control on the fan noise/throttling tradeoff. But that’s immaterial to the point of the iMP having further heat dissipation potential that’s not being used. The fans may have a point at which spinning them faster has a zero or negative effect, but I think that we can both agree that point is at least somewhere in the audible range.

There is also a difference between being able to modify the backside of an iMac (that nobody spends much time looking at), and having painted yourself into a corner by restricting your form factor to a specific size of cylinder that your customers are constantly looking at. Change the latter, and you lose your product’s visual appeal. Change the former and if anyone outside of a tech blog notices at all, it will be the couple of pounds of extra weight.

 

[barmann]

You seem to suggest that every component in the iMac Pro will/must be used at full power at all times.

Let me rephrase .

Do you seriously believe that an iMac is not significantly compromized by it's design ?

Is there any technology in this world that makes a flat rectangular box, housing a display and all the computer components, anything but the least efficient desktop case design known to man ?

I think iMacs are absolutely brilliant for what they are, don't get me wrong, but you have to face some facts here .

 

[ThatSandWyrm]

Let me rephrase .

Do you seriously believe that an iMac is not significantly compromized by it's design ?

Is there any technology in this world that makes a flat rectangular box, housing a display and all the computer components, anything but the least efficient desktop case design known to man ?

I think iMacs are absolutely brilliant for what they are, don't get me wrong, but you have to face some facts here .

Is its theoretical maximum performance compromised? Yes, almost certainly.

Is its *practical* maximum performance significantly compromised? Yes, due to the no-noise throttling. But without that throttling, I don't think that you'd see more than a few percent difference in practical performance between a standard fan-filled PC box and the current iteration of the iMac Pro. PC boxes aren't nearly as well designed from a thermal management perspective as you might think.

As for "efficiency", that depends on what you're measuring. There are a lot of ways in which a business could find an all-in-one design to be more "efficient" from a deployment and maintenance standpoint than a box + monitor setup. Having a computer be all-but-silent in its operation can also do a lot to increase the "efficiency" of its users. Especially for audio/video work.

Having a separate box is obviously more "efficient" for raw performance if you're aiming to put in a DIY liquid cooling solution, for instance. But the number of systems that ship with liquid cooling is almost non-existent.

 

[Joe The Dragon]

There have been Xeon SKUs with integrated graphics before. I'm certain Intel could provide one with sufficient punch to drive the upcoming pro display. Don't hear this discussed much. Or, since there will be PCIe slots, there's room for a GPU on a mezzanine card somewhere in there.

The more I think about it, that's a very Apple solution. Baseline GPU options are Vega 56 and 64 (or equivalents) on a mezzanine card. Apple can customize it to do whatever trickery to handle video over TB3, enthusiasts can upgrade as new cards hit the market (and if not, who cares we have PCIe again).

Hmmm... coupled with my belief that they're really into this thermal core language... imagine the same thermal core. Your configuration options are (per side)

1. CPU
2. GPU
3. 2nd GPU or CPU

Only DIY would attempt upgrade of these core parts.

The rest of the chassis could house 4 full-sized Nvidia GPUs and still be reasonably small. So a mini Z? They must have something more in mind. Some high speed fabric to tie the core components together could be something. AMD was talking about bringing an open source NVLink competitor to market, but it's been a while...

lot's of servers have IPMI/low end video chips on board (pci or pci-e X1 bus).