NVIDIA Ampere Cooling Solution Heatsink Pictured, Rumors of Airflow Magic Quashed

[Valantar]

Its not a joke, its to help illustrated how this design functions, how its pretty much exactly that sapphire card but with one of the fans instead on the other side.

... You didn't actually read the last few posts in the thread, did you? I was referring to this:

Sapphire Pulse RTX 3080 TI lol

I very sincerely doubt that was meant as an illustration of the design principles of this hsf. I might of course be wrong, but I do believe my reading of that post is a bit more likely than yours, @ZoneDymo.

Also, if you're asking for pedantry, you'll get it: that isn't the only difference; the fan on the opposite side is also flipped, reversing airflow. As such this is more of an evolution of the sapphire-style design than a new take on the same style (no idea if it's an evolution to the better of course...). The Sapphire designs also use horizontal rather than vertical fins, which makes a lot more sense for the part of the heatsink that is obstructed by the PCB.

 

[ppn]

Well Im saving for anything on 6nm with 66Mtr/mm2 density, GDDR7 instead of this GDDR6x.
400 watts power means this is still on 10nm and a total joke. or something like the 7nm used in NAVI 40Mtr/mm2. the 7nm used for Ampere is something else 66Mtr/mm2.

 

[ARF]

Well Im saving for anything on 6nm with 66Mtr/mm2 density, GDDR7 instead of this GDDR6x.

GDDR7 doesn't exist. GDDR6X doesn't exist either.
So, it is GDDR6 or HBM2/2e/3..

 

[1d10t]

That's debatable - this design does maximize fin area for a two-slot card in terms of having fins "all the way out" rather than having a shroud around them. What is questionable about it on the other hand is (as @RH92's illustration above shows) the middle section of the fin stack having next to no airflow through it. There will be some, as the fins aren't nearly as dense as those directly surrounding the fans, but most of that air will take the quickest possible way out of the fin stack, leaving most of the middle/diagonal fin stacks with near zero flow. Nvm, didn't spot that the fins were entirely closed off. Passive VRM cooling fin stack?

Still, as an SFF enthusiast I have to applaud Nvidia for moving to smaller reference PCBs. Should allow for some pretty nice AIB partner ITX cards, even if this cooler design will be terrible for sandwich style cases.

Edit: see above.

That exactly what I meant.Those fin array arrange diagonally, with their end being blocked. It would be useful if they stick to regular fan instead of Axial.

Two regular 92mm fan will provide best cooling and also keep noise level to minimum.But hey, what do I know

 

[Midland Dog]

for verticle mount this would go off, just need red rgb

 

[Valantar]

for verticle mount this would go off, just need red rgb
View attachment 158568

This would be quite terrible for a vertical mount - the inner/secondary fan would blow all its hot air directly into the glass, leading a lot of it to be sucked back into the primary fan. Vertical mounts are already terrible for cooling, but this design would likely be even worse than regular open-air designs - at least those have relatively consistent flow directions.

That exactly what I meant.Those fin array arrange diagonally, with their end being blocked. It would be useful if they stick to regular fan instead of Axial.

Two regular 92mm fan will provide best cooling and also keep noise level to minimum.But hey, what do I know

"Regular" fans are axial fans, axial just means it has an airflow direction parallel to its axis of rotation. This is in opposition to radial or "blower" fans that expel air outward radially from the axis of rotation. What they seem to be using here is just a sort of enclosed axial fan, likely as a move to improve static pressure/avoid air escaping out the sides and forcing it through the fin stack instead. Overall this design seems like a bunch of weird compromises put together, though I guess they might have stumbled onto something brilliant?
I'm not an engineer, so I frankly have no idea, but it sure seems overcomplicated and likely to not be very good.

 

[Casecutter]

Is it just me or do those Heatsink look to be 3D printed? New types of laser powder bed fusion (LPBF) materials make it possible to produce components with a wider range of complex geometries for optimal heatsink efficiency, that are not possible with traditional manufacturing processes. However I don't such a process being cost effective for even $1000 consumer graphics cards, but they could prove me wrong.

Edit: Looking at this more I believe the tubes could be printed and then after filled and swedged under vacuum r... that to be would be tricky cause if you make a mistake you loose the whole printed component. Or, on the side we don't see they have a sealed plate that connects everything in one vapor chamber.
More i consider looking at the cost the idea from a total manufacturing point this might have advantages and offsetting costs in assembly. While upfront you design, and optimize in the software, send it to print and have the working design to go actually evaluate in hours. IDK I'm starting to see the value in this ability.

Comparing Optimization Methods for a Heat Sink Design for 3D Printing

To find the best possible heat sink design, researchers from Fraunhofer IAPT compared different optimization and manufacturing methods. Get the full story.

www.comsol.com

 

[Valantar]

Is it just me or do those Heatsink look to be 3D printed? New types of laser powder bed fusion (LPBF) materials make it possible to produce components with a wider range of complex geometries for optimal heatsink efficiency, that are not possible with traditional manufacturing processes. However I don't such a process being cost effective for even $1000 consumer graphics cards, but they could prove me wrong.

Edit: Looking at this more I believe the tubes could be printed and then after filled and swedged under vacuum r... that to be would be tricky cause if you make a mistake you loose the whole printed component. Or, on the side we don't see they have a sealed plate that connects everything in one vapor chamber.
More i consider looking at the cost the idea from a total manufacturing point this might have advantages and offsetting costs in assembly. While upfront you design, and optimize in the software, send it to print and have the working design to go actually evaluate in hours. IDK I'm starting to see the value in this ability.

Comparing Optimization Methods for a Heat Sink Design for 3D Printing

To find the best possible heat sink design, researchers from Fraunhofer IAPT compared different optimization and manufacturing methods. Get the full story.

www.comsol.com

There is nothing in this design indicating that it is manufactured in any exotic way. Both end heatsinks are standard fare with stamped fins packed (relatively densely) together, crimped in various places, and pressed onto their respective heatpipes or cold plates. The middle portions are extruded aluminium profiles cut to length, with any necessary cutouts CNC'd into them afterwards. All fins are clearly far too smooth, straight and even to be printed - there would inevitably be printing artifacts of some sort if it was printed. The article you linked seemed to be about designs optimizing surface area far beyond conventional fin stacks, while this is clearly nothing but a (couple of) conventional fin stack(s).

 

[Casecutter]

There is nothing in this design indicating that it is manufactured in any exotic way. Both end heatsinks are standard fare with stamped fins packed (relatively densely) together, crimped in various places, and pressed onto their respective heatpipes or cold plates. The middle portions are extruded aluminium profiles cut to length, with any necessary cutouts CNC'd into them afterwards. All fins are clearly far too smooth, straight and even to be printed - there would inevitably be printing artifacts of some sort if it was printed. The article you linked seemed to be about designs optimizing surface area far beyond conventional fin stacks, while this is clearly nothing but a (couple of) conventional fin stack(s).

I would say look at the blow-up that RH92 gave in his post #8 that does not appear all that smooth. You should look around on the internet and see what's happening in the field... Direct Metal Laser Sintering (DMLS) (Printed metal) and/or bed fusion materials have come a long way.

Are those 8mm heatpipes?
Dang this thing must be HOT

They probably look like that as they are printed, not traditional copper heat pipes

 

[Valantar]

I would say look at the blow-up that RH92 gave in his post #8 that does not appear all that smooth. You should look around on the internet and see what's happening in the field... Direct Metal Laser Sintering (DMLS) (Printed metal) and/or bed fusion materials have come a long way.

They probably look like that as they are printed, not traditional copper heat pipes

Sorry, but that just looks like a poor phone camera to me, combined with the typical minor dents and bumps seen on any heatsink. I know full well there's lots of crazy stuff done with metal 3D printing, but this would be an incredibly stupid use case for it. Not to mention that it would drive the price of this heatsink far beyond the rumored $150 level. This is a mass produced part for a mass market GPU; 3D printing is not suited to mass production, it is suited to small runs of high complexity parts. While this is a somewhat complex heatsink, it is entirely possible to manufacture with cheap and widely available techniques and equipment. If Nvidia was crazy enough to print a heatsink like this, why on earth would they go with a straight-finned design? Doesn't that defeat the only purpose of using 3D printing - the ability to make complex, non-flat metal structures? There is literally nothing in the structure of this indicating that it needs exotic manufacturing methods. I agree that it would be cool if it was true, but there is no way it is. Period. Even for a $5000 GPU that would be highly unlikely simply due to necessary production volumes.

 

[ARF]

Who knows what this GPU is, might be RTX 3090 or new Titan.

Exclusive first look at Nvidia's Ampere Gaming performance - HardwareLeaks.com

116,984 total views, 7 views today Next-gen Nvidia Ampere gaming GPUs are one of the most anticipated pieces of hardware in the PC and gaming scene this year. Over the course of the last few months, a number of leaks surfaced detailing various aspects of the much anticipated gaming GPUs...

hardwareleaks.com