Saturday, June 6th 2020

NVIDIA GeForce RTX 3080 Pictured?

Here are what could be the very first pictures of a reference NVIDIA GeForce RTX 3080 "Ampere" graphics card revealing an unusual board design, which is the biggest departure in NVIDIA's design schemes since the original GeForce TITAN. It features a dual-fan aluminium fin-stack cooler, except that one of its fans is located on the obverse side, and the other on the reverse side of the card. The PCB of the card appears to extend only two-thirds the length of the card, ending in an inward cutout, beyond which there's only an extension of the cooling solution. The cooler shroud, rather than being a solid covering of the heatsink, is made of aluminium heatsink ridges. All in all, a very unusual design, which NVIDIA could implement on its top-tier SKUs, such as the RTX 3080, RTX 3080 Ti, and in a cosmetic form on lower SKUs. We get the feeling that "Cyberpunk 2077" has influenced this design.
Sources: ChipHell Forums, HXL (Twitter), VideoCardz
Add your own comment

225 Comments on NVIDIA GeForce RTX 3080 Pictured?

#151
Mats
Vya Domus
Anyone with 2 brain cells probably figured out what I said, what solid metal, what are you talking about ? The air is exhausted through the sides of the heatsinks because the PCB is in the way, so yes it's designed that way. For part of this design the air supposedly is meant to go directly perpendicularly through the heatsink, except that, like I pointed out, the that area looks closed off.
There is no PCB in the way in the front of the card IMO, it ends where the arrow shaped shroud begins. Picture link here.

Remember, the fan is on the rear side = PCB side, so there's no room for both a PCB AND a fan at the same spot. (Unless there are two PCB's, not very likely.)

The whole reasoning came from the idea that it would work like this. (Old GTX 960 as an example)

Air going straight through, just like I and at least two others tried to explain with excellent Paint skills. Then I realized that there's probably a vapor chamber in the way, so it's not possible.
This isn't about me being right or wrong, just a chain of thoughts.

Do you feel up to date now?
Posted on Reply
#152
Vya Domus
Mats
There is no PCB in the way in the front of the card IMO, it ends where the arrow shaped shroud begins
I feel like giving up but I'll try one more time. Usually in open air coolers the PCB is in the way, like in that 2080ti example you gave which I was clearly referencing.
Posted on Reply
#153
Mats
Vya Domus
Usually in open air coolers the PCB is in the way, like in that 2080ti example you gave which I was clearly referencing.
Yup, you're right there. I did some editing.. This time the fan is on the back, so it can't be any PCB there.
Posted on Reply
#154
Valantar
Mats
Yup, you're right there. I did some editing.. This time the fan is on the back, so it can't be any PCB there.
There are two fans, one on each side. The front/bottom fan, just like any traditional open air cooler, blows straight into the PCB (and thus whatever is on top of it, whether heatpipes or vapor chamber, with the air then turning 90° and going partially out the slot covers and partially back into the case through the exposed fins on the front. The back fan sits on the same plane as the pcb and base of the cooler, so its fins can't be connected to that vapor chamber - vapor chambers are flat, after all. Which is why I brought up the possibility of heatpipes connected to the vapor chamber earlier, though that is a terribly inelegant solution. Either that, the whole thing uses heatpipes (a regression from earlier designs) or the entire other half of the cooler cools only the VRM (which would be beyond stupid given that the GPU+VRAM likely outputs 200-225W of heat and the VRM 25-50.

Either way, this design seems problematic.
Posted on Reply
#155
Vya Domus
It's also very unnatural to have the hot air going both up and down, the heat naturally rises up.
Valantar
VRM 25-50.
That's way too much for some VRMs, they usually top out at about 15-20W for that kind of current.
Posted on Reply
#156
EarthDog
Valantar
Either way, this design seems problematic.
For what, exactly? If this is true/the design...IF... i'd imagine this was tested. I doubt it will be a "problem" though (surely it will work as designed and keep the gpu running where nvidia wants it to).
Posted on Reply
#157
Valantar
Vya Domus
It's also very unnatural to have the hot air going both up and down, the heat naturally rises up.



That's way too much for some VRMs, they usually top out at about 15-20W for that kind of current.
That's true, I was trying to be generous and add in some headroom for overclocking :)

As for convection, even the weakest fan can beat that easily, so the absolute direction of airflow only really matters if you're designing specifically to utilize convection in the first place. If you have a front intake and rear exhaust on your case, convection inside of the case is near meaningless.
EarthDog
For what, exactly? If this is true/the design...IF... i'd imagine this was tested. I doubt it will be a "problem" though (surely it will work as designed and keep the gpu running where nvidia wants it to).
Problematic in terms of maximising cooling potential per volume while maintaining compatibility. I obviously also assume this has been tested extensively if it is real at all, but nonetheless it stands out as an example of engineering for the sake of engineering (and partially for the sake of looking like engineering, of course) rather than aiming to make as optimal a design as possible. Design simplicity is not a quality Nvidia tends to favor though, what with the near infinite number of screws on their founders edition coolers, so I wouldn't put it past them to say "screw it, let's ignore complexity altogether and just make something we think is cool".
Posted on Reply
#158
Mats
Either way, the PCB looks crazy short for a high end card.
Posted on Reply
#159
RH92
Mats
I changed my mind. As someone pointed out, this cooler has probably a vapor chamber. If so, how can it possibly exhaust air in the underside of the card? (see pic in quote) Unlike heatpipe heatsinks, vapor chamber heatsinks doesn't let air straight through. Like someone else said, we can't see anything between those fins on the underside anyway, which is highly unlikely if that indeed was an exhaust. Maybe it's just fins on the underside of the vapor chamber, which makes this part (front, underside in case) passively cooled. (Yeah it can be a combination of VC and heatpipes, but still, we can't see anything in there anyway.)
Here's the vapor chamber of the 2080 TI FE. Pretty much no air, or light, comes through.

Yep i came to the same conclusion . I was the person who suggested the vapor chamber but i agree that on the '' VRM '' section if there is a vapor chamber there is no way for air to exhaust downwards .

Maybe as you said it's a mixed cooling solution , vapor chamber for core/memory and heat pipes ( which will allow the fan to blow downwards ) for the VRM .
Posted on Reply
#160
cucker tarlson
RH92
Maybe as you said it's a mixed cooling solution , vapor chamber for core/memory and heat pipes ( which will allow the fan to blow downwards ) for the VRM .
that'd seem reasonable.
who puts a vapor chamber on vrm ?
Posted on Reply
#161
alchemist83
birdie
This looks fake as hell and also the card is far too early - rumors have it pinned at Q4 2020 which is at the very least three months away. OEMs mustn't even have the chips at this point.

Edit: some interesting analysis hinting that the card/leak could be very much real:
Well make your mind up>? Real or fake? Cos looks real on first view. But you in the first case, so adamant its a fake. Explain more why GURU. Or dont. Sounds better to me actually. Flip flop. Back bone made of jelly.
Posted on Reply
#165
RH92
Honestly the more i look at this design the more it grows on me . It combines maximum heat sink area with minimum useless plastic shroud all this with 0 compromise to aesthetics !

Posted on Reply
#166
Chrispy_
Makes sense to me in almost every way, apart from the fact that I think the front fan should be a radial/blower outward blowing fan and not an axial down-blowing fan.

The front fan obviously exhausts out via the angled slots in the centre section. I'll assume this section is covered in a vapor chamber and with closed sides around the fan as well, the choice of an axial fan instead of a radial fan is an odd one - probably chosen for looks rather than functionaliy.

The rear fan obviously exhausts through the front of the card - you can just about see the heatpipes in the photo of the rear of the card.

Personally, I think the PCIe power cables will go along the top edge, right by the slot cover. I don't see them dumping PCIe power connectors on the end of the card like they did with the RTX 20x0FE because that was a mess even with a simple single-sided cooler and regular rectangular PCB.
Posted on Reply
#169
dyonoctis
Vya Domus
I like how they made it such that the fan can be seen through the fins even though in the pictures that isn't the case at all :).
You could argue that since the fan is black, the fins are dark, the surface where the gpu was put down is black and the low quality of the photos, as well, there's no much way for the light reflecting on those dark surface to come through.
Posted on Reply
#170
mechtech
Hmmm

After exchange rate to Cnd$ and 13% tax, I think I'd rather have a whole set of new appliances instead.
Posted on Reply
#171
cucker tarlson
mechtech
Hmmm

After exchange rate to Cnd$ and 13% tax, I think I'd rather have a whole set of new appliances instead.
I feel you.
was thinking about cpu upgrade.3600/10400f + mobo + ddr4
for what it would cost me I got a new case (with extra fans),new heaphones with a dac and a new printer.

ampere/rdna2 is a must buy tho.I bought a card every gen cause I just like new stuff.started with 7870-290-980-980ti-1080-1080Ti-2070S.
Posted on Reply
#172
Turmania
I would not be surprised if AMD, copies this design on their next gen RDNA3 once again...
Posted on Reply
#173
cucker tarlson
Turmania
I would not be surprised if AMD, copies this design on their next gen RDNA3 once again...
if it's not patented and locked by nvidia then they're free to do so.
Posted on Reply
#174
chodaboy19
I really want to see this designed released! It's certainly piqued everyone's curiosity. :)
Posted on Reply
#175
EarthDog
Valantar
That's true, I was trying to be generous and add in some headroom for overclocking :)

As for convection, even the weakest fan can beat that easily, so the absolute direction of airflow only really matters if you're designing specifically to utilize convection in the first place. If you have a front intake and rear exhaust on your case, convection inside of the case is near meaningless.

Problematic in terms of maximising cooling potential per volume while maintaining compatibility. I obviously also assume this has been tested extensively if it is real at all, but nonetheless it stands out as an example of engineering for the sake of engineering (and partially for the sake of looking like engineering, of course) rather than aiming to make as optimal a design as possible. Design simplicity is not a quality Nvidia tends to favor though, what with the near infinite number of screws on their founders edition coolers, so I wouldn't put it past them to say "screw it, let's ignore complexity altogether and just make something we think is cool".
If only we knew the results.... if this was real. ;)
Posted on Reply
Add your own comment