Looks a bit blocky but I guess that's needed for cooling down the GPU especially the 64 if they plan to try and reduce throttling for those chips due to thermals and such. Also seems to be one 6-pin and one 8-pin connector feeding power to the GPU. And I guess they aren't going to call it the XFX Vega Thick. 😀 EDIT: And it seems gains might be minimal but I suppose that was to be expected, can't really push these GPU's that much higher it would appear and the stock config isn't bad though you can get them a bit quieter and perhaps there's room for some other improvements too.

I Don't like the positioning of the power connectors, but lets hope it can cool the 64 well.

HH:

"...Odd to see is the PCIe power connector placement in the middle..."

Barry J:

I Don't like the positioning of the power connectors, but lets hope it can cool the 64 well.

It's because card's PCB is so small, just take a closer look at the card from the backplate side. But indeed those power connectors in the middle of the card are ruining the aestetics of the card.

If you look at the 3rd pic, the back side of the card, you can see that there are large holes in the back that go all the way through to the cooling fins - basically this means it's just a really short PCB, which is why the power connectors are in the middle of the card, because that happens to be the end of the PCB (it's only just over half the length of the card by looks of it). The good thing about those holes in the back plate that go all the way through to the cooling fins - it will allow air to pass more quickly & efficiently through that part of the cooler (less restrictions), so that could help lower temperatures further. EDIT: didn't read your post before I posted mine (oh yeah, yours wasn't there, we posted at the same time), CrazY-Milojko, apologies for looking like I copied!

Robbo9999:

If you look at the 3rd pic, the back side of the card, you can see that there are large holes in the back that go all the way through to the cooling fins - basically this means it's just a really short PCB, which is why the power connectors are in the middle of the card, because that happens to be the end of the PCB (it's only just over half the length of the card by looks of it). The good thing about those holes in the back plate that go all the way through to the cooling fins - it will allow air to pass more quickly & efficiently through that part of the cooler (less restrictions), so that could help lower temperatures further. EDIT: didn't read your post before I posted mine (oh yeah, yours wasn't there, we posted at the same time), CrazY-Milojko, apologies for looking like I copied!

I'm gonna sue you for this 😉

Good to see some custom Vegas coming out., I was beginning to wonder. I sort of like the card, most only see the back-plate side any ways. With the power plugs indented, you can tuck the power cables making them less visible.

Vega 64 Double Whopper Edition...

Embra:

Good to see some custom Vegas coming out., I was beginning to wonder. I sort of like the card, most only see the back-plate side any ways. With the power plugs indented, you can tuck the power cables making them less visible.

I'm sure they were waiting until the miners calmed down, which right now seems to be the case. No point in spending money designing good custom coolers if the crappy leafblower reference cooler sells out within minutes.

On a slightly different topic, I do know multi-chip cards have some troubles on their own, but since the fab process is tuned for low power and clocks, wouldn't it make sense to put 2 lower clocked chips in the same card? It's not like AMD never did that :P Is the HBM memory controller getting in the way or something?

TieSKey:

On a slightly different topic, I do know multi-chip cards have some troubles on their own, but since the fab process is tuned for low power and clocks, wouldn't it make sense to put 2 lower clocked chips in the same card? It's not like AMD never did that 😛 Is the HBM memory controller getting in the way or something?

I figure the problem ultimately comes down to being too expensive to engineer while being highly undesirable products. From what I noticed, multi-GPU cards from both AMD and Nvidia have generally sold poorly. The thing is, DX12/Vulkan and features like HSA ought to make such products attractive again. But, I don't think we're going to see a GTX 1090 or Vega64x2 any time soon.

Unless mGPU is trully transparent for game devs (ZERO work to be done) and users multi GPU card will remain a thing from the past.

Well, it looks like we'll probably be going Multi-chip module rather than really Multi GPU, so any development towards that is diminishing, except on the Pure Compute side (Pro Users).