EVGA GeForce RTX 3080 FTW ULTRA features 1.8 GHz boost clock (highest so far)

https://videocardz.com/newz/evga-geforce-rtx-3080-ftw-ultra-features-1-8-ghz-boost-clock

For reference, the same model for the 2080 had a factory boost clock of 1860 MHz

Not necessarily, since, for instance, Fiji was 596 mm² versus 352 mm² for the Tahiti chip with half as many CUs and an earlier iteration of the GCN architecture. There is a lot of machinery inside the chip that won't double with twice as many CUs. Looking at the new consoles, I'd expect less than 500 mm² unless Big Navi includes some surprises that the consoles don't have (but I think the RT solution will be similar and use some form of minifloats).

True, true, I forgot that not everything doubles. Things like the I/O or, if the leaked pic is real, the 256-bit memory bus will ramain the same.

I have decided to wait for RDNA 2... Assuming my trigger finger doesn't tap on the buy button prematurely. One of those, they probably won't have an answer but if they do... Or if Nvidia decides to maybe lower the price of that 3090... Mmm

Hearing a lot of rumors/leaks lately about 3060 and 70 ti's due soon makes me think that 80 ti might not be far off either. Think I will (at least try to) wait til early next year to make my purchase decision.

Possible Big Navi leak:

256-bit
128 ROP
WTF

Most likely Navi 22 but that is one hell of a ROP count. Or it's Navi 21 but the Engineering Sample might be old which could result in a low speed bus. Some more discussions if yall want to have a look:

AMD ES with 256 bit interface:@CapFrameX noted that *maybe* big caches can be helpful. @Locuza answered that this will not close the gap. Others believe, that we seeing a Navi 22 board here. @scottiefour4 noted that the GPU on that ES is 320-360mm² big, so probably not Navi 21. https://t.co/bidavi5ueC

— 3DCenter.org (@3DCenter_org) September 11, 2020

If the die size(505mm2) and bandwidth(256bits) of Navi21 is true, someone said Navi21 will have a huge cache to explain the number. How do you explain it? There is a lot of discussion @3DCenter_org .

— kopite7kimi (@kopite7kimi) September 11, 2020

Big Navi might just be a strict doubling of Navi 10 in terms of functional units... And if that's the case, it would very well explain the size of the chip + Ray Tracing cores.

Shouldn't forget AMD is also just cooking qith water. Just no way I'm gonna skip RTX and DLSS for a few fps more from team Red. I doubt anyone will be disappointed with a 3080.

It's less about being disappointed with 3080, specially with those performance numbers, and more so about getting the right card. What I mean by that is, if RDNA 2 turns out to be good and Nvidia releases the rumoured 20GB version... I am 100% gonna regret buying the 10GB 3080.

Like if they release the 20GB model during their refresh next year, fair enough. But the last thing I want is to buy it now and in a couple months, Nvidia releases the 20GB version. Of course, the ultimate solution would be the 3090 but that could also get a price cut if RDNA 2 is good enough.

I'm quite impatient to see what AMD has to offer - if they're targeting 3070 market (and bellow), like they used to do in some of their past battles with nVidia when they were at each others throat, and have price/performance win (which is, I believe, quite possible actually) it will be one helluva fight.