Agree but surely the only interest here is 4k benchmarks
The unigine superposition optimized benchmark at 4k , I got 11605 default clock
and 13059 overclocked = increase of 12.5% o/c on the 2080ti
from above video the o/c on the 3080 was about 6%
so the % improvement that the 3080 has over the 2080ti should be reduced by 7% if both cards are overclocked
edit and my overclock setup is very very basic some guys are getting 30-35% more by overclocking 2080ti’s , which matches the 3080 with current overclock
Maybe spend a few quid water cooling / overclocking the 2080ti to get equivalent fps
I’m not going to watch any more speculation/test version videos right now. But if the guy is comparing 2080ti OC with a 3080… I wouldn’t say it’s bad. The 3080 is so much cheaper and a better comparison would be 3090 (which ofc noone has yet).
So I guess a 3080 is not an upgrade if you’re sitting on a monster 2080ti. If you however have any other card, then it’s a hell of an upgrade… and “cheap” too.
In any case: all the 3000-series will be sold out for months, so we all we can do is sit and wait for some real comparisons and benchmarks
I have not bothered to check out much on the 3080 vs 2080ti, but by the looks of it, the 3080 does run cooled than a 2080ti over clocked, so there could be some improvements to be made on the 3080.
I skipped the upgrade to 2080ti, as My 1080ti water cooled does every thing I need it to do.
But I am intrigued to seen if there is going to be a 3080Ti very shortly, may be worth waiting for this one.
I’ve come to the conclusion that no computer part should be worth this much anxiety.
If you can’t secure one today, you’ll get one in the next few months - the Turing launch was no different.
Sounds like the AIB boards might even have a power limit advantage, so we just need to give it some time.
If you’re out of a GPU right now because you sold your 2080 Ti… well… sh*t happens.
there is still too few tests to conclude anything but still, it’s going from 191% to 203%, averaging 196%
OC 2080TI still wins over this extremely small 3080 sample.
worth to sell a 2080ti for 400-500€ to buy a 700-800 out of stock 3080 that is maybe as performant as OC2080ti or just a bit more out of the box ?
the question is not trivial.
for 1080 owners, it’s a no brainer, obviously unless you are in a budget, then 2080ti are available and not too expensive anymore on the second hand market.
Edit: there is a 4th test, still on the same range.
Edit 2 : 6 samples now, still averaging 195%, max a 203% VS 204% for OC2080ti
Ok, have preordered the EVGA 3080 XC3 Black for 699€. This card doesn’t have a backplate, thus 2 slots and fans all in the outward direction.
Read that the 20 GB version, if it comes out this year, can only work with RAM modules on both sides of the PCB because 2 GB modules will only arrive in 2021. So it is unlikely that a 2 slot version of that would be possible. And I need a 2 slot card for my SFF sandwitch case that bows air outwards, not into the power supply, like the FE cards.
Status is “Delivery time unknown”. Let’s see how this goes…
These forums are full of recent 2080ti purchasers trying to justify their purchase. The fact is the 3080 is amazing. People are comparing it to a 2080 ti but they should compare it to a 2080.
Folks, 1980MHz is an awful result! If that is all we get out of the RTX 3090, it is likely we will see only a 40% improvement in real performance (taking the difference between core/transistor count estimates).
I think it is now obvious why NVIDIA has been up to marketing shenanigans.
NVIDIA wasted their transistors on ‘new features’, specifically RTX and DLSS, leaving VR performance crippled for another generation.
