• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

nVidia 3080 reviews thread

BFG10K

Lifer
Written:


Video:

 
Last edited:
They say Nvidia is following Apple and trying to make it a premium, luxury product. It does look really nice, but I would expect the AIB cards with 3 fans (some of which honestly look much uglier) to run cooler and have higher boost clocks.
 
Anything that is not actual numbers is just noise to me. While I will obsess and waste a lot of time about even speculation about performance.. unboxing is still completely uninteresting. But each to their own I guess..
 
What's the point of these previews? "Ooh look, it's so premium feeling", great, I'm really going to care about that when it's in my ATX case. 🙄
People are excited so they watch whatever is avaiable. The casters know they'll get views as long as they are talking about RTX 30 series. So they are getting some extra views before they are allowed to release the benchmarks.
 
I have a small stockpile of old videocards in my closet, I think going all the way back to my old Trident 2d and Permedia Labs something or another.

I've repeatedly considered gerry-rigging some shadow box displays for the cards to show off in the office.

From that perspective, the appearance of some cards is definitely more appealing than others, but I'll likely "defrock" the cards of their coolers before displaying so... Yeah.
 
Reading the Computerbase review, average is 30% faster in raster than 2080 Ti FE and exactly twice as fast as 5700 XT.

Edit: Held to 270 W, the average is about 25% faster. Going from 270 W to 320 W on the 3080 only nets you 4%.
 
Wow @ ComputerBase Vega noise test - that thing is a horrendous jet engine. Actually all the AMD cards do extremely poorly overall compared to nVidia.
 
Around 30% faster than RTX 2080 TI while using 25% more power.

Hmmmm...
Computerbase's review tests the 3080 power limited to 270 W. The extra 50 W doesn't really get it that much more performance. When compared with a 2080 Ti at the same 270 W power limit, the 3080 is about 20-30% faster.
 
AMD cpus seem to have at least 10% performance deficit @1080 and 1440. And that was with 1866 1:1 setup. Average Joe will have 15-20% deficit instead.

But then again it is 4K card?
 
Worst new generation on a new node after Fermi ???
Actually according to TPU GTX 285->GTX 480 is exactly the same uplift as RTX 2080 Super->RTX 3080 at 2560x1600/1440p resolution. This ain't no Fermi though, NVIDIA isn't six months late to the party and it does run pretty cool for a card that draws power right at its TDP limit.
 
As expected, it uses more power than nVidia claims, per TPU's spec. 348W peak gaming. The thermal limit is only 83C, which is fairly low, but can be manually set up to 88C with third party software. They also say that in stock form, it basically is always being power limited. Raising the power limit from 320W to 370W gives an additional 1% performance, so not much point there.

The card does really well at 4K, but is basically a complete waste of money for resolutions below that. Oh, and the cooler works well in an open bench setup.
 
Top end cards normally have a bigger gap at higher resolutions, but this time that seems more pronounced. So reviews have somewhat more mixed results, so might serve as Rorschach test. If you expect/project to the card as lackluster, you will note the lower resolution results, and if you expect/project it to be good, you will note the 4K results.
 
Reading the Computerbase review, average is 30% faster in raster than 2080 Ti FE and exactly twice as fast as 5700 XT.

Edit: Held to 270 W, the average is about 25% faster. Going from 270 W to 320 W on the 3080 only nets you 4%.

They took it upon themselves to adopt that last 3-4% strategy AMD has been using since Tahiti. I'd personally set the power limit to 270w and give up that 3% performance. I'm guessing the 3070 will have similar characteristics.

This is the worst node jump perf/w increase by Nvidia since Fermi. Disappointing.

I'll be in the market for RTX 3070 ish performance this spring. I'm looking forward to seeing if AMD can bring something better to the market.
 
The 3080 is 31% faster than the 2080ti and 67% faster than the 2080 according to TPU (4k). IMO the dissapointing aspect is power consumption at just 17% better than 2080ti at 4k. All in all, definitely not the best jump ever, but what was expected. Perf/$ is much better than Turing, obviously.

IMO AMD can definitely be competitive with the 3080.
 
Back
Top