We also have a separate upcoming piece involving Schlieren photography of the RTX 3080 FE card, but we can show a short clip of that here as a preview.

Note: This is a transcript from our video review of the NVIDIA RTX 3080 Founders Edition card.

The RTX 3080 is a $700 card, and the Founders Edition with its flow-through cooler launch today.

We will be reviewing the Founders Edition card here, but we know there are a lot more requests for specific tests with this video card.

We’d like to quickly address one point before proceeding: A lot of commenters have expressed concern about the 10GB framebuffer on the RTX 3080, remarking that 10 is lower than 11, and therefore somehow worse than the 1080 Ti or 2080 Ti.

The Strix 2080 Ti and FTW3 2080 Ti are significantly more consistent in coldplate flatness, with medians at about 29 microns and 34 microns, respectively.

The 3080 FE leads the 2080 Ti Strix stock card’s 67.8FPS AVG result by 25.5%, it leads the 2080 Super stock’s 54FPS AVG by 58%, the 2070 Super by about 77%, and it’s about 90% ahead of the 1080 Ti.

Tested at 4K, the 3080 held a lead over the 2080 Ti of 25.5%, but that lead drops to 20.6% here.

Leadership versus the 1080 Ti FTW3 stock is 80% here, down from 90% at 4K.

Versus the 2080 Super, the 3080 stock leads by 48% (as opposed to 58% previously).

The 5700 XT runs 78FPS AVG, allowing the 3080 a lead of 63% -- but it’s not like the 5700 XT competes at the same price class as the 3080.

The 3080 leads the 2080 Ti by 18% here, down from 25% at 4K; it leads the 1080 Ti by 71%, down from 90% at 4K.

The 3080 FE stock had an 8% improvement against itself with Dx12, or 10% when overclocked.

Interestingly, the Pascal generation -- we can highlight the 1080 Ti and 1080 non-Ti entries -- and the 5700 XT both show no real change between the two.

The 2080 Ti Strix ran 144FPS stock, with a much bigger 12% gain from an overclock.

OC-to-OC, the 3080 leads by 14%, or 24% stock.

The lead against the 114FPS AVG 2080 Super is 58%, or again about 90% versus the 1080 Ti, at 94FPS AVG here.

1440p had the 3080 at 329FPS AVG, the 2080 Ti at 271, the 1080 Ti at 187, and the 5700 XT at 185.

That establishes a reduced 21% lead over the 2080 Ti and 76% over the 1080 Ti, down from 90%.

The stock result runs 431FPS AVG, 10% over the 2080 Ti and 50% over the 1080 Ti FTW3.

The lead over the 2080 Ti OC is minimal, at 13%, and so panic-selling 2080 Tis for $500 is probably not the best idea.

The lead over stock is 25%, consistent with previous results.

At this point, we’ve established an obvious trend: We’re 25% over the 2080 Ti, 73% over the 2080 Super stock and similar for the 5700XT, and 82% over the 1080 Ti.

At 1440p, the 3080 runs 23% faster than the 2080 Ti, 70% higher FPS than the 1080 Ti and 5700 XT, and 47% ahead of the 2080 Super.

This is clear for a few reasons -- like identical performance between the OC 2080 Ti and the 3080 -- but also because of the dropped lows on the cards bouncing off of the CPU bind.

The 3080 leads the 2080 Ti by about 20% here, it leads the 2080 Super by about 46%, and the 1080 Ti by 60%.

The 5700XT is close to the original GTX 1080 FTW, both of which are doubled.

The 2080 Ti OC’s 74FPS AVG result illustrates that you can get close-ish to 3080 performance with a little bit of work, while baseline allows the 3080 a more impressive 30% lead than in previous tests.

The 2080 Super is led by 61%, the 1080 Ti is led by 78% here, less than in some other titles.

At 1440p, the 3080 OC lead is reduced to 4.6%, the 2080 Ti lead remains about 28%, and the 1080 Ti lead is reduced to 68%.

The 3080 ran 87FPS AVG, leading the 2080 Tis’ 65FPS AVG by 36%.

The 2080 Super stock is led by 75% here, with the 2080 FE led by 82%.

At 1440p, the 3080 runs 130FPS AVG, leading the 2080 Ti Strix by 32%, still above previous averages.

The 2080 Super and 2080 FE are near one another, led by a split difference of 71%.

At 1080p, the 3080 FE gains to 163FPS AVG -- sadly, the highest we’ve seen in Minecraft RTX -- while still leading the 2080 Ti by 34%.

The 2080 Super is led by 67%, so there’s stagnation in the gains versus the Ti and Super.

At 4K and with DLSS on, the 3080 runs at 65FPS AVG, with lows consistently behind.

The RTX 2080 Ti runs 49FPS AVG, allowing the 3080 a 32% lead.

The 2080 Super and below are all under 40FPS AVG. .

At 1440p, the 3080 runs 111FPS AVG, 28% ahead of the 2080 Ti Strix and about 63% ahead of the 2080 Super and FE

It’s not quite there at 4K, but the interesting part is that the 3080 leads even harder than it did previously: It’s now at 45% higher average framerate than the 2080 Ti, or 90% ahead of the 2080 Super card, which is insane

1440p sees a reduction in lead, because we’re stepping away from the memory bandwidth advantages afforded to the 3080

This is not total system power and is instead total card power

Starting with FurMark power virus tests, the RTX 3080 stock GPU consumed 323W total card draw, ahead of the 2080 OC by about 30-40W

The efficiency is way up here, despite power consumption also being up; for instance, the 1080 Ti consumed 325W when overclocked, and yet the 3080 stock card was routinely in the 70-90% uplift range versus this card

The 2080 Ti pulled 264W stock and 330W overclocked, for comparison, with the 2080 FE stock at 235W

With that 330W OC number, the 3080 was often still 10-15% ahead of the 2080 Ti OC while being a few watts lower, proving its efficiency improvements

To get the actual 2x perf/watt number marketed, you’d have to power limit them artificially to about 240W, so that scale converges towards the mean as power increases on the 3080

The 3080 FE OC draws 367W here, proving exceptionally inefficient as compared to its stock setting, particularly thanks to a lack in gains

The 3080 stock pulled 314-320W, with the 2080 Ti Strix overclock at 306-331W, depending on workload

That’s a lot of power to allow a 13% lead to the 3080 stock

The 1080 Ti stock card was closer to 283, illustrating how much more is being done per watt with this generation

Finally, here’s an overclock stepping chart from our process of overclocking the 3080 FE card

Editorial, Testing, Test Lead: Steve Burke