If I'm reading his benchmarks correct, @ 1440p gaming the 4090 is 28% faster than the 3090 Ti, and @ 4k gaming it's 33% faster than the 3090 Ti. The rumored RTX-4090 being 60% faster, I think was meant in reference to the RTX-3080
In quite a few different videos it shows between 30% and 80% faster, depending on game. If that isn't impressive, I don't know what to tell you.
https://www.techpowerup.com/review/nvidia-geforce-rtx-4090-founders-edition/32.htmland with all the software trickery enabled. ~30% sounds right.
The power draw is the same, and sometimes lower, than the 3090ti. As seen in the hardware unboxing video.No, it is impressive. But it better for a whopping $1600 and a power draw like a space heater.
I'll wait for the RTX-4080 16GB price drop next year and grab one when pricing is back to reality
I'm not a 4k gamer I play at 34"Ultrawide 120hz refresh, so I don't need a card that gives more than 120fps @ 3440 X 1440
No kidding. Not entirely sure where they keep getting "30%" from. The hardware unboxing video (second one posted) shows 50% or more in 4k, and it's without "software trickery".https://www.techpowerup.com/review/nvidia-geforce-rtx-4090-founders-edition/32.html
45 percent faster at 4k, no raytracing or dlss. (100 divided by 69 relative = 45 percent faster Tha. The 3090ti).
Definitely overkill for most things... Honestly with the games on offer right now I sort of agree with Domingo too. We need harder games to run to really stretch this thing's legs.For 34" Ultrawide gamers the RTX-4090 seems overkill. I don't need 200fps
The power draw is the same, and sometimes lower, than the 3090ti. As seen in the hardware unboxing video.
The $1600 price tag is nuts, I agree, but the power isn't worth mentioning.
Being similar and producing 30% @ 1440p and 50+% @ 4k would actually mean it uses less power than it's predecessor to do the same amount of work.The 3090 Ti was ridiculed and knocked because of it's crazy power draw.
So the RTX-4090 being similar is not a good thing.
I thought hardware was supposed to progress to smaller and more efficient and still have faster performance.
All of the HWUB tests were with a 5800X3D and he had cpu bottlenecks on all of the 1440p tests. Without that bottleneck, who knows how far it could have gone. Still, 30% difference even with the bottleneck? Yikes.Noticed tech power up only used a 5800x cpu (non 3d). I'd imagine there are cpu bottlenecks in play there too .
Yeah but right now we have 30 series inventory clearing pricing for the entire 40 series. Wouldn't be surprised to see a price drop near end of year/ Q1 2023.To be fair, it's still a good card, considering I can get one sub $1000 CAD and this thing is $2200 CAD, before taxes.
The OCing by HU was unimpressive to say the least. Much more power draw for very little performance increase.I'm a bit disappointed in the performance as reviewed, but that's because this is now the 3rd gen we've gotten 30-50% at 1440p/2160p over previous gen flagship where we used to get north of 60% typically. A lot of that disappointment is due to the huge increase in shaders along with the node transition from Samsung 8nm to TSMC 4nm, I guess I was expecting gains more like Maxwell 28nm > Pascal 16nm which did a lot more with just clockspeed uplift and very little in terms of shader increase. Ada has both, which makes me wonder why its only seeing 30-50%. Honestly I'd expect to see 30-50% just from the increase in shaders from 10K to 17K at the same clocks, or just die shrinking GA102 to 4nm and increasing clocks from 2GHz to 3GHz.
I'm guessing there's a big difference due to the clockspeeds being tested at. AD102 was reported to boost to 3000MHz relatively easily on even stock cooling with 2520MHz default boost clocks, but in the few reviews that show sustained clocks its only getting 2400-2500MHz? I wonder if that's because AD102 can't sustain those Boost/Power Draw figures or it was mandated by Nvidia to not emphasize overclocking?
There's a lot to like about the 4090, like the temps, power draw, performance at 4K with or without DLSS, but it'll probably still be a game time decision for me. Not going to line up or anything, if I happen to see one online and available I'll probably snag it but I can't say I'm not a little disappointed and expected more.
Anyone see a review that shows sustained benchmark clocks at stock vs. overclocked? The only one I saw cite their clocks was TPU at 2400MHz which seems really low, and the "overclocking" coverage is basically non-existent, which makes me wonder if this is an Nvidia embargo on top of the embargo that was lifted (also heard no AIC reviews).