https://hexus.net/tech/news/graphics/121535-nvidia-geforce-rtx-2080-3dmark-time-spy-scores-surface/
RTX 2080 - 10,030
GTX 1080Ti - 9,508
RTX 2080 - 10,030
GTX 1080Ti - 9,508
Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
The score is just going to make the 1080ti look good until it runs out of stock if that ever happens.
If the gfx score was 11,000-12,000 I would consider getting a 2080 but not for that slight bump.
The score is just going to make the 1080ti look good until it runs out of stock if that ever happens.
If the gfx score was 11,000-12,000 I would consider getting a 2080 but not for that slight bump.
The 1080 was the same way compared to the 980ti. The 2080 is probably better off tbh.
This and the nVidia charts all suggest the 2080ti being 50% over the 1080ti in traditonal games ignoring the new features. That’s better than I think anyone imagined. These things will sell like hotcakes if the reviews match up.
RTX series are a flop.
What nvidia charts? The only chart I saw was the non-ti 1080 vs. non-ti 2080.
What I was thinking. People will lose their minds if that's true. No where does it state definitively that this is a 2080. Granted, stock Boost clock is higher on the 2080, but we've been told that Turing has more overclocking headroom than Pascal.What if that's a 2070 ?
Interesting. Don't cancel that pre-order just yet, but keep your finger on the button.
A 300 MHz overclock to make the 2080 beat a 1080Ti would align perfectly with the shader math assuming zero IPC improvement over Pascal.
1080Ti at stock boost clocks = 11.3 TFLOPS
2080 at FE stock boost clocks = 10.6 TFLOPS
2080 at 2025 MHz = 11.9 TFLOPS
That would align perfectly with the Time Spy result and clock speed.
I'm not saying it's real, I'm just saying what the arithmetic equates to.
The 1080 was the same way compared to the 980ti. The 2080 is probably better off tbh.
FWIW, here's my 1080 Ti at 2025 MHz core and 12GHz on the GDDR5X side by side with that timespy result (my nvidia driver was from back in April, to be fair):
https://www.3dmark.com/compare/spy/4293425/spy/3628077
Show me a stock 1080 only posting a 5% better timespy graphics score than a 980 ti. In games, the 1070 went neck-and-neck with the 980 ti, the 1080 blew it out of the water. [H]'s recent mutli-generational testing showed the 1080 sitting with an over 30% average increase of the 980 ti. So, no, the 1080 was not "the same way" compared to the 980 ti.
980ti OC’d way better. In some games only a few points apart.
I remember all the people complaining then, too. It’s like ground hog’s day.
that's not what he asked. Stock Vs Stock.
And people were complaining because the 1080 FE was a ridiculous price and people could see that it was only going to lead to further price rises in the future.
that's not what he asked. Stock Vs Stock.
And people were complaining because the 1080 FE was a ridiculous price and people could see that it was only going to lead to further price rises in the future.
well, the scores aren't to far off the Tflops ratings..
a 1080 gtx is 9 TFLOPs
a 1080 ti is 11.3 TFLOPs
a 2080 is 10.1 TFLOPs (according to anandtech)
a 2080 ti is 13.4 TFLOPs (according to anandtech)
so.. i would guess that a 2080ti is approx 20-25% faster in old non-RT games.. i really hoped for more - so please surprise me nVidia!!!
but it sure would explain why Jensen thinks that we need to look at performance in a new way..
980ti OC’d way better. In some games only a few points apart.
I remember all the people complaining then, too. It’s like ground hog’s day.
Not going to do the math on everything but from rough estimates it looks the 980 ti OC's is under 5% from a stock 1080. With both OC'd that brings the difference up to around 15%. With custom coolers I imagine the 1080 ends up being a little closer to 20% better and the 1070 probably catches up to the 980 ti (or, at least, gets close). In games that make heavy use of VRAM the 8GB on both cards will put them over the 980 ti. The OC to OC difference is still 200% more than the 5% difference shown by these. That said, these could every well be on non-final drivers so that difference could change.
And yes, people complained. Because prices went up and people worried that it would mean they would go up more. Guess what, they went up more and people are not happy. Unless these cards show Pascal level improvements (highly unlikely) then people are going to be, justifiably, upset at the price increases.
I kinda wonder if Nvidia is sandbagging a bit till Amd, or other alternative company, puts something out to challenge, then bam! Why go all out when you can dole it out piecemeal to stay on top? Plus we have all seen improvements on the driver front.
I think it’s part of the marketing. You know, like the crack cocains. Once you hooked.... Just a hypothesis. Surely major tech companies wouldn’t take advantage of us scmucks and squeeze us for what it’s worth.
Different architectures to get to the same point. RTX is new tech, it brings it's focus on Ray Tracing. 3dmark isn't even that relevant anymore to be honest ...I wish it was (I miss me some 2001 SE)...
6 months or one year from now a killer app may show a much more distinguishable difference.
They said that with bulldozer and vega.
We've been through this before. The cards may end up being similiar in fps and performance due to different ways of getting there and doing it, but this one has Ray Tracing.