On the subject, doubt AMD is that worried, isn't any point. nVidia is certainly taking Big Navi seriously though.

People tend to forget that 5700XT is only a mid range card with 40CUs. A 60CU option would have been ~ as fast as a 2080Ti, and 80CU option considerably faster. We also know that a 1.8Ghz 52 CU unit uses maybe 130W of power in the nextbox, so there's lots of scope for scaling the number of CUs up.

Disclaimer of course being that we don't know anything much solid about Navi2 yet except for the nextbox/ PS5 stuff, and neither MS nor Sony are going to talk down their own hardware.

An related to the last point: It is nVidia's video though. They're not exactly objective.

Case in point, claimed up to 1.9x performance/ watt over Turing. So far as I can tell the performance/ watt improvement is nowhere near that*. Using nVidia's own figures, since that's what is available.

2nd case in point: RTX I/O. It's MS technology (DirectStore, iirc) that won't be available for computers until next year anyway. Much like G-Sync compatible == Freesync they've just given someone else's thing their own branding.

*3070 ~ 2080Ti in claimed performance. 220W TDP vs 250W TDP actually gives 12% performance/ watt improvement, not 90%. And to put the tdp in perspective that's only a bit less than the top line and top hot meme GTX 480

Can I have a suggestion in an AMD flavor? Nvidia is worthless to me as a Tuxian, sorry.

nVidia has dominant market share and just announced a 70-80% performance improvement at the same price point along with DLSS exclusivity and you don't think AMD are concerned about that? Come on.

They'll be fine of course, Ryzen is great and they have both console contracts, but let's not pretend this wasn't a blow.

Does anyone know if these improvements drip down to Tegra processors?

If the rumored Switch Pro coming up has a similar performance jump, that'd be dope.

Concerned is a far better term than worry, certainly. End of the day though AMD can only do as well as they can, so there's no point worrying about what nVidia can put out- always better to focus on your own performance since you can control that. There's also not much point worrying about mindshare, it's obvious that a lot of people want big navi competitive solely so nVidia prices are forced down rather than because they'd buy it but there isn't anything that can be done about it.

I'm extremely skeptical about the long term success of RTX and DLSS as differentiators, though I'd freely admit that is a personal position. We also have extremely limited 3rd party information about Ampere performance- the nVidia product launch is a one hour advert, after all, and they tend to be... selective in what stats they share.

In terms of whether AMD is concerned, it really depends on what they bring to the table. If they're sitting on a well scaled 80CU RDNA2 card with 30% 'IPC' gain over RDNA1 and better efficiency, considerably smaller die than Ampere, good raytracing etc then they may well be happy about their position because they'd have a very good chance of full spectrum competition. OTOH, if they were bring Vega64mk2 to the table they'd be worried, but they certainly aren't doing that.

I'm not a big fan of DLSS either, but most people seem to love it and the media is pushing it hard. Whether it gets support when consoles will surely use a different upscaling technique I don't know, but every title coming soon seems to be using it. nVidia being in such a dominant position on PC gives them a lot of leverage.

As for ray tracing, it's the future whether people like it or not. The consoles are pushing it, and if the 3000 series doubles its performance like nVidia say then there's nothing really preventing it from catching fire on PC. It was kind of a joke on the 2000 series because of performance, but it's not going anywhere long term. I'll certainly keep choosing framerate over it on my 2070, but it sounds much more viable on a 3070.

Might upgrade to the 3090, the headroom for keeping future games at 4k 60 without upgrade to keep up with them is grabbing my attention. 8k 60 fps is a big flex and I hope benchmarks show this when this card comes out. if it fails to impress I will jump to the 3080 and stick with that. gonna give my sister the 2080ti and get her into some Witcher and cyberpunk when it drops here.

the 3090 here is 200 dollars more after tax than the 2080ti Canadian.

2200 after tax.

Sorry, you're stuck. Single slot you have the gt1030 and dual slot the gtx1050ti/1650. No real option from AMD. You can try to find a Msi Rx550 (not single slot and very rare), maybe a weird Rx460 but that's it.
The low end has been prety much adandoned and for good reasons. Even the integrated Intal GPU are getting close to a gt1030, for the price of a low-end GPU you can get a free modern, probably faster processor as well.

Yes, I'm skeptical only about the proprietary nVidia (TM) RTX (R) implementation specifically, not raytracing in general. On windows a 'generic' DXR solution seems a far better bet, since it will be on xbox as well, and a more hybrid approach of using general compute power which AMD seems to favour seems a better bet since it's more flexible- the specialised cores on RTX cards comes at a significant space premium, the one fairly solid bit of info on big navi is die size and at 505mm^2 its 20% smaller than even the GA102 in a 3070 (and probably 3060).

Note the "8k" stuff is with DLSS, so you're rendering at 4k. Rendering at actual 8k will be a pipe dream for many years. Even 4k is largely being ignored by PC gamers in favor of 1440p and high refresh.

This time Orkhepaj is sorta right in posts 3 and 5....as beyond as certain range(which varies a bit in the population, but is around the same for most) the human brain/eye cannot notice much difference.

---------

To be fair, many such companies just make minor changes to old tech and sell incremental improvements....it's a very lucrative business model, sadly.

well you can't argue with the developments between the last 2 generations, even if the tech is old or improvements are ' only ' incremental

Well yeah, there are improvements(never said there weren't).

The thing I was trying to point out is more how they drip feed people small updates(via new models and such) to get more money out of people...especially those who like to constantly buy such to stay on the cutting edge.

Their thinking: why should they release say a GPU with say a 75% better performance over the last one when they can release it as 2-3 more smaller increases in performance?

well i do agree that the setup is aimed at making the tech companies wealthier with less eye for the customer within foreseeable limit. i'm not sure if the eye for the consument did lessen or actually improved over the years, maybe that is also director limited. the improvements over the years do seem to follow moore's law so again i can't be entirely sure if the products could be made a lot better without to many cost involved. The whole release of tech does already seem to be ahead of the software side of life so again i'm not sure if the hold back improvements, if any , are also there for the software side of life trying to prevent a to big of gap between software utilization and hardware possibilities.

in a way the technology branch did also became of the same level of importance as say the oil winning branch, not saying that it is a good equasion but the price increases and production levels set by an alliance as opec does seem to behave in maybe similar ways as the tech advancements seen

Okay, what are these then?