3 ms·
RTX 2060 has Tensor cores on a mainstream, desktop offering, which might be of some interest if you want to play with deep learning at home Edit: I'm also stic
by ohithereyou 8y ago
RTX 2060 has Tensor cores on a mainstream, desktop offering, which might be of some interest if you want to play with deep learning at home
Edit: I'm also sticking with a GTX 970 right now, mainly because I game at 1080p. When I can't get a consistent 1080p@30FPS (though I prefer 1080p@60FPS) then I will consider upgrading. Right now I'm CPU bound, not GPU bound, in any title where I get sub-60FPS.
- wlesieutre 8y agoTrue, and I think these are also used for their Deep Learning Super-Sampling system, which is another potential difference between the 1070 and 2060. Render at lower resolution and upscale with a specially trained neural network for each game.
- eklavya 8y agoIs inference fast enough to not have latency issues?
- wlesieutre 8y agoI assume there's some latency since it would need a fully rendered frame to operate on, but other postprocess AA techniques would suffer the same issue. Not a graphics expert, but I think only doing real supersampling would happen during the initial frame render instead of in series afterward, and that's going to tank performance much further than DLSS (in terms of individual frame time rather than latency, but much more significant overall). The DLSS neural network is trained off of renderings with 64x supersampling and tries to apply that to a low resolution image. In practice I wonder if there will be situations where the neural network screws up and makes strange artifacts, but the demos look impressive.