Reputator

Yeah good luck with that.

"The power to impeach wasn't put in place for Presidents that break the law!"

This statement contains a bunch of assumptions about the future you literally can't make. "Remaining top tier" isn't an argument in favor of buying a new graphics card. I don't care how much it titillates your I/O ports. No one was disappointed at the performance of the 10 series. Prior to that, we were stuck on 28nm for three generations, so there was some stagnation, albeit out of the hands of any IHV. Aside from you, however, I've not heard anyone lament the stagnation of features though. You have a very utopian-like concept for the role of consumers. Be realistic. Few people have the money to throw away on an investment in the future with very few short-term gains. Most of us aren't venture capitalists. We don't "have to" encourage anything if NVIDIA and the developers fail to give us incentives beyond empty promises. If you think otherwise, I have some snake oil you might be interested in.

You'll be confused at any community disappointment if performance remains virtually the same, for the same cost? You GREATLY overestimate how much people care about the potential of raytracing. Again, purchasing a first generation of an unproven technology because it will encourage the technology to advance isn't the job of consumers. Bring the horse, then you can have the cart. It's not on us to assume the risk, which given the pricing, would indeed be the case for all but the top-end 2080 Ti. This is a next generation product. New features are nice, but improving the performance of last gen cards is expected. Not a bonus.

The 2080 Ti gives you more performance at least, but that's not the case for the 2080 or 2070 based on specifications.

No, consumers don't spend money purely based on an ideal. Telling people they need to buy these cards, whether or not they get their money's worth out of it, because it will encourage worthwhile products later literally makes no sense. If you weren't ignoring my earlier responses you'd already know.

Consumers are not investors.

Sorry, I'm a bit too cynical for all that. A push forward? Sure. To a future NVIDIA controls. This is yet another attempt to stack the deck against would-be competitors, seeding the software landscape to grow in their favor. This is not some altruistic quest for the next frontier, and NVIDIA assumes no risk in implementing these unproven technologies with their market dominance. They can offer stagnant advances in performance, throw in some features that currently amount to marketing hyperbole, and stick consumers with the bill. And they are.

This is only true if you look at the flagship card. If you're looking at the 2080 and 2070, you're actually getting worse performance for the money compared to the GTX 10 series. Tensor cores and RT cores are costing not just consumers more money, but a large part of the die space that could have been given to more substantial gains in traditional game performance. Actually around half the die space (more if you count INT32) based on the block diagram showed in the presentation. That isn't a trade-off most gamers would have chosen, if they in fact had a choice on the matter. None of the new portions of the Turing architecture have any proven benefit for gamers, aside from sketchy demos and some vague promises.

Maybe that's why NVIDIA announced a new form of AA (adaptive temporal) just before this launch!

LOLra Croft Worse than PhysX.

Just wait for reviews to come out for this next generation. That will be crucial to any sort of decision making.

Oh yeah. It's gonna be the most exciting time in graphics since the late 90s.

Or $400? You lower middle class bastards better start getting rich!

Stuff like this can and has been faked with simple texture tricks for ages now. Like PhysX, the performance hit of something like this can be substantial, I fear. Yes, NVIDIA created raytracing units inside the GPU to achieve these effects in an effort to keep the rest of the GPU from diverting resources, but the rendering pipeline still has to wait on those units to finish calculations before outputting a finished frame, which means you can still get nasty stalls.

Begun, the price gouging has.

RTX 2070 = $499 https://wccftech.com/nvidia-geforce-rtx-2070-official-launch/ RTX 2080 = $699 https://wccftech.com/nvidia-geforce-rtx-2080-8-gb-official-launch-price-specs-performance/ RTX 2080 Ti = $999 https://wccftech.com/nvidia-geforce-rtx-2080-ti-11-gb-official-launch-price-specs-performance/ $400 option? Too bad. Fuck you. Should have given NVIDIA a better competitor.

The investigation will be passed on to someone else. Also there will be lots of sign picketing.

I'm pretty sure they will anyway.

I almost always guess my own state.

That's nice, but it's a little unfair. We're talking RISC here. The whole point of RISC is to be more efficient, without the baggage of x86 support. And the drawback is always the same: you're not going to have nearly the same application support.

I always assumed it was real.

No, it's this: https://www.npr.org/2018/08/16/639125774/hundreds-of-newspapers-denounce-trumps-attacks-on-media-in-coordinated-editorial

To be fair, this may be the first time he's ever credited his source.