Nvidia shows signs in [2018]

[sonen]

They are trying to keep the average selling price high so people stay used to it as normal. IMO

It's the nature of any publicly traded company.
Once you reach certain numbers (margins, revenue, profit) anything less than the last quarter means you are under-performing.

 

[manux]

Nvidia's/Jensen's siggraph stream is about to start


Anandtech has liveblog here: https://www.anandtech.com/show/13215/nvidia-siggraph-2018-keynote-live-blog-4pm-pacific

 

[Malo]

I like Anandtechs live blog format.

Turing is a monster.

Has Volta basically been replaced already?

 

[McHuj]

Volta is old. However, Volta has really high double precision performance. Doesn't see that NVIDIA is stressing double precision with Turing. It definitely looks like an architecture split for different markets now.

 

[silent_guy]

Volta is old. However, Volta has really high double precision performance. Doesn't see that NVIDIA is stressing double precision with Turing. It definitely looks like an architecture split for different markets now.

I see no reason to believe that Turing isn’t Volta with FP64 removed and RT engines added. In which case it’s as much an architecture split was it was for GP100 and GP102...

 

[Malo]

And the name difference is to differentiate target market only?

 

[silent_guy]

And the name difference is to differentiate target market only?

That. And the fact that it has ray tracing.

JHH talks about TIPS (terra integer per second) in the presentation, and states that they can’t do this concurrently with TFLOPS. That’s a core Volta characteristic right there.

 

[Deleted member 2197]

NVIDIA Open Sourced a Video-to-Video Translation Technique using PyTorch – and it is Super Impressive
August 20, 2018

But research in the area of video processing has been painstakingly difficult. For example, can you take a video sequence and predict what will happen in the next frame? It’s been explored, but not to any great avail. At least until now.

NVIDIA, already leading the way in using deep learning for image and video processing, has open sourced a technique that does video-to-video translation with impressive results. The goal of this research, as described by the researchers in their paper, is to learn a mapping function from a given input video in order to produce an output video which depicts the contents of the input video with incredible precision (as you can see in the above GIF).

If you were impressed with our last NVIDIA article on converting a standard video into slow-motion, this latest research will leave you stunned. And it’s not just limited to recreating real-world scenarios, it can even predict what will happen in the next few frames! When compared to baseline models like PredNet and MCNet, the vid2vid model produced far superior results.

https://www.analyticsvidhya.com/blog/2018/08/nvidia-open-sourced-video-to-video-translation-pytorch/

 

[DavidGraham]

Now that the mining frenzy is over, NVIDIA's Q2 2018 market share is back to the pre mining era levels (~70%).

https://www.jonpeddie.com/press-rel...arch-releases-its-q2-2018-add-in-board-report

 

[trinibwoy]

Was there ever confirmation or acknowledgment of excessive Pascal inventory? Don’t recall anything besides the Digitimes rumor article.

Pascal prices have gradually come down to MSRP levels but I would expect deeper discounts if there was oversupply. It’s either stock levels are actually ok or Pascal is still going to be around for a while and there’s no rush to clear inventory.

 

[DavidGraham]

Pascal prices have gradually come down to MSRP levels but I would expect deeper discounts if there was oversupply.

There are theories out there that NVIDIA priced RTX cards high enough so that they don't interfere with the prices of the overstock of pascal cards.

 

[Kaotik]

I wouldn't be surprised the slightest if there will be a big price cut on RTX's before the holiday season - given they get rid of the assumed excess stock of Pascals

 

[Malo]

I wouldn't be surprised the slightest if there will be a big price cut on RTX's before the holiday season - given they get rid of the assumed excess stock of Pascals

Apparently the AIBs are making very little off RTX. Combined with the FE pricing, I highly doubt prices will be going down anytime soon. The only way that would happen is if Nvidia cut FE pricing and give discounts to AIBs but Nvidia are in too strong financial position to need price cuts.

 

[troyan]

Apparently the AIBs are making very little off RTX.

AIBs are free in the pricing. If they are making very little than their prices are to low.

 

[Malo]

AIBs are free in the pricing. If they are making very little than their prices are to low.

Probably, but we don't know that do we? And they can't just throw an extra $100 margin on the cards. They're competing with each other AND Nvidia FE. All which are way above this supposed MSRP. The point was that supposedly their costs from Nvidia and making the GPUs is very high and their margins are in the 3% area which makes the MSRP absolute bullshit. That or the AIBs are spinning misinformation.

 

[Geeforcer]

There are theories out there that NVIDIA priced RTX cards high enough so that they don't interfere with the prices of the overstock of pascal cards.

Competing with yourself is actually a lot less fun than it sounds....

 

[CarstenS]

I wouldn't be surprised the slightest if there will be a big price cut on RTX's before the holiday season - given they get rid of the assumed excess stock of Pascals

Cut or drop?

 

[Kaotik]

Cut or drop?

Not a native English speaker, so let's go with a drop. Thought cut would mean the same (because the same term in finnish means price going down)

 

[Malo]

Not a native English speaker, so let's go with a drop. Thought cut would mean the same (because the same term in finnish means price going down)

It does mean the same thing in english.

 

[DavidGraham]

It does mean the same thing in english.

I think what Carsten meant is:

Cut: an official price cut from NVIDIA across all partners and AIBs.
Drop: an unofficial price cut transmitted through the back channels and distributors