08:47PM EDT - We're here in Austin, Texas for NVIDIA's GeForce 2016 presentation

08:48PM EDT - Taking place in Austin this weekend is the Dreamhack conference, and while NVIDIA has not explicitly commented on why we're in Austin, it's a safe assumption that at least part of the reason is to tap into the Dreamhack crowd

08:48PM EDT - It will be a full house tonight with a number of press, along with the public coming over from Dreamhack

08:49PM EDT - NVIDIA is also broadcasting this event live on their Twitch channel: https://www.twitch.tv/nvidia

08:50PM EDT - Run time for the event will be roughly an hour and a half, though it's likely we'll start a bit late

08:52PM EDT - NVIDIA has not officially commented on what will be presented, but the wide expectation is new GeForce cards based on the Pascal architecture

08:53PM EDT - NVIDIA's GP100 GPU is already in production, of course, being used for the Tesla P100

08:54PM EDT - These GeForce products will presumably be a smaller, more economical GPU that yields better on TMSC's 16nm process

08:55PM EDT - Just waiting for everyone to get seated at this point

08:55PM EDT - NVIDIA staffers are all wearing "Order of 10" shirts, NVIDIA's promo stunt preceeding this event this week

08:56PM EDT - So expect to see something about an order of 10 tonight

08:57PM EDT - NVIDIA evangelist Tom Peterson is currently warming up the crowd and throwing out t-shirts

08:58PM EDT - The big question tonight is going to be performance expectations. GTX 980 Ti was built using a massive 600mm2 28nm GPU

08:59PM EDT - Any consumer GPU this early into 16nm is going to be a lot smaller

09:00PM EDT - So it will be interesting to see the tradeoffs made on die size savings versus manufacturing costs and the performance target NVIDIA needs to hit to suitably beat GM200

09:02PM EDT - Tom is leaving the stage, so we're about to begin

09:05PM EDT - Still waiting to start

09:07PM EDT - Alright, here we go

09:07PM EDT - First on stage: Jen-Hsun Huang

09:08PM EDT - Opening up saying that they couldn't think of a better place to do this than Dreamhack

09:09PM EDT - History lesson time. History of gaming on the PC

09:10PM EDT - For NVIDIA of course, PC gaming is a massive market, even with their efforts to diversify

09:10PM EDT - "We've dedicated ourselves to advancing this platform"

09:11PM EDT - 4 things: New Art Form, New Sound, New King, New Tech

09:11PM EDT - "A brand new technology you've never heard of before"

09:12PM EDT - Taking screenshots as an artform

09:12PM EDT - Introducing a 3D in-game camera system: Ansel

09:13PM EDT - Instagram for the PC gamer?

09:14PM EDT - Free-moving camera, filters, capturing at greater than the screen resolution, and 360 degree stereo captures

09:15PM EDT - Demo time

09:15PM EDT - Looks like NVIDIA is locking the render state and then letting the user move around

09:17PM EDT - So this doesn't require the game engine to diredtly participate, but it wil be interesting to see what compatibility is like

09:18PM EDT - Demoing viewing a photograph on the Vive

09:19PM EDT - Photos can also be displayed on phones via Google Cardboard

09:21PM EDT - Listing several games that will be supported

09:21PM EDT - Next NVIDIA Works project: VRWorks Audio

09:22PM EDT - Sounds like an on-chip audio processor, ala AMD's TrueAudio

09:22PM EDT - Jen-Hsun is comparing it to physics modeling

09:22PM EDT - Based on the company's Optix technology

09:23PM EDT - Acoustically accurate modeling

09:25PM EDT - Still unclear if this is dedicated hardware, or just an application on top of the CUDA cores, with NVIDIA focusing more on the middleware

09:26PM EDT - Now rolling a video of a new demo, NVIDIA Funhouse

09:27PM EDT - Second announcement today: major upgrades to VRWorks for more physically-accurate simulations

09:28PM EDT - All current consoles are based on one current architecture: x86

09:28PM EDT - Not mentioned: that all of the consoles are also based on AMD's GCN architecture

09:28PM EDT - Talking about how a common platform has improved production values

09:29PM EDT - Demo time: The Division

09:29PM EDT - Just showing off the graphics quality

09:30PM EDT - More games; Tomb Raider, Mirror's Edge

09:31PM EDT - All three games were on maximum quality

09:32PM EDT - They were running, of course, on NVIDIA's new card

09:32PM EDT - GeForce GTX 1080

09:33PM EDT - Pascal, of course. Several thousand people have been working on it for over two years

09:33PM EDT - R&D budget was several billion dollars

09:34PM EDT - Most efficient and advanced architecture NVIDIA has ever created

09:34PM EDT - 16nm TSMC FinFET, GDDR5X

09:35PM EDT - Focus on card craftsmanship as well

09:36PM EDT - Energy efficiency goes hand in hand with performance. "Moore's Law is running out of steam"

09:37PM EDT - Discussing power delivery efficiency

09:38PM EDT - GPUs are primarily fed at 12v, so power needs to come down to around 1v in today's GPUs

09:40PM EDT - GTX 1080 has to beat GTX 980's power delivery efficinecy

09:40PM EDT - 1080 is faster than 980 SLI

09:40PM EDT - (No comment on under what game, given than AFR's limitations)

09:40PM EDT - Faster than a Titan X

09:41PM EDT - Titan X: ~3.6. GTX 1080: ~4.3

09:41PM EDT - So around 20% faster?

09:42PM EDT - "The Pascal family is going to be pretty amazing"

09:42PM EDT - Not charted: GTX 980 Ti. Which woiuld be at around 3.5

09:43PM EDT - Now on stage Tim Sweeney

09:44PM EDT - Back of the envelope calculation says that GTX 1080 should be around 25% faster than GTX 980 Ti, using NVIDIA's numbers

09:44PM EDT - Though for NVIDIA, they're going to want to focus on getting GTX 700 series users to upgrade

09:45PM EDT - Jen-Hsun and Tim chatting about the importance and progress of the PC

09:46PM EDT - This segment is supposedly unrehearsed

09:46PM EDT - Tim's response: this bridges the gap between photorealistic graphics and real-time graphics

09:47PM EDT - Pitching Epic's Paragon

09:49PM EDT - Showcasing real-time photorealistic rendering of Paragon models

09:50PM EDT - (For varying definitions of photorealistic, since the art style is slightly exaggerated/cartoony)

09:52PM EDT - "The future of graphics, available today"

09:52PM EDT - 2.1GHz GPU clock

09:53PM EDT - 11Gbps memory clock

09:55PM EDT - Now talking about display tech

09:55PM EDT - And specifically, multi-projection

09:56PM EDT - "Simultanious multi-projection pipeline"

09:56PM EDT - Up to 16 independent viewports

09:57PM EDT - What can we do with that?

10:03PM EDT - Demo time with Tom

10:04PM EDT - Correcting a multi-monitor display

10:08PM EDT - Fixing the warp/projection errors with an uncorrected setup

10:08PM EDT - Now multi-projection for VR

10:10PM EDT - Pre-distorting images to counter lens distortion

10:12PM EDT - Not clear how this is different from multi-projection acceleration

10:16PM EDT - 2x perf and 3x efficiency vs Titan X wheb using Pascal's special features

10:17PM EDT - The new king: GTX 1080

10:18PM EDT - $599

10:19PM EDT - $699 founder's edition

10:19PM EDT - Available May 27th

10:20PM EDT - GTX 1070, $379, June 10th

10:20PM EDT - Looks like reference cards are the founder cards?

10:20PM EDT - Wrap-up time

10:22PM EDT - We're done here



View All Comments

  • exileUT - Saturday, May 7, 2016 - link

    The last two Ti models released on average 7 months after the x80 model. Reply
  • Dug - Friday, May 6, 2016 - link

    Why wait when it's as fast as 2 980's in SLI and is priced at $600?
    Once the Ti comes out, are you going to wait for the 1180?
  • RussianSensation - Saturday, May 7, 2016 - link

    It's just logic. If 1080 is 25% faster than a 980Ti and that took 1 year, GP100 with 3584 CCs would be up to 40% faster once it's also overclocked and it'll take about a year to get here. If NV released a 3840 CC full fat chip 780Ti successor for $699, then 1080 for $599 looks really bad for anyone that has a decent GPU that's powerful enough to grind out over the next 12-15 months.

    No doubt, 2560 CC 1080 with 1.733Ghz Boost and 2.11Ghz overclocking is impressive but that makes $700-750 GP100/102 EVEN more impressive in light of that!
  • Yojimbo - Friday, May 6, 2016 - link

    Jen-Hsun Huang should think of more adjectives besides 'amazing'.

    The simultaneous multi-projection seems pretty awesome for VR. If it has a similar output quality it seems to undercut AMD's focus on multi-GPU rendering for VR. I imagine they can still output to the various viewports at various resolutions (multi-res rendering). But it seems to me that the SMP method of correcting for lens distortion might allow for less accurate correction than the standard method (potentially) does. With SMP there are just 4 different projections per eye, each projection itself monolithic. But with the standard method it seems one should be able to correct as accurately as one wishes/has the information for. I have no idea what is done in practice, though. That's if I'm understanding the two methods correctly.
  • Yojimbo - Friday, May 6, 2016 - link

    Oh the other thing I noticed is that, despite saying there can be up to 16 view ports at a time, Huang only proposed to use 8 of them for VR. What's the application for 16 then? A Light field HMD? Reply
  • D. Lister - Saturday, May 7, 2016 - link

    VR probably incorporates two view ports simultaneously, so 2x8=16. Reply
  • Yojimbo - Saturday, May 7, 2016 - link

    In his demonstration he showed 2x4 = 8. Reply
  • Mr Perfect - Friday, May 6, 2016 - link

    Jen-Hsun will probably stop using"Amazing" when he no longer considers leather jackets business wear. Reply
  • D. Lister - Saturday, May 7, 2016 - link

    Seeing him on stage, I alway get this feeling that the next time he would ditch the whole biker thing and just go full awesome with green spandex, and a cape with nvidia logo on it. :| Reply
  • xaueious - Friday, May 6, 2016 - link

    Nvidia's home page is showing 1.7GHz as base clock, but in their presentation their sample was clocked at 2.1GHz on the stock cooler. If this is any potential indication, even though the 1070 is severely cut down, it might be able to make up for a lot of headroom with overclockability. Reply

Log in

Don't have an account? Sign up now