NVIDIA CES 2014 Press Conference Live Blogby Anand Lal Shimpi, Brian Klug & Jarred Walton on January 5, 2014 10:52 PM EST
- Posted in
- CES 2014
12:35AM EST - and we're done! Thanks for following
12:34AM EST - "more about Denver in the coming months"
12:33AM EST - Looks like we're done!
12:31AM EST - I'm looking around the room trying to figure out who from NVIDIA will be most likely to spill details on Denver
12:29AM EST - Jen-Hsun confirmed that the crop circle from earlier today was NVIDIA's doing
12:27AM EST - "Tegra K1 is probably the most ambitious project that we have ever worked on"
12:27AM EST - "Maybe this is more of a confession"
12:26AM EST - One more surprise
12:26AM EST - "I was so impressed it took my breath away"
12:25AM EST - oh no Jen-Hsun is losing his voice
12:25AM EST - NVIDIA's press release calls Denver's two cores "super cores"
12:24AM EST - I think disclosing that it's a 7-wide design was very deliberate in light of what we know about Apple's Cyclone
12:23AM EST - I want to know so much more about Denver
12:23AM EST - Showing real time customization of the gauges
12:22AM EST - Using Tegra K1 VCM to render a high quality digital cluster in a car, customizable using an Android app
12:21AM EST - physically accurate automotive gauges
12:20AM EST - Project Mercury: intended to make a configurable/customizable/photo realistic environment for industrial designers to use on a digital dashboard in a car
12:18AM EST - Still showing more K1 demos
12:15AM EST - Cortex A15 version of Tegra K1 will be available in the 1H of 2014, Denver version in 2H 2014
12:13AM EST - Tegra K1 VCM - automotive version of the Tegra K1
12:12AM EST - Now talking about NVIDIA automotive
12:07AM EST - Also if we're talking about chips back now, the Denver version is likely 28nm still, 2.5GHz is still quite high frequency
12:06AM EST - "I've got a few more surprises for you"
12:06AM EST - er technically a little wider than that, but it's unclear whether we're talking about ARM instructions or some internal instruction format
12:06AM EST - 7-way superscalar puts it on par with Apple's Cyclone
12:06AM EST - 7-way superscalar, up to 2.5GHz, 128KB + 64KB L1 cache (I+D?)
12:05AM EST - Running Android 4.4
12:05AM EST - YES they are running CPU-Z
12:05AM EST - Can I point out that the Denver version only uses two cores :)
12:04AM EST - Running the Denver version of K1 right now
12:03AM EST - dual-core Denver version has been back for a few days
12:03AM EST - Very high performance single and multithreaded processor
12:02AM EST - Denver is ARMv8 64-bit
12:02AM EST - Second version comes with quad-core Cortex A15s
12:02AM EST - First version comes with dual Denver CPUs
12:02AM EST - Tegra K1 comes in two versions
12:02AM EST - One surprise?
12:01AM EST - NVIDIA claims almost 3x the performance on the next-gen GFXBench 3.0 Manhattan test as Apple's A7
12:00AM EST - Now NVIDIA is comparing Tegra K1 to Apple's A7
12:00AM EST - I'm guessing that's a multithreaded SPECint estimate
12:00AM EST - Tegra K1 has more GPU horsepower in terms of GFLOPS than an Xbox 360 and PS3, and higher estimated SPECint than either console
12:00AM EST - Here we go, GFLOPS comparison
11:57PM EST - Next up is Trine 2, UE game
11:56PM EST - No word on how much power the platform is using though
11:56PM EST - This is a really good looking demo running on Tegra K1
11:54PM EST - Still running the UE4 K1 demo
11:53PM EST - "this is real time computer graphics ladies and gentlemen, on a little tiny mobile chip"
11:51PM EST - it's a tech demo from Epic
11:51PM EST - Frame rate is clearly sub 30 fps, but impressive nonetheless
11:51PM EST - live
11:51PM EST - Unreal Engine 4 running on Tegra K1 now
11:49PM EST - Showing last gen graphics on mobile, now Tegra K1
11:49PM EST - More Tegra K1 demos coming
11:48PM EST - Ok we are now 47 minutes into the keynote and all we have is a marketing name for Logan and a confirmation of CUDA core count
11:46PM EST - Still running Tegra K1 demos, showing what mobile Kepler can do
11:44PM EST - er sorry, Ira not Ari
11:43PM EST - Still watching the Ari demo running on Tegra K1
11:41PM EST - Tegra K1 demo running on what looks like a Tegra Note 7 running NVIDIA's Ari demo
11:38PM EST - One singular architecture, one singular platform and one engine, can now span PC, console and mobile
11:38PM EST - "we have reduced the time for next-gen games to reach mobile devices"
11:37PM EST - Ok so this is a graph of DX feature level over time, not really performance
11:35PM EST - Now we're seeing a graph of historical mobile, console and PC GPU performance
11:34PM EST - Tim Sweeney, Founder, Epic Games - "We can take absolutely anything that runs on PC or high-end console and run it on Tegra...I didn't think that we'd be at this level on mobile for another 3 - 4 years."
11:34PM EST - Epic Games is bringing Unreal Engine 4 to Tegra K1
11:34PM EST - Note that other than the Tegra K1 name and confirming it (Logan) would be 1 Kepler SMX, NVIDIA hasn't announced anything new yet
11:33PM EST - It appears that NVIDIA is saying Tegra will now be on the same GPU roadmap as desktop GeForce
11:32PM EST - Wow so pay close attention to the slide below
11:31PM EST - "We've brought the heart of GeForce and the soul of Tesla to the Tegra family"
11:30PM EST - It's almost inappropriate to call it Tegra 5, it's not linear, we decided to call it Tegra K1 because it's based on the Kepler GPU architecture
11:29PM EST - 192 CUDA cores, a single Kepler SMX
11:29PM EST - he's counting CUDA cores
11:29PM EST - UGH
11:29PM EST - "world's first 192 core processor"
11:29PM EST - Tegra K1
11:29PM EST - What's next for us after Tegra 4?
11:28PM EST - Cmon Jen-Hsun, announce it
11:28PM EST - we believe the car will be your most important mobile computer
11:28PM EST - Android will also disrupt the car industry
11:27PM EST - "we believe Android is going to be one of the most important platforms for game consoles in the future"
11:27PM EST - Android is going to be an important factor in televisions
11:26PM EST - "There are some new industries that Android will continue to advance into and make contributions to"
11:26PM EST - Jen-Hsun is recapping the historical Tegra roadmap
11:25PM EST - "[Android] is wonderful for the world"
11:25PM EST - "The next step...a giant leap"
11:24PM EST - I wonder what comes after Tegra 4...
11:24PM EST - Tegra 2, Tegra 3, Tegra 4 all listed on a chart
11:23PM EST - Ooh starting with Android discussions
11:23PM EST - Now we're on to chips
11:22PM EST - We're still in recap mode
11:21PM EST - G-Sync looks awesome, as always, it's a dramatic effect here because of how inconsistent the frame rate is in this particular scene
11:20PM EST - This system must be super slow, my SC2 experience is way smoother than this
11:20PM EST - Still explaining the conditions that pave the way for G-Sync to make a difference
11:18PM EST - but in this case we're looking at relatively variable frame rate
11:18PM EST - Starcraft II can be a poor example of G-Sync, at least on a high-end system, given how well modern GPUs can run the game
11:17PM EST - Running a Starcraft II demo with v-sync on, showing stuttering
11:17PM EST - I have come to enjoy NVIDIA's CES keynotes, but there's way too much recapping at these things
11:16PM EST - Jen-Hsun is explaining how G-Sync works
11:15PM EST - I want a 27" 2560 x 1440 IPS G-Sync display that ships color calibrated
11:15PM EST - G-Sync is awesome, it just needs good display support
11:15PM EST - Our review of G-Sync here: http://www.anandtech.com/show/7582/nvidia-gsync-review
11:15PM EST - Now on to a G-Sync update
11:14PM EST - NVIDIA's GeForce GRID virtualized server, streaming from France to a Shield here in Las Vegas
11:13PM EST - The demo was actually streaming from a server in France
11:12PM EST - Now we're seeing a PC streaming wirelessly to Shield and playing on the Shield's attached display
11:12PM EST - Still running a Gamestream demo, just did PC wirelessly streaming to Shield in console mode connected to a 4K TV
11:09PM EST - Gamestream demo time
11:09PM EST - Jen-Hsun is now talking about using Gamestream to stream directly to twitch
11:07PM EST - Still recapping NVIDIA's Gamestream tech
11:07PM EST - ...and we're back
11:07PM EST - Now he's sipping the water
11:06PM EST - Jen-Hsun feels like his voice is starting to go, he's searching for water, he found the water
11:06PM EST - (seriously there's a lot of illness at CES this year, this week isn't going to end well)
11:06PM EST - Now recapping Gamestream - streaming real time computer graphics from PC to Shield, Shield connected to TV or tablet
11:05PM EST - GFE is one of many tools NVIDIA has to bridge the gap between console and PC gaming
11:05PM EST - GeForce Experience leverages NVIDIA's own test data to automatically define quality settings for your installed games
11:04PM EST - Recapping GeForce and GeForce Experience
11:04PM EST - "next-generation gaming has started, so has next-generation PC gaming"
11:04PM EST - We will get sick from this
11:03PM EST - Jen-Hsun is feeling a little under the weather, he doesn't know if he's still contagious
11:03PM EST - We're going to get updates on all of them
11:03PM EST - Games, Chips and Cars
11:03PM EST - Jen-Hsun is talking about NVIDIA's focus on visual computing
11:02PM EST - Jen-Hsun is taking the stage
11:01PM EST - We're starting!
10:55PM EST - Thanks for joining us, we should be getting started here in a few minutes
10:55PM EST - and I'm (anand) on the regular text
10:55PM EST - Jarred is on snarky commentary
10:55PM EST - Brian is on photos
10:55PM EST - We're seated and connected
Post Your CommentPlease log in or sign up to comment.
View All Comments
blanarahul - Sunday, January 5, 2014 - linkHope to see some Tegra 5 coverage. Tegra 4i is dead.
Quick question: Is it possible to build a 32-bit ARMv8 CPU core i.e. a ARMv8 core capable of running a 32-bit OS without using a hypervisor? That would really ease the transition to 64-bit for Android.
tipoo - Sunday, January 5, 2014 - linkThis sounds like it may go like their last event where Jen Hsun kept going back and forth between explaining Vsync on and then Vsync off for a really long time :P
Lunyone - Sunday, January 5, 2014 - linkIs there any way to reverse the list above, so you can read it from beginning to end, instead of from the end to the beginning?
madwolfa - Sunday, January 5, 2014 - linkThat trashcan looks like Mac Pro.
ViperV990 - Monday, January 6, 2014 - linkYou guys should consider switching the live blog format to chronological. Currently when new entries appear at the top, they push everything else down, making it difficult to follow.
chizow - Monday, January 6, 2014 - linkDenver K1 looks to be a monster. All the CPU perf of Apple's A7 with unseen GPU graphics from a full Kepler SMX. Only question is WHEN. Could be a great year for Nvidia in the mobile space.
skiboysteve - Monday, January 6, 2014 - linkIn my experience add a full half year to all chip maker "shipping in" quotes because the device makers have to put it in a product before we can use it. So 2H 2014 for the A15 ver and 1H 2015 for the Denver ver.
that's what's nice about apple... None of this year early tease crap. They talk about it when you can buy it. (and btw... I'm no apple fan.. But I appreciate this aspect)
chizow - Monday, January 6, 2014 - linkI don't think it will be that long, A15 TK1 was already scheduled for Q1/Q2 from Jensen's GTC interviews, Denver TK1 may come as soon as end of Q2 if they are getting back initial samples now. Keep in mind T4 sampled last year at CES, was pushed back by Nvidia but still launched at end of Q2 in Nvidia's own Shield handheld.
I think that may be part of Nvidia's push to start producing some 1st party designs and products (Shield/Tegra Note 7), so they can get their SoCs to market sooner without having to wait for 3rd party designs similar to your Apple anecdote.
theCuriousTask - Monday, January 6, 2014 - linkEven if we are optimistic on Nvidia's ability to execute, what does that say that it is coming out on 28nm instead of 20nm? It seems 20nm is pretty delayed and may affect the rollout of 64 bit krait/cyclone2.0 products in the future.
B3an - Monday, January 6, 2014 - linkWheres it say that? Only the sample they have now is 28nm. Consumer devices could be 20nm.