nvidia @ twitch tv

>inb4 muh self driving cars muh deep learning muh cuda and nothing the consumer centric /g/ cares about

This is how it's going to go

>muh AI
>Deep learning
>Something about how GPU's save lives/used in medical applications
>unmanned vehicles

Don't despair, at least AMD is likely to actually release some new hardware. same time tomorrow

volta

a second stream

https://videocardz.com/69989/watch-nvidias-ceo-keynote-at-computex-2017-here

>>60651957
You're forgetting all the mandatory botnet in GPU drivers they're adding.

File: wheresmymic.gif (2MB, 320x240px) Image search: [Google]

2MB, 320x240px

>>60651897
maymays.

File: 1440969217941.jpg (20KB, 945x388px) Image search: [Google]

20KB, 945x388px

>>60652040
ebin

File: BLACKED_100461-NICOLE_2160P.mp4_snapshot_03.48_[2017.05.29_22.52.53].jpg (93KB, 1600x900px) Image search: [Google]

93KB, 1600x900px

10 minutes

>>60652793
30*

>>60651897

Lol@AI Forum a whole lotta shit no one cares about.

Jensen is such an insecure bitch.

>>60652019

It's deep learning to their AI. To identify your PC usage habits. Geforce experience is literally spyware.

>>60652934
>/g/ - Technology

>>>/v/

File: happy-wojak-women-s-t-shirt.jpg (44KB, 800x800px) Image search: [Google]

44KB, 800x800px

>>60652793
>nicole aniston is blacked

>>60652960

Lol. AI is meme.

Until they get serious with asics all the GPU "Tensor cores" are a joke.

>>60652996
You said the same shit 3 year ago about deep learning

are they still making those spy eyes you plug into every room in your house?

>>60653018
People who underestimate "deep learning" and "strong AI" are retards who's only experience with shit even remotely close this are fucking markov chain bots so ofc they think it's not srs business, but the reality is you're all fucked. Google just recently fired all the programmers working on DeepMind because it can program itself better than they can.

Theyre live

File: 1483618349400.png (49KB, 808x805px) Image search: [Google]

49KB, 808x805px

ITS TEIM FGTS GET IN HERE

Twitch/nvidia
Twitch/nvidia
Twitch/nvidia

BUUGGEN

twitch tv nvidia

>>60653126

LMAO.

Yeah, we are all impressed with it's Go championship. Deep Blue won jeopardy too. Who then had to have its memory wiped after it learned how to swear and kept cussing at scientists when it absorbed Urban Dictionary.

Real cutting edge stuff. Maybe in 50 years we can have an AI with the functional intelligence of a small mammal.

File: 1445234321804.jpg (46KB, 640x473px) Image search: [Google]

46KB, 640x473px

>>60651957
I'm 3 outa 4. C'mon automotive industry, don;t meme me now

>>60653126
>Google fired it's programmers working on DeepMind

The fucking cancer of clickbait headlines

No they fucking did not you retard

>>60653401

Self driving cars!

File: 1301186505416.jpg (33KB, 500x328px) Image search: [Google]

33KB, 500x328px

Woodscrews when?

>>60653532
go bed Raja

>>60653540

Pascal passed off as Volta when?

>>60653585

No discussion of how bad the 1080 throttles? Too bad.

>>60653554
Is Vega will beat the 1080 and the 1080ti?

File: screenshot-www.hardware.fr-2017-05-30-00-06-42.png (14KB, 635x313px) Image search: [Google]

14KB, 635x313px

>>60653600
You live in 2014

ASUS LAPTOP COMES WITH TV LUL

>amdfags: I don't care about the power draw

AMD ARE HOUSEFIRE, THATS WHY THEY SUCK ON LAPTOPS. RADEON BTFO

>>60653622
But does it come with Ryzen Mobile?

>>60653631
NVIDIA invented housefires.

>>60653631
Macs have radeons.

>>60653649
>facebook machine

File: advanced_thermal_solutions.png (73KB, 237x221px) Image search: [Google]

73KB, 237x221px

>>60651897

Unless they manage to BTFO AMD in the next few months and the GTX 1080 TI and XP who the fuck cares.

>next-gen efficiency
>no HBM2
Lmao gookman of woodscrews.

>>60653696
>hbmeme

>meme learning
Not this shit again.
>>60653706
Is used by everyone, including Intel and NVIDIA. Where's HMC?

File: 1484548311184.png (252KB, 595x654px) Image search: [Google]

252KB, 595x654px

>>60653706
DELET

lel banned for saying CHING CHONG when he was speaking chinese

>>60653696
Nvidia are the 1st who use HBM2, how ironic amdrone
>>60653720
Expensive af

>>60653759
>Nvidia are the 1st who use HBM2, how ironic amdrone
Too bad AMD has patents for it. Just like patents for GDDR3/5.
>Expensive af
Where's HMC, porkshoulders-kun?

>>60653768
HBMx for pro
GDDR6 for gaymen

Duuuude AI lmao

where is my volta

>>60653782
No. Not even close. HBMeme is for everything. It takes less space and it more efficient.

>>60653796
>where is my volta
GV100. :^)

>>60653797
$$$$$$

Yay! We bolted on asics!

It's wood screws!

S K Y N E T

All I'm hearing is
>cloud cloud cloud
>skynet
>cloud cloud cloud
>cloud skynet cloud
>skynet skynet, cloud + skynet
>skynet skynet skynet
>cloud moar, skynet good.

1 petraflop card, wew lad.

AMD on suicide watch.

>>60653959
It's a server node, not a single card.

>>60653959

Wrong. A petaflop node.

Something AMD has already beaten.

http://semiaccurate.com/2016/12/12/amd-shatters-petaflop-rack-barrier-instinct/

>>60653983
>Something AMD has already beaten.
H-hayai.

>>60653983
>AMD and Inventec were pretty tight about the details of the box

Even semiaccurate aka redteam are not fooled

>>60653995
please say that to me while I pound you

>>60653983
>http://semiaccurate.com
DO YOU KNOW WHO IS CHARLIE DEMERJIAN?

File: 1471115989420.png (933KB, 903x1027px) Image search: [Google]

933KB, 903x1027px

>>60654001
>semiaccurate
>redteam

>>60654020
Read this

https://twitter.com/cdemerjian

>this music
that's it

im selling my nvidia stock

>>60654027
>Like I said, Xpoint is broken, endurance isn't there, and Intel should not have released it
But he's right.

>self-crashing memes
>meme mearning
Maybe they'll talk about something that actually gives NVIDIA it's revenue?

This is some interesting stuff

>>60654104
they stop the video before that drone hit that green guy

isn't this the same presentation from a few weeks ago. I have the feeling I have already seen this

>>60654147
Every NVIDIA keynote is the same since forever.

hhahaha good one jensen

File: stdin_snapshot_50.47_[2017.05.30_00.01.51].jpg (143KB, 1920x1080px) Image search: [Google]

143KB, 1920x1080px

>Activate Windows

>>60654190
lol

>>60654101
Deep learning workstations sell for $130k a pop and Facebook/Microsoft/Amazon buy Nvidia FP16 deep learning GPUs in bulk. So yeah, they do get revenue from that.

>>60654241
It's almost nothing compared to something that actually gives NVIDIA it's revenue.

File: 1494046109724.jpg (58KB, 500x500px) Image search: [Google]

58KB, 500x500px

Wow robots playing golf, nvidia wins again.
AMD is about to destroy two big companies in the same year.

Has anything interesting happened?

>>60654260
No. It's NVIDIA keynote ffs.

File: nvidia-q4-f2017-divisions.jpg (61KB, 807x389px) Image search: [Google]

61KB, 807x389px

>>60654241
Gaming is bulk of their businesses

>>60654190
Can't M$ sue them for this?

>>60654283
It's really sad how much gaymers will overpay for shit like the 1070 and 1080.

CIOs are less gullible.

Is AMD going to present new stuff soon too? I have hopes that their framework for neural networks on their GPu will be decent and with less botnet.

>>60654490
22 hours from now.
It's expected to be Threadripper and Vega RX announcements.

>>60654490

Tomorrow night.

>>60654505
>>60654512
Alright. I still have time to return the nvidia GPU I bought yesterday. You better deliver amd.

>>60651897
any links to it? I wanna watch it

>>60654489
Not really. Nvidia jews them more. Tesla P40 has 12 TFLOPS single precision performance and 47 TOPS INT8 performance and is sold for $5,699. 1080 Ti has 11.3 TFLOPs single precision and 45.2 TOPS INT8 and is sold for $699. Both are dual-Slot, full-height and has TDP of 250W, and incapable of half precision inference. And P40 has passive cooling meaning it can only be used in rack units

>>60654490
AMD has no chance of competing in deep learning anymore. Anything they put out will be too little too late. They don't even have something akin to cuDNN.

>>60654578
Hardware is what matters.

>>60654649
Not if there's no one to develop on them. Non of the big deep learning libraries has high-profile OpenCL branches. It it has (like Caffe and Theano's) it's being put together by enthusiasts and are no longer supported/updated, while things like Tensorflow and (Py)torch have backing of big companies (Google and Facebook respectively)

>>60654562
But they don't buy them, hence the revenue numbers.

>>60654690
For now. Major ML libraries will be ported to OpenCL as the part of MIopen.

>>60654692
Well they certainly aren't putting consumer cards in their server farms saves for some cloud gpu computing startups

>>60654578

AMD doesn't have the money to gonfor deel learnjng and AI. They realy don't give a shit as ASICS are the beginning of the endgame for AI. and since they will go after cost effective solutions that will have better than 2% silicon yields they are focusing on content creation like movies and games and modeling.

>>60654742
Vega is deep learning capable, though it's still general purpose.

>>60654450
You need to be 18 or older to post on 4chan

fuck brand loyalty turned this board into fucking /v/

"nvidia wins"

This statement is brought to you by nvidia.com from paid actors.

>Announces new GPU
>It's called the Titan XP
>Not to be confused with the Titan Xp or the Titan X Pascal or the Titan X or the Titan.
>It's 10% faster than a Titan Xp
>It only costs $2000
>You can only buy it from Nvidia special order and it will take 3 months to get to you by which time they will have the NEW and IMPROVED Titan xP!

>>60652950
>Geforce experience
Botnet is built into driver now, you don't actually need experience for it. I block telemetry processes on firewall, but it's bullshit that I have to. Never buying jewvidya again.

>>60655329
YOU need to be 18 or older to post on 4chan.

>>60656008
Nvidia still has a big advantage over and due to cuda. Sucks really hard man.

>>60651957
>AMD is likely to actually release some new hardware. same time tomorrow
not really, end of june was confirmed by raja, and only enterprise cards

>>60651917
>>60651957
I'm interested, but there's probably nothing new here that wasn't already presented on GTC.

>>60652996
>Until they get serious with asics all the GPU "Tensor cores" are a joke.
Found the butthurt Intel shill.

>>60656008
You can remove telemetry before installing so you don't need to firewall them because they don't get installed in the first place.