With NVIDIA, now you dont need driver!
Other urls found in this thread:
barrons.com
youtube.com
youtube.com
youtube.com
cnbc.com
en.wikipedia.org
twitter.com
What does this computer has to be better for driveless cards than any other real time inferencing problem?
>What does this computer has to be better for driveless cards than any other real time inferencing problem?
Glorious nVidia™ ecosystem.
The thing doesn't even exist in reality now anyway.
>meanwhile amd is making chips for tesla that has marketable self driving features right now
wow its nothing
Why do you fags create a thread for every wccftech article?
Nigger I already have this in my model X and it still won't get me my coffee from the shop down the street.
so AMD killed a man?
>greater than 1TB/sec
what did they mean by this?
>capable of Level 5 Autonomy
and what is this shit?
what is the highest level?
>self crashing meme
pls no
It's probably only something like 3.5GB/s
barrons.com
>Comprehensive view is that TSLA and NVDA have currently a 5-year contract in place, which was renewed just recently. AMD is not a Player. AMD is DoD (Dead-on-Departure) in DML (Deep Machine Learning) Workloads. Google threw AMD out of its GPU Training Cluster, as AMD had extremely poor performance on GOOGL TensorFlow framework. Later, as a courtesy to AMD, GOOGL redeployed the AMD GPU's for only VDI (Virtual Desktop Infrastructure). So far we have attended no less than 60 DML (Deep Machine Learning) conferences...we have not seen even a single benchmark of any DML Framework running on AMD GPU's for production workloads. Investors optimism is completely misplaced that AMD will become a significant DML player. There is only going to be one GPU player NVDA, just like there is only one CPU Player INTC ....rest all the players will be in the others category, which will be about 10% of the market.
AYYMD is not making anything for Tesla, their garbage architecture and garbage software cannot do driverless cars
What are those ports?
I bet musk is regretting that because this one is level 5 and more power efficient..
It's 500watts and they don't exist yet.
At first glance i thought this meant we wont need graphics drivers.
>just like there is only one CPU Player INTC ....rest all the players will be in the others category, which will be about 10% of the market.
All this while EPYC killed the entire Scalable Family.
more like idiots who had level 2 automation but for some reason though their car was level 5.
NO!
the fuck is TOPS
>LEVEL 5: DRIVERLESS ROBOTAXIS
these buzzwords are starting to get a bit too autistic
Probably has required processing power for enough spatial awareness not to crash into anything in situations that may be challenging for regular navigation computers.
Likely the total data throughput the system can handle from all of the vehicle's sensors.
They're for all of the various sensors, cameras and radars required of self-driving cars, as well as power delivery and interface with the ECU.
The fact that this isn't semi-custom just proves in my mind that Nvidia isn't going to stick around in AI for much longer
Level 4 can drive safely in city traffic at speeds 35mph or lower without your hands on the wheel at all. Audi has the first Level 4 car out there.
Level 5 is basically full on AI chauffeur.
>now you dont need a driver
That's what novidiots with linux say
That's what Tesla would say :)
oh fuck i read it as ROBOT-AXIS not ROBO-TAXIS
Sounds like stock manipulation
Who is the theorycist behind this spectacular deployment of knowledge?
>tops
What the fuck is this? Teraflops?
Trillion INT operations per second?
Inferencing is INT and not FP.
Yes.
It's the retards that unironically cited Intel's FUD slide deck.
>electric cars already get shit mileage
>hey let's add a fucking 500W computer
B-but it makes it self-crashing!
Now you can die in a car crash without going full retard!
Just think of the possibilities!
THANK YOU BASED NVIDIA
NOW I CAN DIE IN A CAR CRASH WITHOUT BEING INVOLVED
>For the deep learning aspects, Nvidia provides a separate classification of performance, called Deep Learning Operations, or DL TOPS (Tera OPS, i.e., trillions of Operations Per Second).
500W is nothing for a car, just your ac takes several times that much.
>implying
Nvidia already shat their pants after instinct got released
Bullshit. Electric cars already struggle to get halfway decent ranges, and now you want to draw an additional 500W? Yeah, that's totally nothing.
Many thanks!
MI25 is laughable for ML/DL though.
But yes, it's SR-IOV implementation makes GRID irrelevant.
car engines are on the order of hundreds of kilowatts
how seething mad does a company have to be at a meme image to make a new product line to try looking like the opposite viewpoint
Electric cars have no engine
you sound like a butthurt taxi driver,commercial truck driver or uber driver that is threaten by this. nvidia said it would receive constant software updates to improve the system learning abilities. also the hardware is only getting better each year by improving the power consumption and the performance.
>muh electric cars
they still have motors
>Tesla Model X P90D
>259 hp (193 kW) front, 503 hp (375 kW) rear
500W is not a big deal for something as major as a chauffeur replacement
NVidia will say literally anything to propel their fucking bubble stock forward.
Absolutely everything can happen on the fucking road and no meme machine will ever be able to handle it 100% correctly every fucking time.
human drivers crash too
AI can take data from millions of human drivers and use it to learn how to drive better than the best human
Only because amuricans have 2 words for something that converts some type of power to mechanical power. In most languages the word is the same for both electric and combustion ~engines~.
Human drivers can prevent crashes too.
No meme AI will be able to do that though since car crashes are inherently unpredictable.
>inherently unpredictable
No, they're not. They're usually due to someone not paying attention to what someone else is doing.
AI being able to monitor 360 degrees around the car constantly with no distractions minimizes crashes in a way human beings never can.
Not to mention that the more widespread driverless cars become, they can communicate with each other and all cars in a lane will know the intent, velocity, and proximity to every other car.
I'd say 500W is still worth consideration given the car only has a 100kWh battery. Not to mention that when the battery dies it's not just a matter of spending a few minutes at a gas pump.
>TOPS for AI inferencing
>AI Inferencing
>ASIL D Functional Safety
>Functional Safety
>MEMORY BANDWIDTH
is this jargon real? or is this just bullshit to sell a peace of plastic?
500W is at full load, and it's a lot more efficient at 320 TOPS vs 24 TOPS on the drive px 2, it'll consume less power at least when it's just sitting at a stop light or doing tesla-tier cruise control
This is nothing is a car
>unironically blasting these in public
kill yourself
>Vastly more complex than traditional board games like chess & go
lol
tesla has always used nvidia chips for autopilot moron
In an electric car it is. You've only got limited battery, there's no alternator constantly generating power, you have what you have and that's it.
>going 80 down a backroad with like one house per mile
>in public
lol vega keeps on failing
nvidia gpu master race
tesla already use drive px 2, 250W TDP, it's fine
Not as long as our current battery tech's as shitty as it is.
> Not to mention that the more widespread driverless cars become, they can communicate with each other and all cars in a lane will know the intent, velocity, and proximity to every other car.
That's the future of driverless cars.
i'm not the one who insists on the electric car meme
Nor am I, but it's sadly here.
That's UK fags not Americans. Engine is specific. Motor means both.. Just say motor
American here, I've learned it and always known it as motors are electric and engines aren't.
en.wikipedia.org
you're an idiot
Unless you put tape on traffic signs bringing the system to its knees
>you're an idiot
nope, just average amdrone
>In most languages the word is the same for both electric and combustion ~engines~.
Except that the original meaning for "engine" was a mechanical contrivance. As in Difference Engine, Ruling Engine or Siege Engine. A Motor is just one of the class Engine.
Wait, this thread isn't about videogames..