This is what Nvidia GPUs are being used for right now
how can AMD Pajeets even compete?
At least name the image
>nvidia tracks amd users
If you want to shitpost.
AMD's target are poofag /v/ kiddies.
>/g/ calls Nvidia the /v/ brand
>Nvidia responsible for all the GPU related progress in the world including self driving cars
>Meanwhile AMD makes cheap gaymer shit only
>>56378589
When you don't have working OpenGL, all you can hope is DX and gaymurs.
http://vrworld.com/2016/05/23/amd-working-vr-bollywood-movie/
Like this
AMD pajeets are the ones in the poverty levels meng
>>56378548
>console warz
>>>/v/eddit
>>56378637
>>56378584
>>56378589
AMD and nvidia defenders are as bad as each other and you're all /v/ cancer.
false flagging is precisely why I will forever boycott nvidia
>>56378720
boycott them all you want, they have bigger customers doing better things with technology
at least you can support better Bollywood movies with AMD
Nvidia twitter
>we just helped scientists discover amazing things
>were making self driving card and advanced AIs and shit
AMD Twitter
>Were helping girlllzz make games!
>please buy our products!
>>56378861
DELETE THIS
>>56378548
Except that AMD is much better for computing.
>>56378757
>look at how racist i am guise can i join you're sekrit club now?
>>56379575
>>>/reddit/ is more your speed
>>56378548
Thank you NVIDIA
Nvidia doesn't need deep learning for that, they just need to locate the houses which contain AMD cards.
>>56381252
dumb invidiot
>GPUs are helping scientists to predict who has light bulbs by looking at a satellite image of where light bulbs are
Western civilization is fucking dying and all you morons care about is a brand fanboy war.
>>56379487
Bitcoin mining is not the only compute task GPUs are used for you retard.
>>56381252
That was great, anon. Thanks. Here, have a thing:
https://www.youtube.com/watch?v=X5cW3WeKF00
>>56378548
>nolightsugal
>>56379487
No. There are virtually no application or framework for scientific applications that even run on AMD hardware in the first place because everything's written for cuda.
>>56378589
>Not knowing who pushed for HBM
>ITT
>>56381294
fuck off /pol/
>>56382710
Use your time for writing working ayyymd OGL drivers, pajeet.
Jesus christ, University really is a meme. Why do you need satellites to work out which countries are poor? Note: poor/dumb has a strong correlation with the density of shit colored skin.
>>56382777
Nice post Rajeet
>>56382777
>>>/reddit/
>>56382750
#triggered
>>56382651
Are you advocating vendor lock in?
Am I still on /g/?
>>56383370
Pretty sure he's just stating a fact.
>Nvidia is discovering how the universe works
Meanwhile, Pajeet is using his AMD card's compute power to mine buttcoins
>>56378548
the irony here is that, since "deep learning" doesn't really need high specs (fast double-precision FP), nvidia hw (and CUDA) shines.
where raw power is needed, AMD is THE alternative, and then, there is customized hw.
>>56383370
>recognizing reality is advocating for existing practices
Let go of the straw.
>>56382651
>what is OpenCL
Even phones start to support that.
>>56378548
Dear Sir,
I am going to asking you to go ahead and delete this post or we will have to take further legal action or fine you in the amount of 10.000 $ (ten thousands) payable via moneygram.
Best regarding,
Mike Davis,
Microsoft support manager, Panjaab province
>>56383520
>>56384510
And then using that fact to say that AMD is not stronger in compute tasks, yes. Vendor lock in is totally unrelated to the strength of the cards. If anyone's got straw around here, he does.
>>56383855
>Pajeets is using his AMD card's compute power to simulate designated shitting streets
ftfy
>>56384666
What use is this "strength" if it is not utilised by anything other than alt coin miners.
>>56382651
There's a reason they don't mine cryptocoins on CUDA
>Using GPUs to map out poverty levels.
Wasted electricity.
They should just dig up a list of everyone who buys AMD and put those people on the map.
You'd instantly have the poverty concentrations right there.
>>56384906
It means AMD is too lazy to actually support software and expects everyone else to do it.
Which is a failure of AMD's management and budget rather than their hardware.
They also have less money to straight up bribe software manufacturers to support CUDA.
>>56384945
The problem is even Intel and Nvidia make better budget options
its not poverty that leads you to actually buy AMD, its stupidity and blind fanboyism
>>56384963
I know what it means, that doesn't change the question. Who cares if it's not used?
You know, I actually want to like AMD, but every time I'm tempted to buy their product, they do whatever it takes to discourage me. Be it their communist themed advertising or retarded software support.
>>5384945
Or it might be that they are just testing what they have with data that is already known to show that it works.
Their end goal might be to map all solar systems/galaxys/planets that are visible due light to find constelations with high probability of life, but you wouldnt start this project by straight going to this task for obvious reasons.
>>56384934
Theres a reason every deep learning framework uses CUDA
Hashing using different shit you have no idea wtf you're talking about
AMD cards could do about equal for deep learning except they can't because AMD is literally cucking themselves and doesn't even know ehat deep lerning is they're fucking living under a rock
they recent released a GPU with a sorta SSD onboard as a pseudo-RAM which would be FUCKING GREAT FOR ME I COULD FINALLY DO MODELS OF THE SIZE I FUCKING WANT BUT NO
FUCKING AMD FUCKING THINKS ITS BEST FOR THE OIL INDUSTRY FOR CRUNCHING NUMBERS FOR DRILLING OR WHATEVER THE FUCK
REEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE
>>56378589
>Nvidia responsible for all the GPU related progress in the world including self driving cars
toplmao
>>56385186
>every deep learning framework uses CUDA
Except it doesn't, unless you believe everything means Nvidia's libraries
Every deep learning framework has OpenCL support or has support being developed
Right now Wolfram Mathematica runs on OpenCL along with CUDA, which is easily the most important deep learning package right now
>b-but Nvidia's twitter told me CUDA was the only important thing!
wew lad
>AMD cards could do about equal for deep learning except they can't because AMD is literally cucking themselves and doesn't even know ehat deep lerning is they're fucking living under a rock
ATi was pushing for GPGPU's before Nvidia, AMD FireStream was launched months before CUDA, but Nvidia didn't just discontinue their API while pursuing open API's like AMD, so CUDA grew pretty stable way before OpenCL was around
>>56385469
face it, CUDA fucking owns the game right now
if you aren't running CuDNN and CnMem then you're like a fucking decade behind
>>56385549
>if you aren't running CuDNN and CnMem then you're like a fucking decade behind
>if you aren't paid by Nvidia you're a decade behind
>>56385469
>c-c-cuda aint shit
>amd has um o-opencl um and they haven't had as much time so um cut them sum slak ok!!1!
L E L
E
L
>mfw CUDA is 20x (TWENTY TIMES) faster than amd
2 0 x
0
x
T W E N T Y T I M E S F A S T E R
W
E
N
T
Y
T
I
M
E
S
F
A
S
T
E
R
>>56385567
>implying i'm paid by nvidia
id fucking love that but no, I just am not retarded, and I care about what I do, and nvidia seems to at least throw money at it, so yea I like them a whole lot more than AMD.
And of course AMD doesn't mean to be so shitty, its just that they are, they just are so incredibly dumb that they don't know what they're missing. It pretty much is too late, they better fucking pull a miricle or nvidia is gonna swamp them (as they kinda already are)
>>56385587
>just shitposting
How much do you get paid? How can I contact your representatives? I have a bunch of pajeets who usually shill on reddit for Clinton but they ain't paying much anymore