http://wccftech.com/nvidia-pascal-trouble/
Why did you all convinced me to wait for pascal?
Why do you have to be so cruel
>>52372186
I'm so glad that memes are back.
Silly OP Not buying superior AMD GCN architecture from back in the HD series and totally missing out on good game play.
Can't shareholders sue over this?
>>52372241
Why would they care? They know that Nvidia's retarded fanbase will forgive anything, so they're free to fuck up as much as they like.
>>52372240
im buying a 270x just to get used to amd, i head you can play cs go at 4k 80 ish fps with the reference one
Those looks like replaceable card, and is only a showpiece.
It doesn't mean much.
So can I just go ahead with a 980ti build? Fucking sick of waiting
>>52372467
This isn't acceptable work, Amir. Don't use weak words aka "looks like" or "doesn't mean much". Make positive statements that everything is fine, and that Pascal is not only on track, but set to crush the competition.
Consider this a first strike.
>>52372484
Yes goy
>How Nvidia made the big Pascal chip in such a way that both GDDR5 and interposer based HBM memory will work with the same package is going to be very interesting to hear, every other company using HBM has found this impossible to implement too.
Very interesting. How they managed to do something, that even the inventors of HBM could not do is certainly a technological feet in itself. Yet alone the 15TF of compute on a GDDR5 based chip.
Oh. Now this is golden.
Here's a better read though: http://semiaccurate.com/2016/01/11/nvidia-pascal-over-a-year-ahead-of-1416nm-competition/
I can see the lawsuits and stock dives now.
>>52372467
>It doesn't mean much.
Except it does.
Because doing exactly what this Gook of a CEO has done (again), is Illegal.
It also means that nVidia are so threatened by AMD's advances and recent IP and Patents that they are willing to go to the lengths of faking products, again.
>>52372186
So, AMD can't driver
>>52372601
Is HBM and ggdr5 the 3.5 meme revision 2?
Like:
>Buy 4GB HBM card
>Its actually 3.5 gigs of HMB and .5 of gddr5
>>52372720
I can't wait to sell my 970 for $260 when there's new $350 GPU's.
>>52372752
Keep dreaming, cucklord. They already go for less than that.
>>52372800
Anyone is wiling to pay anything.
>>52372716
kek, your self driving car has a watercooled GPU, what could go wrong?
>>52372716
>250w
holy fuck that thing is thirsty. Isn't Tesla using a regular tegra right now, and it's doing pretty well.
>>52373277
I am more interested in what google is using
>>52372467
this post looks like Patel speak
It doesn't mean much
>>52373277
Dude your car's getting a neural network
So the federal govt can track every car on the road in real time and know what your tire pressure and oil temp are at the same time!
IoT!!!
>>52372402
Had one. It was garbage from day one
>>52372218
Yes! I actually lost some of my AMD brand loyalty but I'm glad we are talking about woodscrews and low yields again. I like my Intel/Nvidia laptop though and it runs really cool.
>>52372716
self driving car, wait...so it has no...driver?
>>52373277
It's dual-gpu.
Personally, I'm waiting for the high end Polaris cards. But then again, I have a vested interest in seeing AMD recover.
AMD Fury X master race reporting in
Stupid nvidicucks were memeing to just wait 2 years and buy pascal instead of upgrading
KEKED
>>52372601
>>How Nvidia made the big Pascal chip in such a way that both GDDR5 and interposer based HBM memory will work with the same package is going to be very interesting to hear, every other company using HBM has found this impossible to implement too.
how the fuck could they possibly work on the same package?
are they planning on requiring an interposer ever on GDDR5 solutions?
>>52375962
sorry, buddy.
R9 3xx and GTX 9xx series will be forgotten as the last in a frustratingly long string of 28nm GPUs.
Polaris and Pascal have the potential to have enormous gains and endurance, since we're probably not seeing a jump to 10nm or better for a long time.
>>52372186
that guy has a hot grill friend
I hope Polaris has conservative raster. It's basically needed to do any voxel-based global illumination.
>>52376861
how many of the 12_1 features aren't just HW accelerated/managed versions of techniques already available in DX12 or even DX11?
for the GPU-sorted transparency, tiled 3D textures, etc., you really need techniques that aren't feasible in DX11 but are with HW acceleration.
some of Nvidia's 12_1 demos still look like they were making 980Ti cards sweat for pretty low intensity scenes.
>>52376861
voxel is useless
