nVidia Volta - May 2017

File: purple_haired_slut.gif (2MB, 1294x1338px) Image search: [Google]

2MB, 1294x1338px

AMD BTFO

Intel must be really scared of amd then huh it's about 4 generations ahead.

kys

>>55745073
Kek

>>55745061
>scared
just like they got scared of 480

File: slack-imgs_480.png (195KB, 480x288px) Image search: [Google]

195KB, 480x288px

Nvidia mesters leave

>>55745070
>>55745073
>>55745176
>>55745190
kys nvidiots

Reminder that Volta was supposed to be on 10nm.
Moving it to 2017 means it's still on 16nm and only a minor improvement over Pascal.

Nvidia didn't rush anything, AMD is not relevant at all in GPUs and doesn't even support Feature Level 12_1

Volta is to fulfill the contract to the US govt

http://www.anandtech.com/show/8727/nvidia-ibm-supercomputers

FUUUUUUUUUUUUU

I just ordered GTX1060

File: amd-is-dead.png (132KB, 792x432px) Image search: [Google]

132KB, 792x432px

>>55745061
>nvidia
>scared

TOP KEK

>>55745286
>intel gives them money under the table just to avoid antitrust laws
amd hasn't been relevant since 2006. They had the edge only because intel decided to develop netburst.

>>55745272
Why are you complaining, then?

Just sell it in a year for 60-80% of what you got it for and get this thing.

>>55745341

true, at least im not a >just waitfag

>>55745264
so Volta is only for HPC market while Pascal will have HPC and consumer market? A chance od Titan and maybe 1080Ti HBM2 Volta anyway?

File: 1274581975691.gif (3MB, 445x247px) Image search: [Google]

3MB, 445x247px

>>55745061
>nVidia to finally address DX12/Vulkan async deficiency
>actually believing they won't keep pushing their preemption meme

>>55745061
I was wandering - how they will handle thermal dissipation with all those 3D chips? They staining with it with regular flat chips now, how they will deal with even harder task?

>>55745073
Top lel

>>55745419
Those chips run at significantly lower frequencies.

>>55745286
>18%
man they should be scared at this rate
http://wccftech.com/amd-takes-gpu-share-nvidia-q1-2016/

>get a Zotac 980ti AMP! Extreme
>become kinda sad when the 1080 comes out 8 months later
>it's beaten by the Titan X weeks later
>Volta now coming in May possibly

Whew, looks like this bad boy will get me through the next generation no problem at this rate

>>55745419
by not making AMD tier thermonuclear warheads

>>55745443
>pajeettech
why do you even pretend this shit is credible when you will poo poo it when it's positive news about nvidia?

>>55745440
> Those chips run at significantly lower frequencies.
How that will not harm a performance? Sorry I'm just that illiterate. Maybe you can direct me to some useful resource about those?

Still not worthy of upgrading from 980ti. Oh bother.

>>55745466
because it's not a 'leak' or 'rumor' you retarded child go back to /v/

>>55745061
>vega and polaris are under performing
>nvidia is just accelerating more nails in AMDs coffin
I'm still buying a 1060 because it is completely sufficient and waiting is for fags.

>>55745468
Clock speed does not directly relate to performance, in case you haven't noticed.

Especially when it comes to different architectures.

>>55745545
>raping goyvidia in dx12/vulkan
>buying outdated tech

cuck

what the fuck I just bought 1070

>>55745836
your card is gonna be what the 970 is now by next year.

>>55745849
good, just unexpected. I haven't sold my 970 yet lol

>>55745400
>preemption
Someone explain this shit to me please

vulkan,dx12,async..... how come nVidia still has faster cards then??


EXPLAIN !!!!

>>55745061
>Fudzilla source
lol

>>55745886
AMD suck hard at making drivers. They have excelent hardware but the coders are literally monkeys

>>55745657
>raping nvidia in a benchmark and a single game that will receive nvidia optimizations within a month
>still drawing too much power from the pci-e socket
>playing any games made within the last year anyway

>>55745886
AMD simply has terrible DX11 and OpenGL perfomance, that's why they see huge gains when moving to DX12 and Vulkan thanks to close to metal development and async compute

>>55745886
Driver overhead.

Game devs will not use heavy async compute that makes Nvidia cards (80% market share remember) slower. They will follow the recommended DX12 implementation as outlined and used by 3DMark in Time Lie. Mark my words. Apart from a few outliers using Vulkan, performance in DX12 titles will remain roughly the same for AMD and Nvidia. Game devs want to sell games and penalizing the 80% of Nvidia owners would be stupid.

>>55747130
>i still think async compute is a magical feature that will boost game performance by 50% but only for AMD cards because they have the best implementation!