Wait (tm) Vega

wait (tm) Vega

Other urls found in this thread:

youtube.com/watch?v=iDuAugAykcc
tweaktown.com/news/58635/amd-radeon-rx-vega-56-leaked-benchmarks-gtx-1070-killer/index.html
research.nvidia.com/publication/2017-06_MCM-GPU:-Multi-Chip-Module-GPUs
twitter.com/SFWRedditImages

Kill (You) yourself.

No, wait (R)

Just Wait

soon

minerfags watch your backs I'm going to jack ur ass @ gunpoint

wait™®℠©

Wait(TM) to see if RT fixed the memory bandwidth issues with the HBM2 on Vega. Increased ETH hashrate suggests they have, and Vega was as memory bound in gaming as it was in ETH mining...

*RTG

For what? Every bit of speculation and leaked data so far has it pegged as a less energy efficient 1080. I'm an ATi fanboy and even I'm not holding my breath for Ryzen lightning to strike twice. :(

still waiting

Any updates/info on preorders?

Vega 56 is a day one buy for me, AIB version of course.

jokes on you
i already have a 1070

To find out whether they fixed the memory bandwidth issues with the HBM2, because to the extent that the ~20-30% bandwidth regression vs. Fiji gets fixed, to the same extent Vega will get faster in gaming and ETH mining.

...

Learn English, you stupid cuntwaffle.

wait®

1070 Performance

Well, I am waiting...

Repeat after me

1070 Performance

1070 performance

youtube.com/watch?v=iDuAugAykcc

>quotes the whole thread
>calls someone autistic

Leaked reports on Tweak Town are showing the Vega 56 is a decent margin faster than the 1070:

tweaktown.com/news/58635/amd-radeon-rx-vega-56-leaked-benchmarks-gtx-1070-killer/index.html

wait™ for actual benchmark tests

The entire thread is autistic shitposting. I had no choice.

>I had no choice.
spoken like a true shill on duty

JUST WAIT FOR NAVI ITLL FIX EVERYTHING

You can shitpost that, but even NVidia knows MCM GPUs are the next big step. research.nvidia.com/publication/2017-06_MCM-GPU:-Multi-Chip-Module-GPUs

MCM != GCN being less of a piece of shit. IIRC Vega already has this partially implemented.

W A I T
A
T

you mean the dual gpu version?

literally a 600watt firehouse

we wont get anywhere until they do 2-4 cores on a single die linked with infinity fabric

i dont think you ever read what the hbcc does eh

thats because it should be clear by now that this time around they flipped the whole gaming segment and focused entirely on compute

99% of the new ISA is for compute the cards perform like a motherfucker beating the quadro p6000 at ease they are releasing the ssg in a month the 9100s in 2 and im not suprised to be honest vega is the last card of gcn as we know it it makes sense to just remove it from the line like that

Thing like Creo and Solidworks and other shit are not compute. Vega is all about pushing pixels.
Also fix your English Pajeet.

a gpu is made to push pixels
macaco 2017

last 5 years that's not what they and nvidia focused on

cock ring

hügck

>doesn't know how to ™
wew

I know what it does, but it cannot solve for raw bandwidth issues that were evident on Vega FE. To my understanding that can only be solved by figuring out what's preventing the HBM2 from running much closer to its maximum bandwidth and fixing it.