How can we make RTG/AMD Great Again????

>NVIDIA Quadro RTX with 96 GB of GDDR6

Attached: NVIDIA-Quadro-RTX.png (761x452, 210K)

Other urls found in this thread:

>How can we make RTG/AMD Great Again
Do any anons on this board even work for AMD engineering department?
I thought the only AMD employees here were on the marketing team.
Anyways just wait for it and it will beat nvidia for sure.

nice OP
you already started wrong.
it's 48+48 as in linked togheter. In SLI.

10000 usd
it's not even hbm2

Just wait ™
Worked every turn

By making it so they can use completely free software drivers

AMD's 7nm GPUs will BTFO nvidia next year right guys?

poo in it

My money is on Intel using ancient kike magic to btfo nvidia.

my build from around 4 years ago is all amd, not sure if i should fall for it again when i do another build soon.

Attached: Raja on joining Intel.webm (500x281, 203K)

I want to see the pcb of that thing

Attached: back[1].jpg (1624x600, 413K)

Probably mIlITAry-gAMinG-as-blacked-as-my-soulmate black with some LEDs on it, like all the crap that's coming out these days. Of course with 4 x 8 PIN power supply ports, for optimized fragging at 720p


Attached: Screenshot_20180813-215242.jpg (1034x867, 245K)

>real-time ray tracing
can't finna wait to jerk off to future hentai games brothers...

>Illusion still don’t use SSS.

ITT things Nvidia adds that make no difference for us. The card is great but Nvidia is not. They are just intel except they are winning. For how long though? Will 7 nm BTFO them? Will it be another VEGA like disappointment? It remains to be seen.

AMD RTG already gave up. They are going back to ATI's roots with OEMs via iGPUs and semi-integrated solutions.

And console

Still 48 GB on a single card, wtf...
AMD+NVIDIA is the way to go for 2020, unless you need a budget GPU. 480s give a very good performance for their price point, but if you want to buy anything higher get a 1070 Ti or 1080 Ti.

Well if you can throw random numbers around so can I.

>RTX with 2TB of GDDR9 at 1 gorillion terahertz

Sounds cool on paper but honey, not gonna happen.

NVidia made a demo on stage you dumbass

>How can we make RTG/AMD Great Again?

by making freesync cheaper.

gsync is a corporate joke.

you can't do anything.
On the pro side, amd largely keeps pace price/performance, but so much got locked down to cuda
on the consumer side, amd had 5 generations where they beat out nvidia in a row, and at least 6 if you take into account the 5000 gen for nvidia and that did fuck all

The best amd could do is establish an open ray trace that is backwards compatible because nvidia decided to split the baby it seams, offering ray tracing to a grand total of 6% of its potential player base, if amd comes out with something that works on existing cards, and potentially accelerated on newer ones, you take a clear win from nvidia before they secure a win.

other then this, consumer side, amd is going to struggle until nvidia finally fucks their own ass, or intel/amd partner ship on the gpu side gets more wide spread use, not just some skus. if intel amd takes off, then nvidia proprietary will no longer be used as standard.

but I believe there may be a good chance this upcoming generation will hit nvidia hard as they may be marking gpu prices up higher again, and it's getting to the point its not ignorable, while leaving the mid range to either pay a high price or fuck off.


Attached: 2018-08-14 00_19_45-NVIDIA.png (3840x2160, 2.96M)

they will compete, not btfo
The main issue with amd is a bottleneck, vega was suppose to alleviate bottlenecks, but it performs like a suped up fury which was also heavily bottlenecked.
apparently 2/3-3/4 of the team for vega got moved to navi before vega was close to done, so it may have been a 'fuck it, this isn't the gpu we need' move, and navi will come with revamped arch components that alleviate bottlenecks.
However navi will likely be competing with nvidia's 2070 rather than trying for fastest.
However keep in mind, amd is likely going to go back to the sweet spot mentality they use to have.
They are also making an effort to get chiplet gpus to gamers rather then just the high end, as that would be a game changer for price performance goes, but they have to make it invisible to venders as even amd calls game devs incompetent retards.

at least this is my understanding of the situation.

the thought is marketing, they add the rtx in it because raytracing is the final step for realism in graphics, it creates a clear have and have not, obsoleting everything not rtx immediately, and if they keep the gtx for the mid range, it may upsell people who want to have the best but cant afford it if they know they will miss out.

AMD has completely given up on discreet desktop GPU's, enjoy your Nvidia assrape.

MidRange cards until 2021

According to AMD, there will not be a high-end GPU in 2021.


So it's a quadro series with new software?!
What happens to GV and GP series?!
Do they also will supported by this RT software?

Volta yes, pascal is slooowww
>watch suicide amdfans

>MidRange cards
Not even that. Where is the answer to the upcoming 2050/2060?
>inb4 2H 2019

Why the fuck would anybody buy an AMD card? Stockholm syndrome? Cuck fetish?

>Buying a shitty 1060 over a 580
Sorry, I like my 8GB VRAM and good DX12/Vulkan performance.

amd is starting to entering on traditional nvidia markets already user
a.i(well with navi we will see if they finally have a simple lang like cuda..)
they are about to open their second studio in hollywood..
embedded ryzen with vega is unmatchable more powerfull and uses less power

we will see if lisa can ressurect them

Maybe their next-gen gpu will be a high-end version of navi

Attached: 1-AMD-CES-GPU-Feature-.jpg (750x213, 44K)

Nvidia ray tracing being confined to windows 10 only is too much of a limitation. If they support vulkan they might do a little better. The chances of anyone making a game around it is slim.

They need to simplify their user GPUs pipeline, nVidia is winning and if AMD wants to have it their way they need first be relevant then demand game developers to follow their design.

We probably won't see ray tracing go mainstream for another 4-5 years.

Everyone who has a graphics card right now won't be able to utilize ray-tracing.

And of those who plan to upgrade, the price premium will keep those who own an RTX GPU to be too small a group for game developers to focus the development of their games towards.

It's early adopter technology for the next 2 to 3 years, and it won't be fully mainstream until the next 4 to 5 years.

AMD has time.

>Supporting ray tracing
lul, shit API is shit.

>lul, shit API is shit.

he mad

I did the math, if the 2000 sell in line with 1000, then only about 6% will be able to ray trace if its locked to tensor cores. and this number wont get better if they lock it to X070+

Then you have the fact a 1080ti is enough for 4k 60 some reduced settings, and we have no 120/144 4k monitors yet, at least cheap ones, and mix in some leaked they are decreasing the ram the gpus have news.

let's put it this way, ray tracing could effectively do global illumination as its not something you need sharp details to do, and if this is able to be done on older cards only, i'm ok with it.

I think that they are heavily jerryrigging raytraceing though, but I have never seen a really clear demo of it, and if they do it this way, the first gen of it is going to look great and suck cock a the same time.

It worked for Zen.


Fuck the end seems so close now. I'm very excited to see what the industry will do in the next few years when they can't shrink their manufacturing processes anymore. 5nm is the last possible if IIRC.

the smallest possible on silicon is 3nm actually
don't worry user, we still have some years of nice performance improvements before stagnation

>That HBM2 bandwidth and voltage
Can't wait for HBM3. Is it slated for 2019 or 2020? I think it's gonna debut with Navi.


Oh look a feature 1-2 dev will ever use and as a tech demo.
Unlike nVidia solution

>nobody will use actual free solution working on any hardware
>everyone will use novideo gimping blob instead
And it's not like anyone will even think twice, they will actually go full retard and do it.

AMD haven't had money in a while to send their own devs to game studios to do shit in there with them.
Nvidia has been doing this for ages.
Which one do you think some cowadoody dev will choose if they can literally get on site tech support for implementing features for a specific GPU?
The one that sends devs in there for optimization and hand olding or AMD who tells them to wait for it?

RTG is being made great again though. When Raja was fire- "left", he took all his poos with him so mommy Lisa replaced the top management at RTG with key people that helped make Zen a stellar success. I heard rumors that Papermaster gassed a bunch of people, promoted some good and loyal engineers (non poos didn't get much from Prince Raja) and they are restructuring the whole corporate culture there. RTG will be more like the CPU design team. Lean and flexible, with multiple leapfrogging teams working on present and future projects. AMD has been dumping millions of more dollars into RnD and hiring for the last 8+ quarters. RTG is restructuring and just beginning to ramp up again. 7nm Vega will be a pipe cleaner for the procces and the new RTG. I expect Navi to be delayed as they review the design and manufacturing goals to make sure they can put something out that is competitive in at least one big market, be it enterprise or consumer. If AMD can make a GOOD and supported alternative to CUDA, I think they can take some sweet AI bucks from Nvidia.

how 2 get job at amd bros

put 4 vega dies in a one chip and sell it as a workstation GPU for cheaper