Raja shits the bed so instead of Nvidia working on Volta, we get more Pascal refresh because there is zero competition

>Raja shits the bed so instead of Nvidia working on Volta, we get more Pascal refresh because there is zero competition
Fucking Christ, AMD, kill yourself

Other urls found in this thread:

devblogs.nvidia.com/parallelforall/inside-volta/
twitter.com/SFWRedditImages

1070ti is exactly what I want so I can't complain much

Why? 1070 and 1080 exist. All it's going to do, if the rumours of a 1070ti are even real, is lower the price on 1070. Maybe.

I want a cool power sipper. Thats usually what the TI are

Wait, nevermind, it's not even a refresh. It's the same series. Fucking Christ.

Two years of the same series

It doesn't mean anything. It's literally going to sit in between the 1070 and 1080.

Volta was always 2018, it was never delayed ever

Volta was delayed the moment Pascal appeared on roadmaps.

Except for when it was 2016...

That was before Pascal magically appeared.

Doesn't change the fact that Volta was once 2016, so there statement that 'Volta was always 2018' is a lie.

And Maxwell once was on 20nm and 2013 for the entire family but that didn't happen.
And Threadripper never existed on the roadmaps, period.
Roadmaps change all the day.

That's point of this thread.
nVidia changed their roadmap again due to AMD being fucking terrible.

NVIDIA changed their roadmaps because it needed to be changed and not because of any external reasons.
Same goes for AMD.
Vega20 is not coming any sooner despite nVidia rushing out V100 like there's no tomorrow.

>because it needed to be changed
For what reason user -pixies?
Nah
They changed it because they know they can sit on Volta for a year now.

I would too, I'm not even upset at nvidia for doing the best for their business.
We're upset at AMD for being so fucking pathetic that nVidia CAN just sit on their collective arses for a year.

Volta won't come any sooner no matter how spooky Vega would like if it was launched in working state.
It's bigger dies, they need decent yields on their half-node to launch consumer shit.

This.
Becuase all it ends up in is stupid shit like CPUs now.
>bbut muh Ryzen shaking up Intel
After how many years of Intel doing sweet fucking nothing - because they could afford to since they had literally no competition at all.
Now Ryzen is shitting up Intel's plans, but as consumers we're still sitting on SandyBridge ~+10% IPC at the most across everything and will continue to do so for the next 2-3 years until Intel can finally get something new out and AMD can tune Ryzen properly.

If AMD keeps this GPU shit up were in for the exact same bullshit as consumers.
nvidia will sit on their asses, AMD will suckerpunch them in 4-5 years with equal performance at half the price but GPUs will have stagnated for 5 years in the fucking meantime.

If Nvidia fanboys had bought AMD when AMD actually had better products (yes, that was a thing in the past), AMD would have gotten more RnD budget. It is entirely your fault that there is no competition as all you shills do, is shove money down Nvidia's throat regardless of what they do.
Another example. Gtx 1080 - midrange GPU when going by die size. High end prices. Not even the slightest portion of criticism from customers or the tech press.This problem is homemade and I dont have even the slightest bit of empathy for you.

Navi is AMD's only chance to suckerpunch Nvidia on price in the near future. Nvidia is looking at a MCM strategy of their own. AMD just happens to have planned theirs a little earlier.

no, no it wasn't. the drivers were shit back then and your looking at the past through rose colored improved drivers glasses.

If AMD gets it to a functional state earlier, they can patent the shit out of MCM and get money with every Nvidia GPU sold.

>planned
The interconnect is there already.

Fermi.

>says the woodscrews 1.7 man
Ayy.

fermi was better than whatever amd have at the time on day one.also

mouse cursor bug.

According to techpowerup, fermi had 11% more performance than 5870, a lot more power draw and was 6 months late. Also, GTX 970 still not forgotten.

It really is funny how that bit of marketing is a lie, not only for volta, but even the current generation of nvidia cards. AMD just sort of gave up and decided vega wasn't worth manufacturing in any real quantity.

This. Also shitting on AMD for being uncompetitive for half a decade in CPU market is like blaming a wheelchair bound kid for not winning against Bolt in a 100m. AMD was shafted by patent lawsuits and had no sales because intel bought off all OEM business.
Can't be competitive without R&D budget, can't mint dollaroos without marketing budget.

>Buying AYYMD HOUSEFIRES
>any year

TOP KEK

and the drivers actually worked well. amazing gamers actually care about something like that right? quit crying about customers not buying underwhelming products.

>Fermi
>was better
>than Evergreen
Talk about revisionism.

because they've planned it
what are you trying to say?

Yeah. Nvidia is doing the same thing Intel did for the last 10 years with their I series. All because once again amd is unwilling or unable to compete.

>hardware
>is half
>the equation

talk about denial

AMD is more than willing to compete in pro markets, their Pro lineup is the best it's ever been.
Consumer, well, you all brought this into yourself.
Enjoy!

>reeee why don't customers like the hype lie disappoint strategy of AMD

I'm not a shill or a fanboy. I buy what's best for my money at the time I am building. Unfortunately for AMD it hasn't been them since I bought a 9800 PRO, but wait, that was ATI then.

Wow, AMD has been a real shitter in my timeline

>10% more perf for $150 more
>was
>totally
>worth it gweilo

>Pro lineup
Please buy our expensive cards. You are a prosumer right? You need the best. Please give us money.

>poor volta
>they didn't even beat Pascal

Talk about deceptive marketing

That's how it works for both companies.
And WX9100 is still like half the price of P6000.
>lie
Sorry wrong company.

Like today where a vega 64 costs the same price as a 1080 ti except the vega is 10-15% slower?

don't be bitter, user. think positive. in five years amd will come up with something like ryzen for video cards

Buy intel hd graphics and hope one day intel becomes a formidable gpu contender. It's what I do. leather jacket man would really hate that. he hates intel to the die.

Vega64 costs $499, which is 200 hundred burgers less than gutted GP102.

They are talking about navi you tard.

Why do you think the vega line up is so limited? They cut back on vega to focus on navi.

>Vega lineup is so limited
4 dies is limited?

Honestly, I'd say they already are. I can play a surprising amount of games on my 10 year old iGU.

>Vega64 costs $499
If amd said I could buy a unicorn for $1 I wouldn't believe that either.

If you can find me one at that price I will cede the argument to you. But here in the real world we deal with real prices, and even a month after release it is still the same price as the 1080 ti.

>when you push a poo on stage and demo cards with a stutterfest like that prey 2 demo and wonder why people don't buy your shit then get mad at them after telling them to wait for a competitive product for a year

just AMD things

Before you shitpost me for cherrypicking I just wanna say I'm an nvidia fan that's going to get a vega 56 because of my 970 legal case.

nvidia are straight up liars who pay game devs to gimp AMD hardware and optimize for their cards. AMD clearly has more power in their chips (see FP32, mining hash rates) but nvidia is constantly pressuring game devs to hide that.

How many dies does nvidia have?

1030, 1050, 1050 ti, 1060, 1070, 1080, 1080 ti?

There's more to GPU than shader math.
GCN is extremely weak at geometry.
Vega fixes that, but not now.
GP100/102/104/106/107/108 to Vega20/Vega10/Vega11/Polaris10/Polaris11/Polaris12.
Kinda like that.

nvlink is still too slow, it's only useful for compute.

NVLINK is point to point horseshit aka PCI-E on such steroids it needs retimers every 1.5m of traces.
It's not in any way related to any on-die links nVidia might be developing.

>There's more to GPU than shader math.
>GCN is extremely weak at geometry.
>Vega fixes that, but not now.
It's more than that. See hairworks and crysis rendering objects that can't be seen by the user or excessive tesselation on AMD cards.

Nvidia has not played fair.

NVIDIA played to its strengths. DOOM on Vulkan is an example of AMD playing to its strengths, nvidia simply lacks scheduling power to handle that much async.

no man, nvidia has cheated and bribed its way to the top

see the adoredtv videos on shillvidia

WAHHH THEY DONT PAY FAIY

amd children need a safe space

its not like i would buy volva or pasval refvesh anyway, im a fucking poorfag

t. Jensen

>amd children
Not a fanboy. I just want competition

I'm a 970 user with a pending legal case on nvidia for false advertising. I was promised 4 GB of vRAM not 3.5.

nvidia doea not only cheat and lie in benchmarks but to their paying customers as well.

Well you finally got one.
In Pro market.

Please make a POOR RAJA pic

devblogs.nvidia.com/parallelforall/inside-volta/

>The new Volta SM is 50% more energy efficient than the previous generation Pascal design, enabling major boosts in FP32 and FP64 performance in the same power envelope.

>Volta GV100 is the first GPU to support independent thread scheduling, which enables finer-grain synchronization and cooperation between parallel threads in a program.

THANK YOU BASED NVIDIA

>SM
And CUs in Fiji were like 80% more "efficient" compared to Tahiti WITHOUT the new node.

This is Nvidia not AMD.

And?
Big fucking whoop for such a statement.

>Nvidia is doing the same thing Intel did for the last 10 years with their I series
>Pascal was the largest performance boost since before Fermi

>Consumer, well, you all brought this into yourself.
Yeah it's the consumers, never mind that AMD hasn't released a good card since 2011. Consumers should just buy AMD cards by the truckload even if they're worse.

I too love double node jumps.

its a 1080ti downlocked you dingus.

If only AMD loved double node jumps...

What else would they do with peanuts for R&D budget?
Throw money at gaymen market and hope children finally start buying Radeon again?

They are too busy removing that bloat that is DX9 and onwards pipelines.

That's all for deep learning and has nothing to do with graphics.

Vega got huge manufacturing, just 90% of the gpus were already bought by apple or seever farms.

No.

>wants nvidia's only competition to kill themselves

so this is the power of nvidiots

They killed themselves with 2900xt and nothing will bring them back.

>cryingpajeet.jpeg

why don't they want the poo I shit out!?

>nvidia's only competition
WHAT COMPETITION

man you guys need new arguement sits getting really repeditive and using the same arguements.
*yawn*

>nvidia delays their development and starts milking customers with the already existing products like good little jews
HOW DARE AMD DO THIS?! SERIOUSLY WHY CAN'T WE SUE AMD FOR NVIDIA DELAYING THEIR NEXT LINE-UP AND SELLING THEIR 2 YEAR OLD CARDS IN NEW PAINT COATING? FUCKING AMD, WHY YOU DO THIS, WE COULD BUY NVIDIA'S NEXT GEN CARDS BY NOW INSTEAD OF WAITING YEARS! STUPID AMD, STOP DELAYING NVIDIA PRODUCTS!

t. i want amd to make nvidia shit cheaper but will never buy amd even if they are better
its almost like you are a massive cuck that doesn't understand capitalism

>but will never buy amd even if they are better
I would but they're literally never better

this, literally never in history has amd been better than nvidia

ALL HAIL OUR LORD NVIDIA

this is just fake news

>what is the 9000 series
>what is the HD 4000 series
>what is the HD 5000 series
>what is the HD 7000 series
>what is the RX 200 series
>never better

That post was ironic.

oh shit ment to reply to this dumbass

There has to be a company looking to compete with Nvidia in the GPU market. There has to be

No one retarded enough to enter mindshare-driven market.

0 chance of that ever happening, same way with any other company trying to make x86 consumer processors

>Nvidia/Intel/AMD file 300 different lawsuits/deny you access to their retarded proprietary standards, launch 1,000 year long legal battle that costs millions of dollars
>have to assemble team of engineers that have the necessary skills, spend millions of dollars researching and designing new product, have to spend even more money to then buy time on fabricators to print them and assemble everything
>and after all that fucking shit then you need to build a gigantic support structure with drivers, constant updates, need marketing and PR people to suck reviewers dicks all day try to convince normies to buy your new product

And after doing all of that shit nobody buys your product anyway. And then the bigger company with more experience and resources just releases new stuff that completely blows you away. Why would you bother trying to enter a space where Intel/NVidia have been at it for so long, and have such a massive lead that there's no hope you'll ever catch up? That's a bad business decision, hence nobody is dumb enough to try, they invest in other shit instead.

What do we do when RTG dies

>RTG
>dying
?

ITT children talk about shit they know nothing about. The tech landscape is complex. There is a very long back story to all of this. If you cared to do any research you would know there is a fundamental reason for why AMD, Inteal and Nvidia are the way they are.

But keep shitposting. Please do.

They should bring back Killebrew for Navi.
He's totally the dude that won't let the schedule slip.
I sincerely advise you to kill thyself.

>You can tell because of the way it is

>im smart
>u guys r dumb
Really made me think.