https://www.nvidia.com/en-us/geforc e/products/10series/tita

>>59767170
What's happening?

>>59767192
Titan X with 3850 cores.

File: amd.png (42KB, 653x726px) Image search: [Google]

42KB, 653x726px

>>59767170
NOOOO

>>59767215
So an even more retardedly expensive titan for no reason?

>>59767261
It's 10% faster.

AMD can still recover the market.

Wow, wow this is bloody amazing!

It's almost irresponsible performance and horsepower!

>>59767170
>hurr durr

>>59767352
Are you stupid or retarded?

This is TITAN Xp, 3840 CUDA cores.

It's faster than the GTX 1080 Ti, period.

>>59767268
This is a milestone for nvidia, they've hit double digit performance improvements

>>59767170
So basically a Quadro P6000?

File: hitman.png (16KB, 622x581px) Image search: [Google]

16KB, 622x581px

What you can expect is some 10-15% faster than a 1080ti

>Sell Titan XP for $1200
>Gimp it so you can resell the ungimped verison later
Gas these kikes.

>>59767365
Yes, I'm retarded.

>>59767441
Not gimped, yields weren't ready for a full GP102 die, reason why the Quadro P6000 released only recently.

>>59767170
>/v/

>Release Titan X
>Release Pascal Titan X with the same name, everyone starts calling it The Titan XP so it isn't confused with the old Titan X
>Release a new card called the Titan XP so it can be confused with what everyone was calling the old one
Why is Nvidia so bad at naming shit?

>>59767476
Yeah this is egregiously bad.
>When we say the Titan X we really mean Titan XP, but not the Titan Xp.
Fuck. How did they not know people were colloquially calling Titan X Pascal the Titan XP already. Now we have to deal with this shit.

File: 1483406568843.jpg (71KB, 552x661px) Image search: [Google]

71KB, 552x661px

>>59767333

>>59767503
Titan X = Titan XP
Titan XP = Titan XP2

>>59767503
How many people are really dealing with this though?

Only those who watch eceleb fags like Linus ponder over overpriced hardware they will never be able to afford.

No one buys these things so no one cares

I'm more confused why they called it 1080Ti and not just 1090

>>59767170
>>59767476
Holding out for the Titan Vista?

>>59767407
What should I buy if I want to get minimum frame rates up to 30-40 fps?

>>59767597
Whatever comes in 2019 at $1000

So.. Volta +1

>>59767597
a 480p monitor

>>59767610
Are you telling me that right now I have to pay $1200 for single digit framerates? That's fucking insane. Who buys these cards?

>>59767642
Idiots and working people who need a lot of FP16 for machine learning.

so AMD now needs to beat:

980Ti aftermarket variants
Titan X Maxwell
1070
1080
Titan X Pascal
1080 ti
Titan Xp

>>59767543
>I'm more confused why they called it 1080Ti and not just 1090
They probably are saving the X90 name for if they ever need to do a dual GPU card like the GTX 690 again.

>>59767170
fuck I'm mad as hell.
but i guess i need to pile up 3k € to upgrade to the next two Titan in 3 years. btw i have a Pascal sli, ask me anything.

File: amd2.jpg (52KB, 568x612px) Image search: [Google]

52KB, 568x612px

>>59767750
Vega will BTFO and bankrupt nvidia in one go. Just you wait.

nvidiots literaly blacked.

>>/v/

>>59767503
So people with short attention span clear the inventory for the old Titan XP. Implying this wasn't delibertae.

>>59767170
Fuck nv naming scheme right in the ass

1. Release Titan X (Maxwell), everything's fine
2. Release Titan X (Pascal)
3. Tech community starts calling the new card Titan XP to avoid confusion with the old Titan
4. Release yet another card called Titan Xp just to fuck with everyone

just fuck right off nv

>>59767889
not if its cheap enough.... 1080 performance +10-20% for $450ish....

feel sorry for people that need to be on the bleeding edge....

>buys titan x ... 1080ti comes out... buys 1080ti.... titan xp comes out .....

>>59767564
heh
I get it

>>59767750
No, AMD just needs to make a card capable of running the latest games with high settings at 1440p while being cheaper than the nvidia cards that can do the same.

>>59767251
why no, why would nvidia put a 5000$ gpu out for consumers?

I would bet on nvidia knows something we dont when it comes to vega, and they want to still claim to have the most powerful gpu...

This would mean that vega is coming out faster then I was expecting, (33% efficiency put it at 1080, polaris efficiency put it closer to a titan x then the 1080, 17-15% performance disparity puts it above 1080 ti)

So either the production vega is going to be significantly faster then the pro vega, which is most likely the info nvidia got wind of so they are pushing a 5000$ card to mainstream to retain fastest card title, or they got wind of efficiency gains, something that is far less likely.

this points to vega being a monster.

>>59768350
thats pretty much what the 480 8gb does already

I called for a new titan black here.
But I'm surprised that they didn't wait for Vega.

tinfoil on: since I'm sure that Nvidia knows a lot more about Vega than us peasents, this could mean a quick cash grab on ultra high end enthusiasts before Vega launches.

or it doesn't mean jack shit and they just had a surplus of pristine gp102 and p6000 weren't doing that well and fuck vega.

>>59768386
I have a 480, it does not.
It's a 1080p card.

>>59768390
>since I'm sure that Nvidia knows a lot more about Vega than us peasents,
Nvidia has Radeon Instinct Mi25's already, they know how Vega does.

>>59768350
so... an overclocked RX 480?

>>59767333
not till the mcm the gpu and can get full product stacks out every 9~ months, this is also when amd can effectively fight nvidia on all fronts, from the extreme low, to the extreme high, without adding significantly to the cost, while still producing overall larger total die areas as you have no sacrificial dies for the monolithic ones.

>>59768386
>RX 480
>1440p card
Only if you like slide shows.

>>59767656
you get amd for fp16

>>59768417
RX480 8GB here, I can run Witcher 3 in 1440p60 on high preset.

>>59768432
Vega for the consumer market isn't out yet, so no.

>>59767333
BERNIE CAN STILL WIN

>>59768442
>working people who need a lot of FP16

you aren't going consumer in that scenario either.

>>59768436
>RX480 8GB here, I can run Witcher 3 in 1440p60 on high preset.
No, you can't.

https://www.hardocp.com/article/2016/06/29/amd_radeon_rx_480_video_card_review/11

>>59768459
Titans were traditionally prosumer cards, they gave decent FP64 for the price without jumping to insane prices Teslas.

Now the Titans went from FP64 prosumer cards to gaymer crap with FP16 capability

>gimp original product
>rebrand it with the full chip a year later
Sasuga Nvidia

>>59768391
>>59768417

its as powerful as a 290x, and it still does 1440p 60fps

>>59768469
>reference card clocked at 1266MHz, probably throttling even lower
>old review with old drivers

I could post screencaps when I'm home

>>59768507
1440p30 yes, 1440p60 no. Stop spreading lies.

>>59768519
>screencap
>1 frame
This is how AMDcucks measure performance.

>they actually used Linus' idea

File: Untitled.jpg (768KB, 2560x1438px) Image search: [Google]

768KB, 2560x1438px

>>59768405
Sapphire Nitro+OC here thats been overclocked a little bit further

>>59768436
>>59768519
No

>>59768417
>>59768469
Yes

>>59767170
I bought a Titan X (Pascal) recently :(

>>59768531
Were you expecting a full performance review from me just so you can belive me?
I don't need to prove anything to a random fag on 4chan.

>>59767251
>>59767505
>>59767889

Nvidia releases $1200 card
... it all amd fault

>>59768477
ok, lets just put this here

the titan use to have huge double flops, something like 1/4th its single precision, they they gutted it to what 1/32?

now titan x
its single precision flops is 10157
its double precision is 317
and its half (16) is 159

meanwhile the 480, amds card not nvidias
single precision is 5161
double precision is 323 - better then the titan x
and its half (16) precision is 5161, the same as its single, nearly 50 times better than nvidias titan x.

yea, these are gutted to fuck and back gaming cards, no hint of pro left in these fucking things. If you are a pro, you buy them because you are fuckin retarded, if you buy them for gaming, you either got them because you are fucking retarded, or you built a setup that was not playable with the 780, 980/980ti or 1080/1080ti and are fucking retarded for building something that is barely fit for purpose.

>>59768631
Wait, even Titans don't have FP16 anymore?

Fucking hell these are pure gamer cards.

>>59768436
My 1070 barely does that including hairworks

>>59768524
what the 290x? I watch that thing shit out 1440 at 100+fps like its nothing, and a 480 doesnt fall more then 20% behind it at that resolution.

>>59768519
>screencap
Want me to show you witcher running at 1800fps?

>>59768556
so to prove them wrong, you show a screen shot of what has to be the worst optimized AAA game to come out since arkem, that never got fixed like arkham did?

File: image.png (34KB, 636x812px) Image search: [Google]

34KB, 636x812px

>>59768727
>1440p 100fps like its nothing bro
You should sell GPUs on YouTube

>>59767378
even though pascal was a near 50% improvement from the last architecture

>>59768524
bf1: 1440p maxed out: 70-60 fps
rainbow 6 seige: 1440p maxed out fxaa + temporal filtering: 60-90 fps
Dying light maxed out - gaymeworks shit: 60 fps
DOOM: Maxed out: 90-120 fps
War Thunder: Maxed out: 70-100 fps

stop lying you dumb shill a 290x or equivelent (rx480) is enough for 1440p

>>59767170
>https://www.nvidia.com/en-us/geforce/products/10series/titan-xp/
what core is it? bigger version of gp102 or cut down gp100?

>>59767170
something i will never buy released

>>59768649
the maxwell titan may do single, not sure as the source wasn't interested in 16 at the time they did the tests.

But considering that fp16 is one of the most important things you need as of now for many gpu is best purposes, I would be shocked if full gp102 wasn't gimped in half precision.

>>59768792
>war thunder maxed: 70-100
HOLY SHIT what is wrong with your GPU? Must be broken to be that fucking terrible, I get about 350fps at 1440p maxed in war thunder, with the same card that only gets 70fps maxed in witcher

>>59768780
>shill benchmarks
>not saying what aa is used
> Tesselation out the ass that you will never notice

>>59768829
i mean totally maxed out
msaa 8x fxaa on top of that a bunch of other settings too, im talking about ground forces btw too

File: image.jpg (124KB, 1000x847px) Image search: [Google]

124KB, 1000x847px

>>59768831
I thought you said you got 1440p/60fps "maxed"?
I thought you said Witcher 3
I thought you said 1440p
I thought you said 290x
I must've been mistaken

>>59768856
what aa? i never said witcher 3, and the media are all lying kikes

>>59768851
Yeah obviously if I'm getting 350 fps I'm maxing everything possible
You're just mad your shitty 1080p card can only hit 60fps for 3 seconds in witcher
Don't forget amd themselves said the 470 was a 1440p card, just go buy thaT

>>59768856
>putting on msaa x8 when you are at a resolution so high you can barely see jaggies
mental disorder

>>59768780
ultra + aa, yea, that will fuck you over hard, but because most settings don't really change much from high to ultra, shadows don't really change from medium to ultra when I did my tests, you add needless calculations to the game for no other purpose then saying you have it on ultra.

but we are talking high settings, not retarded lets break the engine push it to 11 settings.

you know there is a reason gta5 put those settings on another page instead of in like with everything else right? wish more games would start doing that.

>>59768874
t. someone obviously shitposting and mad that a 4 year old amd card still does 1440p 60fps

>>59768886
>I get 60 fps maxed bro wanna see a screenshot?
>Wait no I mean who actually maxes their settings in 2017 haha nerd

>>59768899
So you get 60fps 1440p on witcher 3 on high and medium settings then?
Just say that, don't be an annoying ignorant twit, because people can and will post benchmarks showing you 200% wrong

>>59768787
it was and it wasn't, it was a die shrink. when you equalize for flops, maxwell and pascal perform equal, if they got more ipc you would have seen pascal pulling ahead.

Yes, through clocks and putting more shit on die, they got more out of the gpus... kind of, an 980ti gave a 1070 a run for its money, and overclock for overclock without throttling, a 980ti gave a 1080 a run for its money too, performing within 15%

>>59768903
You now realize what people here mean when they say they have a 660Ti or HD7870 and say they can still run every game "maxed."

>>59768931
yes a benchmark pushing retarded settings that make nest to if not no visual difference sure showed me wrong when we were talking about high, not 'omg cumming in my pants ulta overkill your friends can't play on this setting' settings.

>>59768350
maybe for you but what you just described isn't the entire market. some people want 4K 60Hz, some people want 1080 144Hz or 1440 144Hz. Some people just want to own the best because it's the best.

Also having the most powerful card in the world is pretty much the same as having the only queen in a game of chess. This is because even if the rest of your lineup is shit, less-knowledgeable consumers (which is most of them) will invest in any part of your lineup because since you have the better high-end card "all your other cards must be better, right?" It explains why people who own a gtx 1050 still think they can call people who own an rx 480 poorfags cos AMD is the "cheaper option" and therefore, in their minds, always worse.

AMD need to beat nvidia this time round. The rx vega has to beat the 1080ti to sell well because even if it comes in at $550, that's a price point where saying "what's another $150 for more performance?" actually makes sense. Maybe not the Titan Xp since most consumers can see that card for the e-peen money-furnace it is but owning the better high-end card will help AMD massively.

Funny how this all started by me saying 1440 at high settings, not even maxed.
Granted I took a screenshot from one of the worst possible scenes that you could look at for an absolute minimum but if that was 40fps+ instead of 20fps it would make a world of difference and i'd happily declare the 480 a 1440p card.

>>59768372
>I would bet on nvidia knows something we dont when it comes to vega, and they want to still claim to have the most powerful gpu
This would make a lot of sense.

>>59768903
>>59768944
show me the image, not zoomed in to make 25x bigger where you can see the visual difference between 2xmsaa and 8xmsaa other then rainbow 6 sieges absolutely fucked implementation and the point will be conceded to you

I personally don't count aa toward maxing a game out, as its no longer the 90's where jaggies are everywhere to the point aa is almost necessary, we are pushing toward if not already at resolutions were aa doesn't even make sense turning on

>>59768903
Using 8x msaa at 1440p point and laugh at the retard

I never mentioned that shit autismal Polak game anyways

>>59768831
Hairworks and HBAO are disabled retard-chan, you would know if you actually checked out the article.

>>59768886
Witcher 3 does not support MSAA at all except for Hairworks. Hairworks was disabled in that benchmark, so there was no MSAA at all. The AA in TW3 is some type of temporal post-processing AA.

>>59769023
i'm willing to bet they know the clock speed, optimization... if they knew that they would have never pulled the 650 on the 980ti.

>>59768982
No, being the best card choice at 1080 and 1440 is exactly what they need because its the majority of the market and it will stay like that until vega is obsolete (AMD obsolete, not Nvidia "we got a new card to sell" obsolete)
4k is too niche for them to bother, they are stupid to even talk about it.

People will still buy Nvidia even if Vega is twice a power for half the price, nothing they can do about that.

>>59767333
This! Just wait!

>>59769036
>Giving clocks to Nvidia shills
Never

>>59768581
>I'll post a screencap to prove my point to a random fag on 4chan
>I don't need to prove anything to a random fag on 4chan.
>I'll just start backpedaling because I got called out for using a screencap as a substitute for a benchmark

>>59769114
>All of this just because someone claimed that the 290x does 1440p maxed out

Pure autism

>>59769070
>People will still buy Nvidia even if Vega is twice a power for half the price, nothing they can do about that.

That's exactly my point. You have to ask WHY do people still buy nVidia cards if they're objectively worse and it's simple. nVidia have the most powerful card in the world so they're the "winners". People like being on the winning team.

>>59767642
>Are you telling me that right now I have to pay $1200 for single digit framerates?

>16x aniso
>2160p

Try launching Furmark with highest filtering and aaliasing, it will shit on every card made.

>>59769167
How delusional are you?

>>59768507
It isn't.
GFLOPs is meaningless outside of raw compute tests.

290X has twice the ~20% more TMUs, 50% more memory bandwidth and double the ROPs of a 480.
480 is a 1080p card that can do shaders and effects out the wahzoo.

>>59769239
He's right though. People justify their purchase of Nvidia cards in the mid-range because they're the kings at the high end, even though the 1060 is a piece of trash compared to the 480 and the 1050 Ti is a piece of trash compared to the 470.

People will spend more for less performance because it says Nvidia on the box, as if they can bask in reflected Titan X performance with their shitty gimped 1060.

OFFICIAL GPU TIERS
[What are you even doing tier]
GTX1050
[900p and below tier]
RX360, GTX1050ti
[1050p and below tier]
RX470
[1080p Tier]
RX480, GTX1060
[1080p future-proofing tier]
GTX1070
[1440P Tier]
GTX1080
[4K Tier]
GTX1080Ti
TitanXP-old
TitanXP-new

This list is objectively correct.

>>59769373
This list is objectively shit

>>59769313
The 1050Ti is much cheaper then the 470 though. As most reviewers pointed out the 470 doesn't make sense, it's too close in price and performance to the 480. This leaves a huge gap between the 460 and 470, which the 1050 and 1050Ti fill. AMD fucked up badly, the 460 is dogshit but the 470 is too expensive. The 480 4GB and 8GB are the only good cards AMD released. (Especially the 4GB since you can flash the BIOS and possibly unlock 8GB)

>>59769173
So the test is shit, basically? How can I know if this card is any good then?

>>59769373
This is factual

>>59768436
no you don't. i have an rx 480 at 1320mhz and i can't keep 60 fps at 1440p unless i turn shit like water, shadows, foliage and terrain to medium with hbao turned down to ssao. even then it frame drops like crazy.

File: 1317726561959.jpg (11KB, 250x250px) Image search: [Google]

11KB, 250x250px

So they released a gimped titan on purpouse just to release a 1080ti and a full titan later.

LOL

File: ee.png (108KB, 800x510px) Image search: [Google]

108KB, 800x510px

>>59767170

File: 1474859180420.png (169KB, 296x300px) Image search: [Google]

169KB, 296x300px

>>59769373
1070 is actually CPU bottlenecked at 1080p, and even in some 1440p games.

>>59767170
Guess those P6000's aren't selling.

>>59769493
*on an i5

File: a47.gif (3MB, 355x201px) Image search: [Google]

3MB, 355x201px

>>59769373
>[1080p future-proofing tier]
>GTX1070


>future proofing.
>nvidia

>>59769475
Expect nothing different while AMD is gracefully allowing NVIDIA to have an exclusive claim on the high-end segment. No competition whatsoever means they can do whatever the fuck they please.

>>59768970
>anti-aliasing makes no visual difference

shut the fuck up.

>>59769475
And 900 series is legacy as soon as 1000 series showed up.
Welcome to jewvidia

>>59769528
28 inch 1440, its getting close to useless autismo

>>59768507
no, it's worse than a 290x

and my old 290x could only do Witcher 3 at 1440p 60 with a mixture of medium-ultra settings

>>59769475
they have plausible deniability due to yields.
But you are probably correct.

>>59769543
I have a 24" 4k monitor and even that leaves some aliasing, I almost always use a light post processing on the games that do run at 4k

>>59769572
>24" 4k
Opinion discarded for being a peerless retard.

>>59769597
I have one 5" 4k display and another 5.5" 4k...

>>59767170
will this run arma3 at over 30fps?

>>59769572
Fucking this, though my UHD monitor is 27" instead, there is still some visible aliasing if you don't use any form of AA at all. 1440p at the same screen size requires at least MSAA 4x (or equivalent) to look decent, no AA at that resolution and screen size is jaggy as fuck. It's barely an improvement above 1080p@24".

>>59769617
Your shitty phone + its screen costs less than his entire monitor most likely

Either way 4k at that size is just a waste of power

>>59769540
Legacy just means they don't produce the card anymore, it doesn't say anything about drivers, idiot

Using aliasing what a cuck, you guy like blurry texture ?

>>59769670
it's not a waste if even at 4k there is still aliasing

>>59769741
What are you even trying to say. Are you drunk?

>>59767170
the key feature is that this has mac support
when these new mac pros come out it's going to be the new platform of choice for most professionals.

>>59769730
Legacy drivers idiot

>>59769730
EOL means it not being produced

File: EVGA-GTX-560-TI-448-Cores-Classified-and-FTW-Now-Official-3.jpg (144KB, 1200x1200px) Image search: [Google]

144KB, 1200x1200px

>>59767476
>>59767503
Reminds me of the "560ti 448 cores". That was actually it's name, not to be confused with the normal 560ti, or the 560, or the special OEM version of the 560 that had even fewer shaders.

Then there was the million different versions of the 8800 GT/GTS/GTX/GTX Ultra, made more confusing when they rebranded it into the 9800. Nvidia used to be pretty bad with naming schemes. They've gotten a lot better with the 700 and 900 series, and for a while it was AMD's turn to have retarded naming scheme with the 200 and 300 series.

>>59769759
>high tech cutting edge GPU
$500
>Nvidia high tech cutting edge GPU
$1200
>Nvidia high tech cutting edge GPU in an Apple product
$2000

>>59769844
>nVidia high tech GPU with Quadro badge in Dell system
$6000

>>59769844
>AMD high tech cutting edge GPU
NaN%

>>59767543
Well I bought the Titan XP1 for my goddamn uber build just to build a 0 fucks given mustardrace machine. Then they release the 1080Ti which is close to the same perf as the XP1 and shortly after that they say fuck you again with the even stronger XP2. I feel fucked over kinda.

>>59770208
it's pretty much a given that you're gonna be fucked over if you buy a titan for $1200 anyway. you've just been slightly more fucked over now that another one came out with 300ish more cores. does it really matter once you're willing to be their cuck in the first place and give them $1200 for a card that should have always been $550?

>>59770146
what happened to HBM2
wasn't the 1080 ti supposed to finally start using that?

>>59767365
Except that it comes with Nvidia's dogshit cooler that will have it screaming and throttling like a bitch after 5 minutes of heavy load.
That's one area where I actually prefer AMD; At least they add a proper AIO watercooling system to their high-end cards (R9 295X2, Fury X).

File: 201601IMG_1891-edJT.jpg (718KB, 1400x788px) Image search: [Google]

718KB, 1400x788px

>>59767597

>>59767365
That's only like 8% more cores. It could easily be outdone by a decent AIB 1080ti with a custom cooler and an overclock.

>>59770451
will i finaly be able to play crysis?
or even watchdogs 2?
or anything at all @ 2160p?

>>59770208
The whole point of a Titan now is for early adopters who don't want to wait for an x80Ti. It's not even good for prosumers since they gimped it to sell more Quadros. Nvidia was even nice enough to delay the 1080Ti for 8 months so 1080 owners didn't feel screwed.

>>59768938
yet you are comparing a high end gpu at launch compared to a mid tier gpu

considering that the 1070 runs cooler, has more vram and was on par to a top end gpu just a year prior and was only(at launch) $350. it's was a huge bump up in performance

>>59770791
>$350
lol

File: looga.jpg (76KB, 633x758px) Image search: [Google]

76KB, 633x758px

>>59767170
DELET

>One thousand and two hundred dollars
>For a single GPU
The pinnacle of computing decadence

>>59770821
>(at launch)
cant you fucking read? it was just after launch when everyone went full jew

>>59769519
>It's AMD's fault that Nvidia do bad things.
(You)

>>59771004
I guess you don't know what competition is

>>59767170
>official drivers for Mac

but none of the current Macs can take a regular form factor GPUs

>>59769759
>>59771856
A few things.
>Thunderbolt eGPU adapters exist
>The pre-trashcan Pros are still quite popular in professional graphics environments precisely because they can use PCI-E GPUs
>Upcoming Pros (2018 or 19) will be more modular/expandable, although it hasn't yet been confirmed whether that means PCI-E slots
Finally, this is a great relief to the hackintosh community (myself included) who had mostly assumed post-Maxwell drivers were never going to happen.

>>59767215
Isn't that just a Quadro with less VRAM?

>>59772103
Firmware gimped so that GPU acceleration turns off when you try to use it with high-end graphics or scientific software

>>59771004
>fault
I merely explained that NVIDIA's actions and business strategy are a result of the current market state.

>2017
>playing games
>having a virginstation
>paying more than $400 for a GPU

when will you children grow up?

>>59771856
Apple recently made an announcement that they are going to make the Mac Pro more professional and actually admitted that the current form factor is crap (although not in those words).

Maybe Nvidia has some insider knowledge about Apple's new direction and is nudged towards making Mac compatible products?

File: 666.png (887KB, 1280x720px) Image search: [Google]

887KB, 1280x720px

>>59767170
>ayyymdpoojeets on suicide watch

>>59772353
What's so special about $400? Is that a magic number in your mind and exceeding it is verboten?

>>59772355
>implying Apple gives a long hush-hush lead up to partner companies
>implying Nvidia prepped Pascal drivers for Mac to coincide with Apple's "we might have a new Mac Pro in 9 to 21 months" non-announcement

Nvidia is just blatantly going after the Hackintosh market.

>>59772423

thats about as high as you go before you eclipse the optimal price/performance ratio and start venturing into diminishing returns just to run games at high detail settings that don't even make that much of a difference

medium settings is fine

a 1070 is 400 dollars and will play everything great, for half that an rx480 which is perfectly fine too

getting meme card 400+ is retarded, because in about a years time it will be worth a quarter of that

>>59772456
Hackintosh market is really small. If they are going towards any current market, it's for external enclosure users.

And yes, Apple does give out hush-hush lead to partner companies. And the timing for these two announcements is super-close.

File: van.gif (344KB, 256x192px) Image search: [Google]

344KB, 256x192px

>>59772423
Fisting is 300 bucks, so I assume he spends the rest on coke.

>>59771856
It's for hackintosh users.

>>59772481
Hackintosh users are extremely over-represented in the AI developer segment.

What AI software developers use to develop their DNN shit greatly influences which GPU they're eventually going to deploy that algo en masse on later.

Nvidia always makes it the path of least resistance to use their cards to develop shit on because they understand locking in developers early allows them to extract Nvidia Tax on software users later.

>>59767170
kek, another Titan? Why?
Either they're scared of Vega, which they shouldn't considering the known performance(unless AMD is hiding even more stuff), or they really want to fuck the consumer up. How the fuck do you support this disgusting company?

File: [Higashiyama Show] Gift [English Translated and Edited by Tonigobe] (Complete Manga)_Higashiyama_Show_Gift_ch07e_159.jpg (100KB, 418x410px) Image search: [Google]

100KB, 418x410px

>>59768412
>not till the mcm the gpu
This excites me so much. Imagine they making a very cute small GPU that can be stacked into a behemoth monster of a performer with 4 of them. It would be amazing since it would cost less for us.
That's the main thing Raja wanted to relay to the public when he revealed the RX480. It was a shitty way of presenting it, but two RX480s with perfect scaling are cheaper and perform better than a 1080. We need this shit already, I hope it comes next year.

Even the fanboy JayzTwoCents is shitting on them lol

>still GP102
You enjoy that cock up your asses, it's still not fully fledged Pascal.

>>59767170
This is legit retarded, only reason it was released is because nvidia was feeling pressure from the amd vega line

File: DROPPED.webm (117KB, 640x360px) Image search: [Google]

117KB, 640x360px

>>59767170
>>59773252
>still not GP100
>still no HBM2
DROPPED

>triple slot

clearly not marketed at the regular consumer

>>59773345
>titans
>ever being marketed at the average consumer
kek

>>59773016
>Imagine they making a very cute small GPU that can be stacked into a behemoth monster of a performer with 4 of them. It would be amazing since it would cost less for us.

with the kind of PSU setup you'd need for that i have a hard time believing it'd cost less

1:1 scaling is a pipedream too.

>>59773330
GP100 is only Tesla-branded and has no variants with video out capability, it's useless except for pure compute workloads. If you wanted compute in the first place, you have gotten some random GPU for general graphics display (or even iGPU) and stacked Teslas for however much power you need. There's no need for the display card to have 1/3rd FP64.

>>59773368
well clearly based on this thread and youtube comments people are talking about how "anti consumer" this is when literally no regular consumer will have one of these anyway.

if this is "anti consumer" then isn't the overclocked rx 480 aka rx 580 which is also being sold at the same price as the 480 like 6 months later "anti consumer" too? this are both exactly the same situation.

>>59773330
lol what a retard

>>59773420
http://www.anandtech.com/show/11102/nvidia-announces-quadro-gp100

You're wrong, but Nvidia would not sell it for less than the price of a flagship Quadro or Tesla card

>>59773420
>GP100 is only Tesla-branded and has no variants with video out capability
Wrong: http://www.anandtech.com/show/11102/nvidia-announces-quadro-gp100

>>59773407
>with the kind of PSU setup you'd need for that i have a hard time believing it'd cost less
Elaborate. Hawaii's 295x was a behemoth of a card that could still be ran with a decent PSU. We're closing in on 7nm as well, so we wouldn't need that much power delivery. They could also make a very small and efficient chip and scale it however they want or go middle of the road like Polaris 10, using less than 4 of those on one card.
1:1 scaling is obviously not going to happen, but getting at least 80% would still be great.

>>59773436
only amdtards bitch about titans being overpriced
people who buy these consider $1k pocket change

>>59767170
Just a quick question:
>Why is there not <$100 GPU for general desktop use (no gayming) that supports 4k 60Hz?
The only GPUs that support 4k are from the GTX line or their AMD counterparts. What happened to the GT line?

>>59773330

>that webm

JUST

Where's it from?

>>59773516
btw I know you can get one used for less than $100 but I don't buy outdated tech.

>>59773523
LinusTechTip's Rypoo review, at the end of the review you can see it drop to the floor, totally wrecking the motherboard

>>59773330
>purple and blue haired "hacker" woman

yeah

>>59773546
>Rypoo
You guys don't get tired of this shit. Why are you still spouting Indian shit memes? It was amusing when DESIGNATED came out, but damn are you niggers beating it to death.

File: 1483626101925 (they_know).jpg (8KB, 280x180px) Image search: [Google]

8KB, 280x180px

>>59773588

>>59773598
I hate this board.

File: 1483500540504.jpg (31KB, 480x547px) Image search: [Google]

31KB, 480x547px

>>59773616
Good.

>>59773556
Beat me to it you cunt

>>59773523
https://youtu.be/9wJQEHNYE7M?t=13m4s

>>59773494
>They could also make a very small and efficient chip and scale it however they want or go middle of the road like Polaris 10
polaris was a pretty big step up and AMD did manage to bring their cards to around the same level of efficiency as maxwell, but i do not see a reliable multi-GPU setup on anything less than 750w. and this is just for two cards. that PSU alone would be close to $100.

>We're closing in on 7nm as well
define "closing in". volta and navi are sure as hell not going to be on it, unless you believe WCCFtech.

>>59767170
>this is Nshillia's response

>>59774098
>same level of efficiency as Maxwell

Fiji was already there, at 28nm

File: power.png (95KB, 602x364px) Image search: [Google]

95KB, 602x364px

>>59774167
sure is the same

>>59774209
Are you fucking retarded?

https://www.techpowerup.com/reviews/AMD/R9_Nano/31.html

>>59773516
All Maxwell GPUs support DisplayPort 1.2 for 4K monitors with single DP cable and Pascal GPUs support DisplayPort 1.4 for 4K monitors & 5K monitors with single DP 1.4 cable, 8K monitors using 2 DP 1.4 cables

>>59774275
But all of them are >$100. Or 100€ because I'm from Europe. The oldest NVIDIA GPUs I can get new here are either 900 series GTX or 700 series GT. The 700 series GT hit the sweet spot in terms of price but they don't support 4k. Jesus Christ I just want a 4k capable GPU with passive cooling if possible. Where are all the current gen low power GPUs for use in home theater PCs?

>>59774240
and who actually bought nano's?

it was an overpriced enthusiast item that AMD had marketed SPECIFICALLY for it's power efficiency, any nvidia equivalent shit on it.

>>59774356
The Fury is only slightly less efficient, we're arguing architecture here, and AMD had an efficient one on 28nm, with pretty much the same efficiency as its 980Ti competitor.

>>59767333
DONATE TO JILL'S RECOUNT KICKSTARTER, WE CAN STILL STOP DRUMPF!

>>59774518
Here's how AMD can still win the nomination.

>>59767170
>have to have Warner Telecom anaylize your browsing habits to determine a price to sell you at
yeah no thanks

>>59774209
Hawaii, dipshit.

>>59769167
>Have most powerful card in the entire world
>"Objectively worse"

lol

https://www.youtube.com/watch?v=EYlMrxIgoME

>>59768372
Nintendo swooped in and saved the X1

>>59767407
This is actually good, from a 980 to a 980Ti the difference was very similar too.

>>59767407
>Quadro is good at games
Wait what, I thought they were absolute shit at anything other than compute stuff

File: 1486389140242.png (6KB, 210x230px) Image search: [Google]

6KB, 210x230px

>>59769488
>Titan X
>Titan Y

File: absolutely-halal-1227329.png (354KB, 500x498px) Image search: [Google]

354KB, 500x498px

>>59769488

>>59773546
t. pootel shill

File: laughing Dalian.webm (63KB, 240x135px) Image search: [Google]

63KB, 240x135px

>>59769488
>TITAN X axis
>TITAN Y axis

File: 1454019841736.jpg (6KB, 184x184px) Image search: [Google]

6KB, 184x184px

>>59773839
>3:29
>8:44
>that end clip

>>59772611
Very interesting. It's a tricky legal situation; of course I want hackintoshes to get more support but I don't want them to get so much that Apple decides to take more aggressive measures against it.

>>59772191
source?

>>59774240
>>59774356
You could just undervolt normal fury cards to that level or near that level, same with undervolting the 290/390

What does this mean for prices of current cards?

I was about to get a 1060, should I hold off a bit now?

>>59776394
It won't affect shit on the lower end. There's enough of a price gap in the high end that they can release 10 more slightly different card without moving the price much.

File: tfw_no_debt_and_owning_your_home.png (534KB, 1186x1222px) Image search: [Google]

534KB, 1186x1222px

>>59767333
>AMD can still recover the market.
No. They can't. But they can price their cards more aggressively and dominate the budget segment. I'm not sure they want to, though, because selling more cards for less may be less profitable than selling fewer for more.

>>59767441
>>Sell Titan XP for $1200
>>Gimp it
I'm not sure how on earth they get away with this year after year, decade after decade. It's not like they made their older cards slower with updates (like Apple do with iPhones) just once or twice.

I'm glad I'm not into bleeding edge gaming and can use a $200 graphics card to play all the 2 year old games just fine.

The idea of paying $1200 for a graphics card that's outdated in a year just seems silly. It's not like it's six times better than a $200 card. Feel free to prove me wrong, please show me a graph that has the Titan XP performing six times better than my brand new (just bough it) RX 470 8GB.

>>59767543
>fags like Linus
In all seriousness, he's gay, isn't he? I know he has a cute asian wife but still.. idk why but it just seems like he's.. that way?

>>59767642
>Who buys these cards?
No idea. I'm guessing middle-aged men? No way some teenager is going to save up $1200 and spend it on a graphics card.

>>59767750
>so AMD now needs to beat:
>$NVIDIA CARDS
No, they don't. They just need to show up with budget cards that have a (much) better price/performance ratio.

I paid $495 for an EVGA GeForce GTX 1080 Superclocked just last week. I am getting 144+ fps locked to the 144hz refresh rate of my monitor in games such as Fallout 4, Shadow of Mordor, and others.

Spending $1200 on a video card is going to get me, what, maybe another 12 to 15 frames per second more than I am getting now? And how many more fps will I get if I spend another $495 and SLI this card I have now? Another 40 to 45 fps? There has to be another reason why companies roll out $1200 cards. It must be to serve as a proof-of-concept signal to investors of the direction the company is going and what the future holds. Surely it isn't because they expect to sell boatloads of $1200 video cards.

>>59775211
No. GeForce GPUs have gimped float64 compute modules. I think Nvidia's workstation cards have more ram and proper float64 compute times plus an insane price tag. Someone correct me if I'm wrong here.

Sooo. Theyre introducing more high end gpu inti tge pool, furtger deviding their profits for what reason? Forcing people, that generally have a limited budget in the case of getting a top tier card, to make choices that make no difference save for "Muh Framz pur sekond" the marginal difference between the 1080ti and this for the %30 increase in price is just idiotic. They arent pushing any new boundaries.

>>59776930
Yeah, buying this for games is not a good idea. This is for compute tasks that don't need 64bit precision. Titan is for gamers with too much money and programmers/scientists/machine learning engineers.

>>59776869
The titanx and1080ti are a hood 30% faster than a base 1080.

It wouldn't Be A 15fps leap, rather a 50fps jump.

I feel bad for Amd

>>59770451
$1,300 per TFLOP? What a joke if that's not 64bit.

This is Novidya's 'BAWWWWWW!' release because they know Vega is gonna stomp at the 1080, Titan XP and maybe even the 1080Ti.

They are scared of any competition.

>>59773407
Four 7nm GCN5 1024-shader units would have
>clocking capability 1800-1900Mhz
>64 ROPs
>256 TMUs
>4 stacks HBM minimum, 8 capable (8-16GB/16-32GB VRAM)
>2 command processors per chip
>145w power draw
>roughly 2.3x performance of Fury X
>total package size ~450mm2 on 600mm interposer

Waitfags, 2019 is your year :)

TITAN X buyers on suicide watch.

>>59777380
pinche dumbass

>FULL GP102
>doesn't bandwidth throttle on GDDR5X

nobody will need hbm for years

>>59777513
with 16 gcn cores this shit will draw less than 100W, HBM is REALLY lowpower. Maybe it will get even below the 75W

>>59777588
HBM is a Meme

File: 1462966876148.gif (646KB, 512x481px) Image search: [Google]

646KB, 512x481px

>>59777588
>imbligayshuns

>>59773016
Excites me too, Anon. That SHOULD make them much cheaper and more powerful, and no longer necessitate having midrange GPUs as

Nvidia will likely be doing the same. Maybe not with Volta, but soon.

7nm and MCMs will be a new era.

>>59774209
>comparing Hawaii to the 970, and not 980, when that anon said Fiji.
Wew. That's some retardation.

File: 1488511023085.gif (4MB, 250x141px) Image search: [Google]

4MB, 250x141px

>>59777597
you god damned cave dwelling troglodyte horse fucker, THE IMPLICATION IS NAVI 2018 OR FURTHER 2019 ON 7NM WILL BE MULTIPLE SMALL GPUs CONNECTED TOGETHER

NIGGER

File: 1490951246275.jpg (117KB, 645x525px) Image search: [Google]

117KB, 645x525px

>>59774356
>and who actually bought nano's?
Tons of people, since they were the best mITX card.

>>59775064
I don't usually like that faggot, but he's exactly right and the most on point I've ever seen him here.

The Titan XP just came out like 8 months ago, right? Which was pretty shit for $1200, but people bought it anyway.
Then 1080Ti a month ago, which is a nice card.
Now this for $1200.

>>59777038
>The titanx and1080ti are a hood 30% faster than a base 1080.
It's like 20% at 1080p, and still under 30% on average at 4k, isn't it?

>>59777777

>>59770791
>1070
>mid tier
are you retarded?

>>59767503
GTX Titan X, Titan X and Titan Xp

>>59772353
>paying more than $100 dollars for a GPU
Your mom said you to stop maxing her credit cards

File: ScumbagNvidia.jpg (24KB, 583x429px) Image search: [Google]

24KB, 583x429px

>>59767170
next build ryzen + vega
enough of this bullshit

>>59777808
That's considered mid tier now days since GPU power has grown so much.

>iGPU that is decent enough
low end
>can do 1080p 60fps minimum maxed or near maxed on 95% of games
low-mid (RX470-1060)
>144fps minimum for many games or 1440p 60-90fps.
mid (1070, maybe the r9 Fury)
>4k 60fps
high end (1080Ti, but still not really in loads of games)

>>59777938

>ITT retarded OCN posters

>>59767170
AHAHAHAHAHA
GET KEKED TITAN "PASCAL" OWNERS

>>59777968
Ocwen Financial Corporation is a provider of residential and commercial mortgage loan servicing, special servicing and asset management services.

>>59777938
That's like saying a Lexus is a "mid-range" car because Ferraris exist

File: 1489022026189.jpg (34KB, 799x533px) Image search: [Google]

34KB, 799x533px

>>59777980
Yes, Lexus is midrange, poorfag. Way to use a shit analogy that just proves my point.

Most of their cars barely manage 0-60 of 5 seconds, and don't handle that well. There's lots of 20 year old cars that are faster.

>>59767170

Now Nvidia is pulling the same shit as Intel by releasing rebadged products every 3 months?

It's a nice way of turning consumers against you desu desu

>>59777811
>>59767503

Don't forget Titan X BLACK(ED)

>>59778008
Like the rx500...also how titan Xp is rebrand?

>>59777979

http://www.overclock.net/

"Mid-range" could mean middle of the pack performance wise, or what a responsible middle income earner would buy. 1070 could be argued to be "mid range" because it's technically in the "middle" of nvidia's product line. But to me, mid range usually refers to the second definition, usually lining up with where the fps/dollar "sweet spot" is. For someone making a median wage in a western country, 470/480/1060 is the line after which you should be asking yourself "should I really be spending this amount of money on a goddamn computer?"

>>59778079
>>59777990
Forgot to reply

>>59778079
>>59778087
I was the poster agreeing with you, though. >>59777938

File: 2017-04-07 07_41_00.png (62KB, 203x152px) Image search: [Google]

62KB, 203x152px

>when literally *THE* Nvidia fanboy is fucking fed up with Nvidia launching new products with marginal speed improvement every 3 months (like Intel does)

THE GLOVES ARE OFF

https://www.youtube.com/watch?v=ZjJUVsmjIj4

>>59778183

Once Nvidia donates two titan's for "testing" (that they forget to ask back) he'll quickly change his mind.

>>59775211
No, that's a lie. Linus did a video on it a while back, I think.

>>59778320
>>59778183
god you people are so gullible and naive, that fucking shit video is just to create controversy and get views

>>59778183
literally
Gimme free Titan Xp

File: 3mLydMU.png (70KB, 243x200px) Image search: [Google]

70KB, 243x200px

>>59767170

>releases titan X pascal foe 1200$
>everyone calls it titan XP

>release 1080ti a few months later.
>500$ cheaper and basicly the same performance

>releases titan Xp a week later for the price of 1200$

Soooo everyone who bought a titan X pascal got shafted? and everyone who upgraded to the "most powerful" card the 1080ti got shafted as well.


its too funny.

Even JayZ is mad.

>>59778433
1080Ti owners weren't shafted.
It's $700 and actually not terribly performance/$ at that. Though $650 would have been better, sure.
This thing is $1200 for a maybe 15% performance increase.

Titan X Pascal owners that just bought that like 8 months ago were shafted though by both, though.

>>59778183
>intel
>launching products every 3 months
kek
also
>buying titans
>EVER
i reckon a $750 aftermarket 1080ti will either match or beat the Xp, considering its choked by having shitty blower style cooler.

>>59778690
I agree with you, but you're a faggot for not understanding you can replace stock GPU coolers and how great this actually is.

You get a graphics card with a good PCB, good binned GPUs, put on an artic accelero or other cooler, and it shits on "muh best stock custom cooler" model.

>>59778433
>>everyone calls it titan XP
Dumbasses, all of you. You should have just called it the Pascal Titan X instead of being a lazy mobile poster.

>>59775539
>posting meme
gay

>>59768856
>>59768903
>>59768931
Read the thread. He said high settings, not maximum.

>>59780271
No one person said high settings for Vega while another person claimed maxed settings was already possible on a 480.

>>59778913

you also void the warranty on your $1200 GPU.

>>59780417
poorfag.

File: hurts.jpg (31KB, 540x567px) Image search: [Google]

31KB, 540x567px

https://www.youtube.com/watch?v=UtyaGppWknE

vega is coming in may

>>59767387
> Visit Nvidia site
> see this:
> The NVIDIA Quadro P1000 enables me to work at the speed of thought.
> Jason Pohl, Senior Designer, Orange County Choppers

> Visit OCC's wiki article
> Paul Teutul, Jr. - former minority owner/chief designer and fabricator (Fired in 2008)

And the old man still is the CEO.
Fucking brutal.

>>59780484
>poorfag.
Not Watercool Titan Xp for better performance, it you who poor