BREAKING: AMD CEO rushed to hospital after potential poo in loo

>>59775521

xDDD
>>>/reddit/

>>59775659
u mad?

Implying amd even care at $1200 + gpu

how about you be a good goy and SLI that shit

>add 256 more CUDA cores
>51 MHz faster
>$1200

TitanXp(before this one)
$1500


HOW WILL AMD EVER RECOVER?!

>>59776059
with dual rx 580s obviously..

>>59776059
>1500
Here we see how the mind of a shill works. Making up data that's not even remotely true.

>>59776688
not by much
https://www.newegg.com/Product/Product.aspx?Item=9SIA0AJ5GX6544&ignorebbr=1&nm_mc=KNC-GoogleMKP-PC&cm_mmc=KNC-GoogleMKP-PC-_-pla-_-Video+Card+-+Nvidia-_-9SIA0AJ5GX6544&gclid=Cj0KEQjwzpfHBRC1iIaL78Ol-eIBEiQAdZPVKrjeK1MnYLzXd4TzcRZBPW6NxNEnOuE3iWtZQvS5k8kaAs018P8HAQ&gclsrc=aw.ds

>>59776059
Pretty obvious this is strictly for the big e-peen crowd. There are people who will pay almost double for marginally better performance just so they can say they have "the best". Nvidia would have been stupid to let a $700 card be their flagship when clearly somebody was buying these $1,200 Titan X and Titan XP shits.

Now how about them Quadros and Teslas...

>>59776815
>Wanting to pay $10,000 for a Quadro card
hahah idiot

>>59777078
I mean that they take the cake for overpriced shit. It's insanity.

File: 1462944576048.gif (22KB, 128x128px) Image search: [Google]

22KB, 128x128px

>>59775521
So it's just a relabeled P6000 without the drivers?

youtuber and shill freak out, were my new Titan Xp

>Nvidia so bored of waiting for AMD that they're competing with themselves and just announced the world's fastest GPU twice in like 2 months

Nvidia is beating Nvidia while AMD is losing to AMD.

>>59775521
You are looking at this wrong, the question isnt will amd recover, its why did nvidia put this out in the first place?

I mean its a chip they can sell for 5 grand, why sell it for sub 1500$, why would a company, generally not known for pushing hardware out till it damn well needs to happen to stay relevant push out this card?

My guess is vega ant 12.5tflps single precision on passively cooled pro units is coming to consumers higher clocked because it can be cooled better. My guess is also that nvidia did some math in their head, and went with polarises efficiency and found out that vega at at least as good at a 1080ti if nothing improved whatsoever.

The one big question in the air right now is will this be another fury where on paper, fucking world beater, in practice, bottlenecked to hell and back with severe saturation issues. Or is this a case of we put all our manpower into getting our efficiency and utilization down pat and we surpassed expectations.

Sad to say, but at face value this is interesting, but once you ask why, vega gets more and more interesting.

File: maxresdefault.jpg (52KB, 1280x720px) Image search: [Google]

52KB, 1280x720px

>It's over, AMD. I have the high end.

File: 1476429960915.png (10KB, 264x153px) Image search: [Google]

10KB, 264x153px

>>59778410
I actually did some calcs off reviews, and scaled up performance/teraflops to 1536mhz 4096 shader configuration.

1 would be 1080 Ti equivalent.

This is assuming no per clock improvements, just fixes to allow better scaling.

lol Buthurt Titan/1080ti owner is funny. i thought those people got money to burn

>>59778500
I went flops based and amd to nvidia equivalents.

the worst amd ever was, they needed 33% more to match nvidia, polaris is somewhere in the mid to eirly 20's for a 480 and 1060, based on that vega (the pro card clock rate alone) puts it at either just above a 1080, or closer to a titan x then a 1080

If amd gets that efficiently delta down to 17-15% its better then a 1080ti

But this is all ignoring that the card could be clocked significantly higher

My thoughts on the gpu are this.
1) they got a new render method that should saturate the gpu better, possibly render overall better
2) they overhauled god damn everything on the card, so no ipc gain would be stupid to think, but its yet to be shown clearly
3) even in the potential down the road 8gb vram is not good enough scenario, they have plans/working counter measures so your gpu is not useless like the 970 in re7

There is some other shit I cant think of off the top of my head, but I cant consider this just a dest bed card now like I was before. Shit is still a test bed for navi, but its also got enough real world improvements to not only be a testbed.

>>59775521
ye, gp100 has been out forever like 2015 or 2016

>>59778454
yea, nvidia users never considered amd an option and get belligerently pissed when nvidia charges a high amount for their cards because amd isn't competing.

Till amd has an mcm gpu, they could beat out nvidia by an order of magnitude, but people would still get nvidia at this point, but the mcm gpu, that allows amd to effectively have a low end gpu to above what nvidia can offer for less then nvidias current titan lineup, this is where amd will likely be un ignorable anymore. 1 die allows them to make one gpu across everything, driving the cost of the low end down, along with driving high end down as well, see ryzen, an 8 core 6900 equivalent an almost all respects, a few cases below, a few cases a 6950 competitor got put out for 320$

now imagine this for gpus, granted, with amds cache thing, they may very well get away with having 2gb of hbm vram or less on low end gpus, effectively taking a 200$ card and pushing it to 120-150$ levels, if amd switches to hbm exclusively, they dictate 100$ of the price as all aib partners will do is put on a cooler and some vrms, they won't be able to mark up due to ram anymore.

>>59775521
>They name it the thing everyone was calling their previous card because it was too confusing in conversation and marketing
Oh well at least i'm sure it's intentional now.

>>59778795
lets just call it the titanX15, titanX16 and, titanX17 to avoind any confusoin

>>59778366
the truth is that nvidia has been competing with themselves for a while now

Fury X is AMD's fastest single gpu card, and lets see, its slower than:

980Ti
Titan X Maxwell
1070
1080
Titan X Pascal
1080Ti
Titan Xp

im not surprised in the slightest that nvidia is fucking around so much with Titans. Even when they had proper competition, Titans were still terrible value for money, as the Ti versions would always come out soon after, and would make Titans feel pointless.

I was watching jayz2shills most recent video, where he is ranting about this, but its clear that the only reason he's mad is because nvidia didn't send him Xp's for free. The shill got BTFO.

>>59778825
You fool!
Now the 15 and 16 series cards are coming up soon, they'll just skip a bunch like with the 300 and 800 series just so they can continue this.
Nothing will stop nvidia on their quest to make sure no one fucking knows what the hell graphics card is for sale on those cheapo and used websites where they don't give specs sometimes.
Hell even amazon does it sometimes.

>>59778547
I don't the reason why 1080Ti owners would get butthurt. It's stupid. The card is $700-$750 and with aftermarket coolers im sure it can easily match or even beat the $1200 new Titan. It's so much better value for money its not even funny.

Of course if those people only buy graphics cardsd just to say "I have the fastest one" then whatever, they deserve to get facefucked by Nvidia.

>>59778909
Those people definitely exist.
I remember when the 980 came out a bunch of first gen GTX Titans flooded the second hand market.
The value of the titan was wasted on those gamer shits.
They were selling for like $600-$700 a pop too which must have hurt nvidia's quadro sales because that is way cheaper than an equivalent performance quadro is even now and the titans never had FP64 again.

>>59778871
jayz2shills was "mad" about every titan card

>>59778972
the real quadro killers were the original titan and the titan z

>>59776688
Prices are very regional. You do laugh when Aussies pay literally twice as much or more for any computer part or accessory, right?

File: 1491192298441.jpg (105KB, 337x434px) Image search: [Google]

105KB, 337x434px

>Meet Jen-Hsun in person at deep learning event at Stanford
>We have a short conversation about Titan X (Pascal Titan was announced last year at that event)
>Casually make a joke about AMD
>He gets serious and gives me a weird look and avoids the subject

What did he mean by this?

File: nvidia-tattoo-09242010.jpg (57KB, 600x603px) Image search: [Google]

57KB, 600x603px

>>59779143
Graphics cards are serious busines, anon

Bow down to the new king PASCAL XP

Who would 2 years ago that amd will be such garbage for poorfags. They have only lowend cards now. Kek.

>>59779143
He amd fanboy, don't you know

>>59779492
>radeon pro duo
>16k tflops
>Still fastest graphics card there is

>>59778972
People that buy quadros and actually need quadros don't buy second hand gpus.

>>59779608
Not everyone who buys a quadro is using it for mission critical 24/7 shit.
Sometimes it's just nerds doing some compute shit for a personal project and you can even justify using it for something more official if it isn't an ongoing thing and you're tight for cash.

In other news Goldman Sachs has put a sell advisory on AMD stock citing the strength of nvidia products and AMDs inability to complete at the high end

>>59779677
Aka ABANDON SHIP THE STICKY OLD SHIT BARGE OF AMD IS SINKING

File: ff.jpg (452KB, 800x960px) Image search: [Google]

452KB, 800x960px

Oh boy and it will be prohibitively expensive for a super small niche market.

Who cares about "DUDE FASTEST GPU LOL"

>>59778874
I had this with a 560Ti back in the day. I purchased a version thinking it was higher clocked. It's a big fucking scam by the jews.

so...
buy a 1080 for half the price
install cooling
crank to 2 ghz
????
same performance and save $500?

>massively expensive card few care about
>massively expensive card fewer can afford
>a big deal is made of it

Why?

>>59778871

No competing with yourself is not what nvidia is doing.

the 980ti came out because of the fury, nvidia was competing with it, and they largely won outside of a few games that can saturate a fury

The titan x is nvidias 'we want 1000$ for this' gpu, they will put them out no mater what.

the 1070 and the 1080 were to have gpus in categories that had no competition, with the titan pascal for that next 'retards with to much money to spend'

Nvidia held the 1080ti so they could steal amds thunder again when they announced vega or talked more in depth about it, but they decided they wouldn't wait anylonger and released it.

Its fully possible around this time nvidia got wind of what consumer vega would do on paper, and decided they want to be nowhere near it.

then the titan Xp comes out, can you tell me what on gods green earth nvidia would put out the 1080ti, then the titan xp in rapid succession rather then put out the 1080ti, wait 3 or 4 months for everyone who was getting one asap to get one for more power, then release the titan xp so they could potentially triple dip the high end buyers?

this screams nvidia got wind of amd's vega, and are scrambling for something so they can retain top gpu bragging rights.

If you honestly think nvidia competing with itself looks liek this, then you are fuckign retarded.

>>59779872
Nvidia owners seem to believe that this card existing makes their 3GB 1060 better somehow.

>>59779143
because, amd is nothing to joke about.

Look at every single gpu they have put out on gcn, they are absolute fucking monsters, what was it, the 7970 put up a fight and could max games out sans aa till what last year there were a few it would no longer 60 at 1080?

The fear that must be in him about the day amd got drivers day one ready, or had hardware that was fully utilized from the get go must scare the shit out of him.

He cant let himself get caught up in amd is behind, look at fucking intel and see the absloute wreakage they left in their wake.

Intel has to admit, they are competitive
intel no longer has a 300$+cpu that is worth buying
Intel no longer has a 150$+ cpu that is worth buying
Intel is on their knees thanking god for motherboard manufactures shitting the launch window bed.
But intel knows it wont last
Amd is faster for watt, ultrabooks could legitimately ditch intel hand have a better cpu, and soon a cpu+gpu combo that will overtake them
The motherboards are getting patched, windows is updated, and ryzel looks better and better each time and there is nothing they can do.

They have literally no answer for 3 years at best, 5 at worst, and amd has yet to blow their lineup wad. workstation/prosumer 16 core cpus, quad channel memory, server 32core octa channel memory. all made at lower costs, for better margins than intel, and unlike intel, amd has no fat left to trim.

Nvidia must be thanking what ever god they worship over there that they learned the threat of amd back with the 4870 and the 5770+ so at the very least they aren't nearly as fucked, they will be deep dicked with the next gen, but they may be far enough ahead to stave off vega stealing the top gpu spot.

>>59779677
ah, that's what it was about. so they ignore the biggest market amd is re entering, the server, and think that amd was only about the gpu?

>>59775521
While in hospital, the nVidia CEO reels in a whiteboard, and made her bubblesort all the letters in her medical report

>>59779143
Because he's a fucking CEO and must live to social standarts, retard. This is why you won't see nvidia ads on the level of commie rebelllion poor volta.

>>59779143
Holy fuck you're autistic

File: 172.jpg (29KB, 645x289px) Image search: [Google]

29KB, 645x289px

Prepare your TearWorks engines Nvidiots

File: no drivers.png (19KB, 1190x119px) Image search: [Google]

19KB, 1190x119px

>>59780074
fix your OGL poojeet

>>59779884
>this screams nvidia got wind of amd's vega, and are scrambling for something so they can retain top gpu bragging rights.

lol no, this is just nvidia repeating the same pattern they've been doing. they did the same with: og titan -> 780ti -> titan black

>>59779982
>Look at every single gpu they have put out on gcn, they are absolute fucking monsters,

every GCN gpu has been an absolute fucking joke. you are either shilling or too young to remember when AMD did good engineering.

File: thinkinglett.png (33KB, 783x295px) Image search: [Google]

33KB, 783x295px

>>59778322
Yep.

>>59780200
Can't wait for vega and amdtards vanishing from this board for at least several days back to /r/amd.

>>59780116
>pcsx2
>opengl

lol what the fuck for, even? Just use the software renderer, it fixes a shitload of bugs. Or even the DX one, if you are that anal about upscaling 15 year old games.

>>59779982
>Amd is faster for watt, ultrabooks could legitimately ditch intel hand have a better cpu, and soon a cpu+gpu combo that will overtake them
Strictly speaking, since Ultrabook is an Intel branding/marketing thing, no they could not. Manufacturers wouldn't be able to call them Ultrabooks anymore.
Branding is a bigger deal than you might think, and losing it could mean losing a lot of normie sales. I don't disagree with you that AMD could be better than Intel in thin and light laptops by now with Ryzen, but Ultrabook manufacturers can't just ditch Intel because they would be ditching the Ultrabook branding.

The most well-known example is probably MacBooks. You might think of these as "ultrabooks", but they are not Ultrabooks. And they never will be, because Apple doesn't adhere to Intel's list of specifications, particularly in their choice of WLAN cards.
Apple also has no need to slap an Ultrabook(TM) sticker on their laptops to get sales, but other companies might not fare as well if they don't have it.

Because of this alone, Intel sadly has that entire market locked the fuck down. If you're not Apple and you don't have an Ultrabook sticker, good luck selling your laptop to normies that see it on display between MacBooks and Ultrabooks, and see the sticker as some sort of guarantee for quality.

>>59779905
I'm an Nvidia owner but I don't give 18 cuckfucks about Titan shit. Samposter you replied to by the by.

>>59780285
t. amdrone
also DX renderer is a joke compared to OGL in terms of accurace

>>59778500
>>59778561
Assuming nothing between Fury X and Big Vega, just the clock speed increase, we get roughly 9% better 4K performance compared to a stock 1080.
>.75 * 1.452 = 1.089

Now, Vega has a shitload of awesome enhancements that just naturally should provide a major uplift in performance. L2$ tied ROPS, tiled rasterizer, better compression features.
The L2 linked ROPs and tiled rasterizer alone practically completely eliminate the RBE/ROP bandwidth issues FuryX had.
Lowballing a 15% performance increase and assuming a 10% clock headroom over the MI25's 1526Mhz we get
>.75 * 1.68 * 1.15 = 1.449
Or nearly 45% better 4K performance than a stock 1080.
Which beats a stock 1080ti by about 10%

>>59780355
true, they would need to lose the ultrabook moniker, but I was referring to it as a form factor instead of label itself. with pc's, especially in more recent years, intel inside holds less and less weight.

so a laptop manufacturer gets to ditch intel's strict rules on hardware, and instead follows amd, who may impose some higher quality parts, but may actually make it cheaper to manufacture then intel side even with intel's deals.

>>59780487
Eh, my math was wrong

>.75 * 1.6 * 1.15 = 1.38
(1050 / 1680Mhz)

Or 38% better than a 1080, roughly barely beating a 1080Ti

>>59780200
so the 200 series relative to when it came out was bad and didn't crush nvidia? the 7000 series didn't last FAR the fuck longer than it had any right to? the 300 series I can admit got dicked by process node falling through, it and the fury were suppose to be 20nm, got to take what you can get, but at the same time, when something uses more ram, those cards still keep up while nvidia need to be turned down, and in the case of re7 REALLY turned down for the 970, and so far polaris has been bad?

gcn's first major overhaul is coming with vega, till now its been relatively minor iterations and small hardware decoders added with some die shrinks. but looking at nearly every single card nvidia put out that amd had a gpu out for, or responded with a gpu, sans the 300 series which caused amd to just terminate a contract with tsmc for fucking them that hard, and thing nvidia came out ahead... thats stupidity

>>59780487
>>59780538
keep in mind, you are comparing it to a fury, a gpu that got fucked over by being built for the 20nm process that fell through, and had to be retrofit onto the 28.

id make the comparison with polaris as a baseline because I don't believe if amd was given time, they would put out a gpu that is less efficient.

>>59780596
I'm comparing it to a Fury X because it has the same number of ALUs, ROPs, and assumedly TMUs

The only thing Polaris really has over Fiji is better color compression and the geometry discard unit.
When testing different GCN revisions it was found that Polaris really only has ~3-5% better per-clock performance, and usually none, compared to Tonga (small Fiji), when you factor out tessellation.

Anyways, worst case Vega is a bit better than a 1080. Best case it is about equal to a 1080Ti.

I know what I'm buying for my birthday.

>>59778500
The same way fury X improves over the R9 390x. Oh wait.

>>59780662
The fuck are you smoking, nigger?