Thoughts? will it succeed? or will it flop?
it'll probably do good. unless AMD pulls something better or they all explode in everyone's computer, it'll probably do good.
>>54521441
> do good
Do well.
>>54521379
A gaymer card for poorfags.
Anyone with self-respect will wait for the Ti.
>>54521379
>10% more powerful 980ti with slightly better async
Overclocking will be impressive but its really nothing new
Wait for volta if you have a 980 or better
>Nvidia GPU
>flop
Did you think you were posting an AMD card or something?
>>54521480
this
>>54521480
If my GPU wasn't six years old I would wait, but I'm done with waiting.
>>54521379
its gonna be weaker than the 980 ti at launch. Nvidia will freak out over it and patchnerf the 980 Ti for awhile till they release more efficient drivers for it, then the 980 ti may or may not be restored and then the 1080 will still only have like a 5-10% lead on the 980 Ti.
the 1080 Ti, if there even will be one (Doubtful) will probably beat the 980 Ti by 15-20% tops.
I guess I really lucked out getting an R9 290 in early 2014 for $215
More than two years later and there still isn't a $200 card that can beat it
Moore's law is coming to an end
>>54521706
>I like making stuff up
>>54521706
Whaaat? Nvidia don't gimp their older generation cards to make their new ones look better than they actually are, right?
>>54521617
Same here, I have a 670 and I want something new already
yet another one of these threads, gotta keep the "awareness" at a high lvl right?
>>54521379
>Will it flop?
yeah, it'll flop.
gflop, that is.
>>54521832
Nvidia doesn't seem to gimp previous card performance despite what conspiritards repeat. http://www.bytemedev.com/the-gtx-780-ti-sli-end-of-life-driver-performance-analysis/
>>54521379
As long as it doest have coil whine its fine by me.
>>54521832
780ti in recent driver updates fallout 4 benchmarks show it performs worse than a 390x even though it should easily win at 1080p(even an AMD guy like me knows this)
They have gimped the 780ti over last few driver releases it /is/ a thing and another reason after a decade using nvidia inmay never go back.
>>54521924
http://gamegpu.com/rpg/rollevye/fallout-4-test-gpu.html
http://gamegpu.com/rpg/rollevye/fallout-4-beta-patch-1-3-test-gpu.html
The problem is most people don't bother benchmarking games after they've been out for this long.
>>54521480
>A gaymer card for poorfags.
how is that?
simply buy this one, have fun with it for at least 6 months and sell it after 1080TI/Titan announcement
>>54521379
Considering it doesn't have any competition it will succeed.
>>54521995
>280X beating 780 Ti
>>54521856
??? My 670 4gb plays everything on high to ultra still. Youre doing something wrong bud
It'll FLOP
9 x 10^12 times per second, to be precise
>>54523858
>everything on high or ultra
I have a 770. Maybe if you're only playing games like DotA or CSGO.
I average at 45 fps in Witcher 3 with tweaked settings. I get maybe 60 in some caves and that's it.
Hell, even Dragon's Dogma drops to the 50s or less outside of Gransys.
Besides, these new cards are great if you own a 120hz monitor and want to play the majority of your games on a framerate suited for those.
>>54523996
Shit, I meant Gran Soren of course.
>>54521706
>le gimped cards meme
kill yourself tinfoil faggot
>>54521546
>what is the 480
Here it is boys. Benchmark at 2.1 ghz
>>54521480
>if you wait 9 months you will get a better card
But if you wait another 9 months there will be a better card out.
Oh but what about the better card released a year after that?
Better wait forever and never buy any tech product, then.
>>54524224
didn't the same website leak benchmarks shortly before the announcement that claimed it's about the same or even a bit weaker than the 980 ti?
fuck this site and the false information they keep spewing out. I'll wait for a credible source.
>>54524682
I don't think so, no.
>>54524224
>20% OC gives 20% better synthetic results
Surprise!
When will people realize that cherry picked press samples OC a lot better than the cards you can actually buy in stores?
>>54522365
>not arranging your gpus on the wall as trophies'
poorfags confirmed.
>>54524720
4870 x2 was the only card I didnt sell because it looked too good at that time. :(
It's still laying besides my mirror.
>>54524561
Non-poorfags are on the Ti cycle.
should i sell my 390 and save for new gpu's?
>>54521480
Is it really that necessary to wait for the ti? I have a 760 and wanted to upgrade to a 980ti but I've waited until this generation and am dying to upgrade, how much longer is it typically before the ti comes out?
>>54525471
it's not gonna come out for another half a year at least.
>>54525196
super non poor fags upgrade for every incremental increase
>>54524224
>110% faster than a stock 980 with an overclock
Nice. This card has good oc potential with just an 8 pin as well. Can't wait to see how the board partner overclocking cards perform.
If the 1070 gets anything like this performance over a 970 I'm sold.
I'm still on that 670 buzz.
Looks like it could be upgrade time. Although not like there is any game that needs it
>>54525196
>ti cycle
>the poorfags you hate so much will have a faster card than you for at least 9 months
>>54525572
I think I can probably wait that long, the 760 is fine for 1080p at the moment, Ive been a poorfag all my life until now so I want to build the best, i7 and 16gb of ram maybe. It won't just be for gaming, it'll also be my work station for developing games.
>/v/ermin would rather be early adopters and take the nvidia dick when they know full-well that nvidia is going to need to do damage control to the 10XX cards within the month and will release versions a month later that run cooler/faster
Which card should I get for 1440p 144 Hz gaming?
>>54525674
gonna need a powerful rig for developing that sweet retro hipster chip tune game
>>54525695
1070
>unironically buying memevidia cards
what the literal fuck is wrong with you people?
>>54525715
Kek, we have to start somewhere, we are planning on getting a vive between us and developing games for that by the end of summer though
Only morons who have more money than brain cells buy videocards that cost more than $300 for gaming only.
I'm waiting for Polaris or Vega; they'll either have better performance/price, or they'll force a huge pricedrop or a better priced new card in response. Then I can make a proper decision on whichever to buy.
>>54525759
And on that note, I'd rather buy a new SSD instead... it'd give a way bigger increase in everyday productivity, eg. it would be more worth it.
>>54525759
>poorfag detected
Just kidding. I've never spent more than $240 on a graphics card. The diminishing returns on graphics cards after about $300 has always been way too pronounced for me to justify spending any more.
>>54521446
i like you.
I can't wait to upgrade from my 660
>>54525788
Thats exactly how I felt when I had less money. But now I'm just like fuck diminishing returns I can afford it
It will succeed for the targeted market.
There is no competition.
I am more interested in the GTX1070.
I hope it reaches GTX980 Ti performance.
I can return my 980ti within 23 days, but I need some gameplay benchmarks soon
>>54525804
> i
I
Any word on when partner cards for the 1080 are coming out? Nigabyte in particular, they're the best 3rd party I heard?
are they gonna do an upgrade program? only bought a dang 980ti around christmas
As always the 1070 will be the most important one.
Look at the 970, it is the most used card on steam.
If the 1070 can match the 980ti I might get one to replace my 780ti.
I'm not that interested in the 1080, too expensive for what it probably offers
>>54521379
It'll flop and bankrupt Nvidia.
Of course it's going to do well, you fucking moron.
>>54524224
>2.1 ghz
This can't be fucking real
>>54521379
What is this founders edition jewery?
Also why are they killing 4 way SLI? Is the SLI necessary when you are making a neural network out of four video cards? Is this a way to fuck those people over so Nvidia can jew them more or are they just acknowledging that it has no place in gaming.
>>54526124
It's so that people buy a Quadro card instead of simply getting gaming cards in SLI for workstation loads.
>>54521379
Oh it will flop............
....it will TERAFLOP!
Both pascal and polaris will revolutionise the gpu market once they hit. The jump from 28nm o 1/14 nm FinFET is huge enough to make it a big leap in performance and efficiency. The architecture of these new cards may be complete garbage, but they'll still be strong because of the die shrink and node process.
Know how AMD and nvidia keep talking about how it'll bring 3x performance and 2x lowered wattage compared to previous generation? This is mostly due to FinFET.
>>54526039
>gigabyte
Kek
>>54526148
Nvidia cards shutdown if they notice they are in a virtual machine. I have wondered if that wasnt for the same reason. You can get around it but it costs you like 5% performance.
>>54526188
How is the architecture garbage? Pascal is basically Maxwell with some improvements and Maxwell was a fantastic architecture
>>54526192
I thought the Gigabyte 980 was the best one, no? Anyway, any word on partner cards?
>>54524720
>buying trophies
lmao look at this millennial here, never won anything in his life
When will I be able to buy a bunch of cheap 980ti's used so I can play with my own neural network?
>>54526209
Evga all the way. Partner cards should be out in a month or two from initial release
>>54526205
You misread my post. I was saying if the architectures were garbage, it's still a relatively safe purchase. These cards will dominate anything out in the gpu market right now. The 1080 was shown to have benchmarks similar to the radeon pro duo, which is a dual radeon nano put in one card running in crossfire. That's seriously impressive and is a testament of how amazing FinFET is over 28nm.
>>54526248
Ah I see
>The 1080 was shown to have benchmarks similar to the radeon pro duo
Source?
>tfw might actually buy a NEW gpu for the first time since my 560 ti
the only question is 490 or 1070
no OP, it won't flop unless it literally kills people. its performance may not by 2x980Ti like Nvidia implied, but it's still better and new, so it'll sell well
I'm looking to upgrade my 2gb 7850 in Q4 for battlefield :ww1. Hopefully amd has their selection by then. The card has held up damn well.
>>54526271
I actually can't find the article right now, weird. It released benchmarks of the 1080 OCed to 2.1 gh and it achieved a score of 7300 in firestrike or something like that, while the radeon pro duo got 7500 if I recall.
>>54526038
Nevermind.
>>54526292
490 imo. cheaper, will run anything you want a 1080p, assuming that's the resolution you want to run things at. it's not a high end card by any means, nor is the 1070, bu it'll work perfectly for 1080p.
>>54526410
>490
>for 1080p
I was hoping for 1440p at least. a 390x can destroy most games, if not all, at 1080p
>>54526234
Is it worth waiting? Or is the reference card usually fine?
>>54526471
well amd said that the 490 is aimed at high end gaming laptops and midrange dekstops, so i think 1080p would be the best resolution for this card considering 90% of laptops are 1080p. it'll probably perform decently in 1440p too, but don't try using it in 4k, it'll probably be unplayable.
>>54526494
It's very much worth waiting. The founder's edition is basically the card itself with no added features, but it just has a glossy, metal plate to make it look cooler. You are essentially paying $100 extra for a metallic backplate, which is a complete ripoff. The 3rd party cards will have more features, better coolers, backplates etc and be cheaper.
>>54526104
They already announced that it'd be overclockable to 2.1ghz
What's actually impressive is that according to EVGA's PrecisionX tool, the card only runs at 67°C while overclocked at 2.1ghz. Pic related.
>>54526540
>blurry a shit
>only 61fps
You gotta be kiding me.
>>54525695
R9 390
>>54521379
>no full DX12 featureset
trash, but fanboys will still buy it, so it won't fail financially
>>54526632
>talking out of your ass
>>54526632
>muh directx 12
I'm actually pro AMD but my god, AMD shills are the fucking worst. They never shut up about anything, even if it's something which has happened literal decades ago. I still see AMD fangays go "lel nvidia fermi housefire cards and woodscrews xdXD" when fermi was like what, 10 years ago? Holy fuck.
Wake me up when they've released 1180, I just bought a 970 because I got tired of waiting
>>54526668
The irony in this is that the latest AMD cards are literal house fire tier
>>54526662
it's already known that it only supports 12_1
>>54526668
>AMD shills
I've only ever had 1 ATi card. but I guess criticizing nvidia fuck ups makes me a shill, huh?
>>54526702
source
>>54526702
So the latest and most relevant feature level?
https://en.m.wikipedia.org/wiki/Feature_levels_in_Direct3D
>>54526717
No, but criticizing fuck ups made 10 fucking years ago, makes you one.
>>54526858
>10 fucking years ago,
Nvidia cards being crippled by DX12 is 10 years ago?
>>54526858
7 years and never forgotten
>>54526868
No, it's 2 years ago, but it's still valid. Pascal now fully supports directx 12, stop trying to jump around and backpedal the discussion, you know exactly what I'm trying to say.
I own a used 295 right now, but seeing you brand loyalty faggots talk about shit like wood screws, 3.5 vram and shit which happened ages ago infuriates me, because you guy represent the kind of people I hate the most; holcaust sympathisers who can't shut up about how le jews had it so bad.
>>54526554
It was the high poly reference model with DoF and other artistic BS.
Go watch that NVIDIA keynote and they explain it.
>>54526907
>it's 2 years ago
it's actually recent, check out the DX12 game benchmark that showed the 390X matching 1080's performance and the 1080 being beaten by Fury X
>>54526907
Alright there Terry Davis. Also
>3.5
>ages ago
Fuck off
>>54526965
Nvidiots have the memory capacity of a hamster when it comes to nvidia fuck ups
>>54526998
Nvidiots. When will they learn?
https://youtu.be/8lTGk0R12gs
>>54526944
>check out the DX12 game benchmark that showed the 390X matching 1080's performance and the 1080 being beaten by Fury X
What benchmark
http://www.techpowerup.com/222450/more-polaris10-and-polaris11-specifications-revealed
gg
>>54527070
So polaris 10 really is inferior to the new 10xx series cards. Literally what everyone was expecting.
>>54523996
>120 hz monitor
There's your problem, you fell for a stupid gimmick, now go buy the most expensive card to get your gazillion FPS on.
>>54527201
he won't even get that with a 1080 with recent games. unless all he plays are games made for toasters
>>54527070
>Industry sources revealed
>SP count of 2,048
Bullshit. There was some leaked benchmark that confirmed 2304 SP.
>>54526944
If you mean these, then you're retarded
>>54527255
I already get 200-300 fps in csgo with my 770.
I get 45 avg. in Witcher 3, a 1080 is going to more than double my framerate. I'm already happy if I'm above 100 fps.
>>54527201
>120hz is a gimmick
(You)
>>54527335
>csgo
hence, games made for toasters.
>>54527335
you don't run Witcher 3 maxed out at 100fps mate.
>>54527291
>leaked speculative benchmark which could have been any experimental hardware
I'd rather believe TPU instead of a speculated, unverified benchmark.
>imagine being an amdshill and seeing these benchmarks completely obliterate anything AMD can put on the table
Must be truly suffering
>>54526998
Rodents actually are said to have good memory capabilities.
>>54527497
Kek. They're used to it though. This has happened the last few gens.
>>54527354
not the same anon but, I do run the witcher 3 maxed out, (minus hair works due to it looking like SHIT imo on everything but the beasts) at 89-98 avg fps
>>54526539
To add on to what this guy was saying, the main selling point of partner cards is the extra fans. The reference cards have one fan all the way at the fucking end, which makes it work harder and thus, higher temps. That's why the GIgabyte 3-fan card, though goofy looking, provides the best performance.
EVGA's my go-to Nvidia card partner, though, because they have a greater variety of clock speeds depending on how many NEETbux you wanna spend. It's gimmicky, but that's what you get with Nvidia.
>>54527580
and Ubersampling? I think not
who the fuck cares about directx, it's nothing more than a meme api that should've died years ago. mantle and vulkan are the future.
>>54527070
That's nothing but speculation.
>It turns out, that the performance-segment chip, which the press has been referring to as "Ellesmere," could feature 32 compute units (CUs), and not the previously thought 40.
Yeah, it could feature only 32 CUs, but it could also feature more. They don't know anything and published this article without an ounce of content to generate clicks.
Just wait for the reveal.
>>54527676
DX12 =/= comparable to DX11/10/9
it's a low level API giving devs more direct control of hardware.
>if nvidia doesn't support it, it doesn't matter
kek nvidiots
>>54527683
>Just wait for the reveal.
may 27th, only a couple of weeks away
>>54526494
*Founder's edition
>>54527690
it's also the worst ai to use in terms of performance, has terrible multi threading and multi core support and developers hate coding for it.
Doom managed to achieve 200 fps on ultra on vulkan, while it only achieved 100 fps on directx 2, it's not even a meme anymore, directx is just fucking shit. It used to be good when the only alternative was opengl, but now that vulkan exists there is no reason to cling to it anymore.
>>54527734
full of shit: the post
>>54527580
Same. I have everything maxed to ultra except for hair works and foliage distance because no one needs to load foliage they can't even see. I have it at high or medium and can save like 15 fps in some cases. I average between 90-110 fps.
wait for 1070/1080 or buy a 980ti now? i just wanna play some games now
>>54527734
It has good support and developers don't hate coding for it but nvidia and even amd have put so much into vulkan research which leads me to believe directx might become 2nd rate if all gpu manufacturers and devs decide to hope on the vulkan train.
>>54527757
fuck off bill gates.
>>54527799
>wait for 1070/1080 or buy a 980ti now? i just wanna play some games now
If you have the money, and lack the self control then buy now. Anyone with half a brain would wait though.
>>54527497
gtx 1080 was nail in the coffin
>>54527070
woah, is nvidia that scared? I haven't seen this much anti AMD articles in quite some time
first they keep saying it's just 470/480(no x) =no source
then they say it has production problems = no source, debunked by AMD representative
now it's not enough CUs
This is getting ridiculous.
On the other hand founders edition, metal layer talk by real specialists not some random tpu poster, gddr5x bottle necked by small bus, lesser than expected performance for such a huge node shrink.
>>54528024
>calls out 'false articles'
>starts damage controlling instantly with points which have been proven wrong
Wew try harder
>>54528024
>110% extra synthetic performance over a 980 when maxed out
>less than expected performance
kek spotted the butthurt amdrone
See
>>54527497
>>54521379
So should I get the 1080 or settle for the 1070? Will I kick myself later for not just getting the 1080?
>>54521379
780? 780Ti is better, but scrap drivers.
>>54527580
>>54527773
>1080p
>2016
>>54528073
>>54528127
Missing the point as usual. My point is that not a single article was made when those rumors circulated full force, not even slight mention - but AMD gets full coverage everywhere on speculations of nobody knows who.
>>54528293
it's definitely weird. amd is hushed up about polaris and vega, they haven't released anything aside from marketing buzzwords like "it can run hitman at 60 fps!!!". Im wondering if amd has something amazing in store or if polaris is trash and they don't want to let it out yet.
As soon as AMD drops Vega """""in October""""" Nvidia can just drop theur 1080 Ti and demolish them. All they have to do is keep one gen in their pocket for whenever AMD plays their hand and they'll always win. If they get eind of AMDs performance they can even play their Ace in the hole early knowing they can't be beaten.
>>54528181
Its always cheaper to move down than up, if you so choose.
>>54521379
Yes.
>>54523896
Exactly what I thought when I saw this thread.
How do I become a shill?
I think life would be much easier when I only have one item to possibly buy on the market. I wouldn't have to waste time researching and just believe any random rumour or benchmark I want.
Is it easier or does it hit hard when your teams product turns out to be shit
>>54526292
>nvidia implied
Nvidia implied that it's faster than sli 980s not 980 tis. It certainly is faster than what they mentioned.
>>54521379
Whst resolution does it target? For 1080p it's overkill. For 1440p its ok but only 60fps not the 144hz that people want. For 4K forget it. You gotta ask which monitors people have. Most are still 1080 60 fps and don't need this.
>>54528536
Just shill Nvidia. It's always good.
>>54528570
they want you to get 2 for 4K 60 FPS
>>54528536
honestly, the pattern for the past few years always has been: nvidia is at the top of the performance benchmarks, while amd is the runner up. this is not to say amd doesnt make good cards, they do, there's only like a 10-5 fps difference between amd and nvidia, but shills will blow it out of proportion and say "KEK AMD CANT REACH 60 FPS LIKE MY NVIDIA CARD, ONY 55 FPS, AMDONE XDDDDDDDD"
TL; DR just buy whatever the fuck you want, /g/ is 90% shills.
>>54528480
1080ti will not be released this year, if it will exist at all. Big Pascal is most likely branded Titan and 1180ti with the 1080 being rebranded as the 1170.
Vega 10 is not the Titan/x90ti competitor either. Supposedly Vega 11 is.
Vega 10 will probably be the same market position as Hawaii, which was bigger than a 680/70 but smaller than a 780/Titan
>>54521379
Nvidia make pretty good cards, but they have the worst, most aggressive fanboys in the world. Never have I seen a group of people so deeply embedded upon a companies dick
>>54528585
That's all 1080 is good for in my mind. They're being tight lipped about how well the 1070 oc's. I bet it's going to be the single gpu to get. Just slower RAM. Core will still probably be 2ghz+
>>54528629
core speed will probably be the same
it's larger maxwell so expect it to be cut down in the same way as the 970 with less 3.5 gb nonsense.
>>54527070
http://wccftech.com/amd-r9-480x-470x-specs-allegedly-revealed/
>according to VCZ this 5.5 TFLOP GPU is not even a desktop class chip but a mobility variant. This would bring R9 390X/390 class performance to notebooks.
Desktop variants will probably have the preivously rumored 2304 and 2560 shaders and offer Fury/980ti performance
>>54527734
look everyone it's someone who doesn't know anything about programming
>>54524224
About 26% faster than overclocked 980 ti. 90% faster than OC 980. Pretty impressive, would get shit on by sli 1070 for price/peformance however.
>>54528742
Big polaris 10 at fury x speed for $300 would make me wet.
>>54528869
Not going to happen
It will top out at 85% of the 390X's performance at stock speed (assumed to be 1200MHz) for the fully unlocked GPU, probably higher if they clock the balls off of them from the factory (which they will have to in order to compete with the vastly superior and faster 1060 Ti that will have 7.5GB of VRAM, not the full 8 GB that the 1070 will have). The Polaris 10 with shaved off CUs will probably stick around a 390's performance at similar clocks.
The GTX 1060 will probably walk all over Polaris 10 given its performance and how much of a gain Pascal has yielded at higher clock speeds. I think AMD's reign as the budget king will be very short-lived once the GP106 series comes out with even better power consumption and performance at higher clock rates.
>>54528979
They just can't win.
>>54529042
Their Pajeets pooed in the loo for the last time
It's back to street-shitting for them
>>54528625
AMD fanboys are worst. They knowingly defend a company that makes bad products.
>>54523858
Seriously retarded person.
>>54528979
>>54529042
>>54529062
>>54529153
>t. nvidiot samefag
>http://wccftech.com/amd-r9-480x-470x-specs-allegedly-revealed/
>299$ for 390/X performance
>and faggots here are speaking bad about Nvidia
Ayyyy
Git Gud
>>54528979
The smaller p10 supposedly has 2304 sp meaning its the 490 series. P11 is looking like its 470 and 460 series.
>>54529270
>literally costs the same as the rebranded and overpriced 390
>you could have bought a 290X for the same money last year and have slightly better performance
ayyyy
>>54529270
>>54528979
This entire post is complete bullshit.
>>54529315
If it's bullshit, then prove it wrong with facts and not speculation.
>>54527626
Not gimmicky just robbery
>>54529331
see: >>54529311
>>54529311
Holy shit, dude, you are one of those totally deluded people.
Let's think logical
>AMD says Polaris will not replace the Furies
>AMD says they try to hit the mainstream market, making VR ready cards
>clearly Polaris are lowend and mainstream cards
<hurrr durrr Polaris@over 9000Mhz clock will beat GTX980Ti and thus make the Fury obsolete
You underage faggots should question your sanity.
>>54521379
25% faster than previous gen.. it will be looked at as "slight upgrade" and most people will stick to 9xx/7xx series.
>>54529331
That entire post was unfounded speculation and shits in the face of what (admittedly little) we already know. We have no idea if the silicon the 1060 is based on overclocks to the same degree as the 1080. We also don't know how throttled the 1060 will be to fit into a certain power envelope. If an overclocked 1060 is "wiping the floor" with Polaris 10 for less money then Nvidia is just going to gut themselves of their own 1070 sales.
>>54529346
I said facts, NOT speculation
We all know that every single time AMD hypes up a brand new product family, they overhype and underdeliver. Anything they say is immediately suspect and should not be taken seriously.
Polaris 10 will not deliver twice the performance over Fiji.
Polaris 10 will not match the performance of a 390X.
Polaris 10 will not come out with 1600MHz clock as standard. In fact, I doubt that they'll be able to get close to 1200 MHz because GCN is so abysmal at attaining high clock frequencies.
We're looking at a 800-1000MHz GPU at best, and even those might be a high-ball estimate.
How much is AMD paying you to make up these bullshit speculations?
Pascal will beat Polaris and Vega at all price points. Accept it.
>>54526216
> caring for prizes
you guys are pathetic, only managing to stay alive for the approval of normies
>>54529270
>supposedly people who know wtf they are looking at.
>aknowledge they have mobile version specs
>still claim its a desktop 480x anyway
>>54529404
>what I predict is fact, what you predict is speculation
Fuck off, nvidiot shill.
>>54529403
>We have no idea if the silicon the 1060 is based on overclocks to the same degree as the 1080
Historically speaking, the fewer SMs Nvidia GPUs have, the higher clock rates they can obtain. Look at the difference between GM200 and GM204. It has more SMMs, but has to be clocked lower. GP106 will have about 80% of the SMs as GP104. It's only logical that they can clock higher.
>We also don't know how throttled the 1060 will be to fit into a certain power envelope
GP106 will be significantly smaller than previous GX106 dies. It stands to reason that they'll have to push those GPUs to the limits in order to close the gap between it and the lowest performing GP104. GM206 was different because they were stuck using the same 28nm die as the previous generation, thus the silicon was bigger than Nvidia would have liked.
>If an overclocked 1060 is "wiping the floor" with Polaris 10 for less money then Nvidia is just going to gut themselves of their own 1070 sales.
>If an overclocked 1060 is "wiping the floor" with Polaris 10 for less money then Nvidia is just going to gut themselves of their own 1070 sales.
The performance difference between a 1070 and Polaris 10 gives more than enough leg room for a seriously overclocked GP106 or a even more cut down GP104. There's a massive gap in performance, as the 1070 is at the 980 Ti level of performance, whereas Polaris 10 can't even match a 390X.
>>54529363
Replacing fury is the not the same as obsoleting it in industry terms. 1080 doesn't replace titan x for instance.
2016 still debating for GPUs while having single monitor 1080p............
>>54529445
AMD promised that the Rx 300 cards would be on a 20 nm die shrink with better power consumption than their Rx 200 cards.
They also promised that Fiji would be "an overclocker's dream".
They promised that they would continue support of legacy cards and prep them for DX12.
They promised better driver support with Crimson.
They promised that Mantle would revolutionize the gaming industry and become the standard API to supersede DX11.
They promised that Steamroller would have better performance than Sandy Bridge and their previous Phenom II CPUs.
I could go on.
AMD has overhyped and underdelivered with the consistency of a Swiss clock.
>>54529538
>They promised that Mantle would revolutionize the gaming industry and become the standard API to supersede DX11.
They were right.
also they never promised that.
>>54529471
Titan X are semi-professional cards.
It's another product segment.
No standard user would buy a Titan X for double the price compared with a GTX 980Ti, if he only uses this card for games.
AMD has confirmed that Vega will replace the Fury series as highend solution.
>>54529404
>Polaris 10 will not match the performance of a 390X.
Yeah AMD its totally going to release a next gen card for $300 thats weaker than its current $300 offerings you fucking mong.
>>54529548
>also they never promised that.
They did when they announced Mantle, dumbass. That wasn't even three years ago. How can you forget it?
>>54529466
>whereas Polaris 10 can't even match a 390X
Where the fuck are you getting this from?
>>54524561
>not being a waitist
Heretic.
>>54529582
>AMD its totally going to release a next gen card for $300 thats weaker than its current $300 offerings
Yep, that sounds about right.
>>54529589
Ironically the AotS benchmarks you AMDcucks love so much.
>>54529153
Exhibit 1
>>54529538
>AMD promised that the Rx 300 cards would be on a 20 nm die shrink with better power consumption than their Rx 200 cards.
AMD didn't personally fuck up 20nm planar. Lots of companies were fucked when this process completely flopped. It made sense to skip to FinFET at that point. I'm not sure how you can call that a lie.
>They also promised that Fiji would be "an overclocker's dream".
Fair point. The cards are worse overclockers than Nvidia's counterparts. I don't know where that quote came from but whoever said it should regret it. Fiji was disappointing.
>They promised that they would continue support of legacy cards and prep them for DX12.
This gets overblown. The late Terascale cards are now over 5 years old and wouldn't be able to run anything properly anyway. I see this argument get brought up all the time against AMD. DX12 is barely even here in 2016 and yet we're going to criticize AMD for not trying to bring DX12 features, through software no less, to 5 year old (read: ancient) cards? Next.
>They promised better driver support with Crimson.
They've delivered on this.
>They promised that Steamroller would have better performance than Sandy Bridge and their previous Phenom II CPUs.
I'm sure they weren't talking about heavily multithreaded tasks here. I'm sure of it.
>>54524224
>benchmark scores have almost perfect scaling with core clocks
>no mention of bandwidth comparison
>4k scaling higher than 1440p
And people are going to believe this.
>>54529585
Oh, I remember and this is why I know you are full of shit.
>implying either company would not want more shekels out of their cards
They have nothing to gift you.
The Polaris 10 will probably be the card to get for 1080p resolution.
The GTX 1070 for 2K resolution and the GTX 1080 for 4K resolution.
Anything else doesn't make sense.
>All specs and benchmarks of the new Polaris chips are for the mobile version
>Mobile Polaris can easily achieve 390X performance
>Desktop Polaris will have more CUs and clock much higher
>Nvidiots still believe it will barely match 390X performance
Nvidiots should really stop using their empathy boxes and return to reality.
>>54529704
>excuses excuses
Thank you for proving my point. AMD has always overhyped and under-delivered.
>They've delivered on this
No they haven't. They only did Day One driver support for a handful of titles and skipped out on the rest. Crimson still hasn't fixed the bugs left over from the previous driver suits, some of which has existed FOR YEARS.
Crimson is an outright failure as a concept because it was trying too hard to emulate what Nvidia did with its GeForce Experience, except they don't have the money or man-power to match up to them.
All you AMDcucks get out of it are buggy drivers, spotty updates, poor post-launch support, and a handful of Day One game support (which are left unsupported or without updates for weeks, just like before).
>>54521379
It will do well.
>for a none Ti product
>>54529758
>shilling this hard
>>54529806
>not wanting to come to terms with reality
AMD GPU ownership is fucking miserable. You have to make so many sacrifices in order to get a semblance of enjoyable gaming experience.
But at what cost?
The headaches associated with its still buggy drivers. The delays for updates and SELECTIVE bug fixes. Poor compatibility with weaker CPUs with DX11, which AMD has all but abandoned despite the market having 99% DX11 and earlier games still being played on a daily basis.
Even your favorite blogsite, SemiAccurate, called them out for this piss-poor support.
>>54529884
>mistaking delusions for reality
>>54530023
Strange, that's not what AMD's own forums are showing. There are a lot of unhappy or dissatisfied R9 owners on them, namely complaining about a lack of bug fixes that existed since 2014
Software developers will probably have to code in support for multi-projection, right?
>>54529704
Overclock and don't overheat
>>54530067
funny how nvidia forums are showing the same shit
ITT:
>Nvidia released a very good card with great temps, power and overclocking capabilities
>over 100% faster than a stock 980 in synthetic tests
>over 63% faster than a stock 980 ti in synthetic tests
>everyone is happy at these advancements, the cheaper 1070 and 1060 cards look like they'll also get the same treatment
>good for the industry and consumers as a whole
BAM
>buttblasted amdrones start flocking the thread
>proceed with damage control of the highest degree
>call everything shit as usual
>"w-watch vega will come out and have 7 gazillion TFLOPS"
>"n-nvidia are done"
>get told to fuck off
>"l-look at these n-nvidiots bullying le amd masterrace"
>play the victim
Like clockwork I guess.
I'm building my first pc. I haven't decided on a gpu and will get one later in Aug/Sept when the dust settles for Polaris/Pascal. I want to get started on the build now and will just use the igpu for the meantime. Here are my questions:
1. Is this a good build?
2. Will I see any comparability issues with either Polaris or Pascal?
3. Is there any reason why I shouldn't buy this is the next few weeks and have it already built in anticipation for the new gpus?
http://pcpartpicker.com/p/ZZYmf7
>>54531255
>not waiting for the broadwell-e refresh and getting a 6950X
What are you poor?
>>54521532
it's 35% more powerful than the 980Ti based on the leaked benchmarks I saw
>>54531282
>"No"
>>54523858
I was playing Blops 3 and my 670 couldn't even run at 1080 with the lowest settings. It really made me think its time to update. Also new COD is shit.
>>54531255
>spending $30 on a wifi adapter
Just use a really long wire or if you're wiling to spend that much money then get a powerline adapter if you're house isn't old as shit.
>>54523858
My 670 plays fallout 4 on high 1080 with some antialiasing reduced, which is unacceptable to me. Time for a new card.
>>54531342
gimpworks takes another victim
I bought a 980ti last month. Would it be a good idea to sell it now and buy a gigabyte 970 for $200 until the 970 comes out?
>>54531416
either keep your 980ti or sell it and live with integrated graphics for a month or two
>>54531361
I live in a detached unit and pick up the WiFi from the building next to me. It's basically a shed. Throwing a wire between the two isn't an option.
>>54531449
Fuck that shit. 970 it is.
>>54531543
You live in a shed?
Hahaha you can't make this shit up. Absolutely hilarious my man.
I still have an 8800gt. Totaly hyped to get a 1070.
>>54531361
i have this wifi adapter
its a pretty nice wifi adapter that also has bluetooth which can come in handy every now and then
my only complaint is that if you ever use bluetooth to transfer files it will drop wifi connections
>>54531854
I have a 7950. I'm not going to buy in again until there is something 10 times faster for the same price. Simply no reason to. You were wise to wait this long. People upgrade too frequently for less than marginal gains.
>>54531695
>2016
>not living in a cuckshed outside your mistress's and her boyfriend's apartment.
>>54531917
What's wrong with a normal usb dongle adapter like a tplink or something. I doubt a $30 one will make soo much difference at normal ranges. Hell, I can connect to my neighbors wifi across the road with my galaxy s4 and get an almost perfect connection. These dongle adapters are only like $10 as well.
>>54527757
>Shilling
About a free alternative that is not N$A shit
>>54521898
Let me know when it teraflops.
>>54524224
So basically overclocked 980 ti matches 1080 performance at lower cost.
>>54532015
>>54531695
I call it my bullpen.
Tbh it's pretty nice senpai. Its on a horse farm with peacocks and chickens. Beautiful scenery with mountains and shit. Cheap rent and all utilities included. Big enough to support my vive space too.
>>54532379
980 ti OC gets to around 21000 on FS. The difference is about one tier.
>>54521379
It beats the titan x
>>54526632
Dx12 is the death knell. Being Microsoft only in a growing mobile market is suicide.
>>54526367
AMD *SHOULD* not only have Polaris, but also Vega out by then.
I imagine Vega will actually launch alongside BF1, like how they did a BF4 edition.
Can someone show me which ways the 9xx series and 10xx series are missing key dx12 features? Genuinely interested.
>>54524224
AMD IS FINISHED
>>54532757
It's literally just one feature that's not even missing in Maxwell and Pascal cards: asynchronous rendering
The AMDrones insists that Nvidia cards can't do async because it's not hardware implemented, but they can through software. If you overclock a Maxwell high enough, it can still out-perform an AMD card even with software-based async rendering.
>>54532830
Is that something to do with multi-gpu stuff?
I'm out of touch, I went back to school 5 years ago and I graduate in 1 week and I am trying to catch up on stuff since I quit gaming in order to get my life back on track.
>>54531079
Nvidia succeeding more than they have already isn't good for anyone at this point.
AMD is Finished!
>>54532916
It's basically multi-threading, but for GPUs.
Allegedly it brings a performance bump in DX12, but no developer apart from the ones in AMD's pocket are willing to make async features compatible with their games. They don't even benefit AMD cards that much (not even 10% on average), so no developer in their right mind would want to code for a feature that will go underutilized with the vast majority of PC gamers. It's not worth the development time.
And AMDrone that says otherwise is either lying or fell for the marketing hype.
>>54529404
yeah no..
480=390x
btw pascal has 1,55 times better performance per watt over maxwell and offers around 1.7 times better performance. Pascal will have higher tdp, that is where the additional performance comes. If polaris gets over 1.55 better perf per watt over fiji then it will be a more competitive chip design than fiji was compared to maxwell. Actually we already know that it is actually over 2 times more efficient so that means Polaris and Vega will be successful
>>54533057
>no developer apart from the ones in AMD's pocket are willing to make async features compatible with their games
Developers in AMD's pocket use it because it gives huge gains on AMD cards.
Developers in Nvidia's pocket do not use it because it produces a disadvantage on Nvidia cards.
Just like how developers in Nvidias pocket ramp up tessellation because Nvidia cards have an advantage there and it makes AMD cards look worse.
Just like when Assassins Creed implemented DX10.1, which gave Ati cards a fps boost, a few patches later it was completely removed, because Nvidia cards did not do DX10.1 at the time.
And the list goes on and on and on.
>>54533042
It's good for me because I prefer Nvidia cards, my last card was a HD7950 and it was a mess.
Nvidia just werks babe.
>>54533057
DX12 and Vulkan are brand new API's and this guy already knows that no developer will use the async feature. It's like all the pioneers of silicon valley are right here on /g/. The future they know.
>>54533120
GPU doing parallel tasks isn't something new, infarct GPU are designed around parallelism, Async is just the name AMD is using,
>>54533110
bullshit. I got one rig with a 7950 and one with 2 7950's. No problems so far and they are waiting for their first DX12 games.
>>54533080
> If polaris gets over 1.55 better perf per watt over fiji then it will be a more competitive chip design than fiji was compared to maxwell.
Polaris 10 was shown running Hitman in 1440p at 60fps, and is said to be using somewhere between 100-170W depending on the source of the news.
However many news said that Polaris 10 is 100W while Polaris 11 is 50W, and we have a pic of Polaris 11 running while plugged into a wattmeter and it measures around 72W from the wall.
So the 170W figure for Polaris 10 could very well be full system draw, and the cards themselves being 100W.
Which could be pretty damn sweet, having 390X performance at 100W. Mostly because it shouldn't need huge cooling solutions. It could be a super small card, not as small as the Nano, but not a 30cm monster either.
>>54533164
I used to think the same way, then I got a GTX970 and realized how much better Nvidia is compared to AMD, my last Nvidia card was a 8800gt around the time Call of Duty 4 was released.
I thought AMD was good but in reality I just didn't knew better.
>>54533147
I never said it was something new...AMD has been doing it 6 years ago since the 7900 series. It comes down to the way it's implemented. oxide games is the first true developer to show the faults of nvidias async and nvida halfway fixed it. Good for nvidia.
>>54533196
>I got a GTX970 and realized...
Don't speak to anyone or give advice on this site or any other site. Just stop talking. When your mother ask you what would you like for dinner just shut and starve to death.
>>54531416
Listen here you little shit. Keep 980Ti. Then if you're want to play 4K 60FPS+ you buy a 1080Ti or Vega. Simple. 980Ti is already overkill for anything less than 4K. 1080 will gain like 6FPS over your card at 4K.
