http://wccftech.com/nvidia-gtx-1060-leaked-benchmarks/
http://videocardz.com/62278/nvidia-geforce-gtx-1060-rumors-part-7-new-cards-more-benchmarks
>>55602590
stop spamming indian street forums
>>55602590
>currytech
Hitman
Rise of the Tomb Raider
Just Cause 3
Black Ops 3
Doom 4
>>55602607
>>55602616
>>55602622
>10% better performance
>20% higher price
>by the time it is out, you'll be able to buy factory overclocked RX480 which matches the 1060 while being more silent and costing the same as REFERENCE 1060
>>55602590
>wccftech
lol
fuck off
>nvidia architecture doesn't gain ANY real performance with async compute on games
>somehow on 3dmark dx12 test AMD and Pascal have basically the same gains
I wonder who is behind this...
>>55603553
>implying you can't oc pascal
>implying the 480 14nm isn't a fail
>>55603591
http://videocardz.com/62278/nvidia-geforce-gtx-1060-rumors-part-7-new-cards-more-benchmarks
>>55602590
DELETE THIS
>>55603594
Its the same company behind extreme tessllation
Watch out, the AMD defense force is not going to like this.
>>55603594
does it really suprise anybody that the company who spent a fuck load improving dx11 to almost dx12 level performance sees almost no performance improvement going to dx12?
RIP AMD
>>55604444
Quads confirmed for truth
I know wccftech are a laughing stock for leaks, but how are videocardz?
>>55604513
In the past they've been shit, but for most of the Nvidia cards this time they've been pretty much spot on.
>>55603553
except the 480 has terrible OC headroom and tops out at 1350mhz, which is less than 100mhz from the stock core clock.
even a 1550mhz 970 beats a 1350mhz 480 lmao
>>55602590
◕ͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥ OFF THE CHARTS
>>>◕ͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥ NVIDIA
>>◕ͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥ AMD
>◕ͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥͥ
![[computerbase.de]_DOOM_+_Vulkan_Benchmarked._-_Page_23_-_2016-07-16_19.54.19.png](https://i.imgur.com/yFabPjZ.png "[computerbase.de]_DOOM_+_Vulkan_Benchmarked._-_Page_23_-_2016-07-16_19.54.19.png")
>Using 3Dmark Time Lie when Doom clearly shows it to be false.
Time spy doesn't actually use asynchronous compute, it uses concurrency which is different.
Read the 1080 white papers you idiits
'I should note that the primary performance benefit as implemented in Time Spy is via concurrency, so everything here is dependent on a game having additional work to submit and a GPU having execution bubbles to fill.'
http://www.anandtech.com/show/10486/futuremark-releases-3dmark-time-spy-directx12-benchmark
Time Spy does not use parallelism for async compute thus it does not support AMD's async compute proper. It is heavily biased towards Nvidia's preemption etc
Basically you can't use Time Spy to compare AMD to Nvidia on a level playing field. You would need it to change the algorithm for each so that Nvidia gets the best score possible using it's architecture and change it to parallel to get the best from AMD. Just using one type of system for both skews the results.
how much will 1060 cost?
>>55604707
>the 480 has terrible OC headroom and tops out at 1350mhz
Funny that the PowerColor Devil card is shipping at 1367MHz out of the box then. Boy are they going to get a lot of returns when people realise that the card doesn't work at all!
>DX12 Benchmarks
Irrelevant.
>>55605116
AMD is not crying they are laughing because Vulkan and DX12 in games clearly shows that 3dmark is not telling us the truth. Otherwise we would not be seeing such significant gains in Doom. RotTR, Hitman, AotS etc.
It is annoying that 3dmark gimped Time Spy though. I shall be posting on their forums asking why this is the case once the site is back up (conveniently it's down right now. It's almost like they are deliberately avoiding the backlash).
>>55605092
>Time Spy does not use parallelism for async compute thus it does not support AMD's async compute proper.
you clearly don't know the difference between parallelism and concurrency
https://stackoverflow.com/questions/1050222/concurrency-vs-parallelism-what-is-the-difference
AMD has true concurrency (SMs can be utilized for general compute work while graphics is bottlenecked by ROPs/TMUs/other fixed function hardware), NVIDIA only preempts between the two which leaves some parts of the GPU idle.
>>55605142
250 MSRP for for cards by 3rd parties
300$ for FE.
>>55605224
So almost tree fiddy for the AIB cards after nvidia tax.
>>55605174
>is not telling us the truth
![RX-480-ABC-81[1].jpg](https://i.imgur.com/igsTCOm.jpg "RX-480-ABC-81[1].jpg")
>>55605320
those don't look real to me. dx 12 rise of the tomb raider, for example
http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/72889-radeon-rx480-8gb-performance-review-19.html
>>55605394
I don't get it
>>55605409
your image says 41.4 fps for the RX 480 in rise of the tomb raider at 1080p but that's not what other benchmarks show. just one example.
>>55605430
Because ALL reviwers have the same setup or preset?....
>>55603607
You can but its a waste of time.
If 3Dmark scores are telling the truth then why does Doom show such huge gains for AMD? Why does RotTR show big gains for AMD? Why does Hitman show such big gains for AMD? Why does AotS show such big gains for AMD? Why does Warhammer show such big gains for AMD? Also.
![index[1].png](https://i.imgur.com/uIARiyI.png "index[1].png")
>>55605452
im just saying that i've never seen a benchmark with scores as low as the one nvidia posted. i don't know how they got numbers that low for the 480
http://www.guru3d.com/articles-pages/amd-radeon-r9-rx-480-8gb-review,10.html
>>55605477
>Galax
>Nvidia
Need I say more.
>>55604997
>mahigan
If /g/ was half as educated as that guy is about gpu's this board would be a much better place.
>>55604997
THE GOYIM KNOW
SHUT IT DOWN
>>55605499
The other pic is newer than the 480 review - tomb raider's patch 7 came out a few days later and a few sites (like hardware canucks and oc3d) retested. Guru3d did not.
>>55603607
>implying you can't oc pascal
But that's the thing, you can't.
>>55604707
>except the 480 has terrible OC headroom and tops out at 1350mhz, which is less than 100mhz from the stock core clock.
We've been over this. The card come with insufficient power distribution and the cheapest possible cooler.
AIB models get a much beefier cooler and 8pin or dual 6-pin power, so they'll be able to squeeze out way better overclocks. Even the biggest anti-AMD shill site on the planet has confirmed that they are easily getting 1350-1500mhz on the cards depending on silicon quality.
>>55605701
>But that's the thing, you can't.
can you elaborate?
>>55605701
the 14nm is a fail
>nvidiamark
>>55605744
>midrange $200 card
>performing close to flagship Fury and 980Ti
yeah ok.
>mfw dx12
>>55605394
SMAA vs SSAA
Well I don't know, anon
so 1060 or 480?
>>55605774
I talk about the 16nm vs 14nm efficiency
>>55605788
>quantum break
That is such a troll image given 1) the software was capturing frametimes wrong and 2) Nvidia suffered the same behaviour.
>>55605843
1060 obviously
>>55605701
>>implying you can't oc pascal
>But that's the thing, you can't.
wat
>>55605476
Performance gains across the board
What's the issue here?
>>55605320
>Material made by Nvidia's own marketing department with suspicious-as-fuck settings (using a type of AA that doesn't allow async compute among others) and outdated drivers for AMD
>"The truth"
We spent last week reminding you n/v/idiots where this came from and how suspicious the settings used were and you're still reposting in hopes people will buy it?
There was plenty of shilling for the 480 before it came out, but this is almost on a whole different level posting the same thoroughly debunked figures over and over again...
>>55605476
The fact that the gains are marginal for the amount of overclocking required.
>>55605850
>I talk about the 16nm vs 14nm efficiency
You cannot directly compare the efficiency of the fabrication method on two completely different chips. Polaris is low-risk design, optimized for high yields on a brand new (and therefore risky) node.
>>55605476
>you cant oc 1080
>>55605843
480 obviously
>>55605843
1060 = meme
480 = supreme
>>55605320
>Using SMAA to avoid async compute on AMD blowing it the fuck out.
>>55605896
>2 minute benchmark runs.
>>55605850
>Make a 232 mm^2 chip and crank it up so that it competes with chips twice it's size
>Somehow surprised that it's not as efficient as it could have been
Right...
>>55605894
comparing a 14nm card that can actually do everything on hardware
with a 16nm card that miss most of its hardware in favor of power/perf
and yet it needs to be overclocked to almost to max to even pass 480 on stock
>>55605933
Do you have a Nvidia card? It's the BOOST, not the overclocking.
>trusting benchmarks before the card is released
RX 480 was supposed to not be that good, but in the end with the drivers and the non reference GPU coming we're probably going to have a very good GPU for the price. You can't compare them at this point, wait a month or two so you'll have something more reliable.
If you want to gimp 3Dmark further towards Nvidia just leave your card running at 100% fan speed for 10 minutes then run 3Dmark with the demo before the benchmark disabled.
>>55605959
You do understand that 1) those are boost clocks and 2) boost is inherently a form of automatic overclocking until the card hits either its bios defined temp limit or power draw limit.
Plus overclocking Nvidia cards is essentially getting the card to always run at a given frequency, rendering the point of boost irrelevant.
but can i buy a 1060 for $350 CAD on tuesday?
>>55605985
>Plus overclocking Nvidia cards is essentially getting the card to always run at a given frequency, rendering the point of boost irrelevant.
wut? no it doesn't. overclocking an nvidia gpu through afterburner would just tack the extra clock speed onto whatever boost clock your card is running at under load.
>>55605985
>Plus overclocking Nvidia cards is essentially getting the card to always run at a given frequency, rendering the point of boost irrelevant.
>>55605985
>has never owned a nvidia card
>shills against it for free
Overtime
>>55606398
*everytime
>>55603622
DELETE THIS
my wallet is dead for buying nvidia shit.
>looking at those bullshit charts
>480 so close to a fury x
lol, fucking horseshit
>>55604997
Saved.
Fuck, he's good.
None of these benchmarks compare 1060 vs 480 on Vulkan Doom with async compute turned on.
Did Nvidia tell these guys they'll be sued if they actually benchmark the cards, or something?
>>55606633
You should see the shitstorm he caused over on OCN when ashes of the singularity made news for its heavy usage of async compute. In a nutshell the Nvidia crowd told mahigan that he is wrong and a fucking shill and that other DX12 titles that aren't AMD propganda will show Nvidia's dominance.
As we now know that didn't happen and mahigan called it right last year. Drama follows his posts everywhere primarily because he does do his research and is generally correct in his claims.
>>55606752
This guy got a YouTube? Podcast? Blog? Anything? AMD fanboying aside, he seems really informative. I'd listen to his podcast daily.
http://steamcommunity.com/app/223850/discussions/0/366298942110944664/
From the horses mouth - async compute is disabled form axwell in Nvidia's driver.
>>55606801
Not to my knowledge - he is just some random dude who posts on OCN and anandtech (he might have an account over at hardocp but lol fuck that place).
>>55606732
They just don't send the review samples anymore if they show them in bad light just like most companies.
Lmao that's a new one AMD subhumans complain about currynigger websites after they've been using them as "credible source" to shill for 480. Ban those Pajeets asap.
>>55606894
Remember when Nvidia blacklisted Tweaktown? Fun times.
>>55605985
>I know what I'm talking about cause I'm a amdfag
>>55606732
>>55606894
>>55607047
the 1060 isn't released....
>>55606846
>You are so full of FM_Jarnis, this is being throughly researched by the best Overclockers in the world.
Holy shit that's a lot of salt.
>>55607123
People expect too much from the benchmark. It does employ async compute, just not in the manner many had hoped after AMD's marketing of the term. AMD has stretched the definition a bit far because they have so much hardware capable of handling it.
>>55607169
I think people are too obsessed with their religion.. I mean brand winning. They'll go to any lengths to make it seem like their product is the best even when it's not.
That benchmark hardly even makes AMD products look bad. The RX 480 was supposed to be faster than GTX 980... and it is. Or maybe it was 970, but I digress. That seems like a pretty good result to me.
>>55607246
It doesn't help timespy is hot on the heels of doom's vulkan patch which uses quite a lot of GCN favouring features which is why you see those enormous jumps (especially the fury x).
With the trend of DX12 titles (and much misunderstood, but hyped async compute) generally running faster on GCN people were hoping to see more gains from timespy given historically 3dmark's benchmarks have been about squeezing as much fro ma gpu as possible.
It is fairly odd in a round about way to have a synthetic benchmark show lesser gains than in real vidya, but here we are.
>>55605294
$420 in europe
so round that up to 420 euros
>>55607317
It's not that odd, synthetic benchmarks hardly ever give you performance that represents games. That's why everyone calls them synthetic.
I don't even look at synthetic benchmark scores for anything other than to compare my card to other cards of the same type to see if it's under-performing or doing better.
>>55605506
I know his name from since that DX12 thing started, wise guy.
>>55602590
>With Async
nothing better than seeing the AMDefence force flailing and squirming again and again
>>55605506
He's technically just bitching about semantics and marketing though. I think the asynchronous meme should die already.
I don't fucking care HOW the GPU delivers performance. I care what the performance is. If AMD Fury X delivers like 50% more performance than GTX 1080, I damn well don't give a fuck how it got there - just that it did.
>>55607525
nothing better than nvidiots sucking nvidia's dick over an overpriced card that doesn't even exist yet
I have had my 7850 2gb since 2012 and while it's tempting to get a 480 or 1060 I think I'm gonna stay put for now. I can get by at 1080p high settings in the majority of games when turning down shadows and some other filters like SSAO in certain games.
Witcher 3 so far has been the only game that I feel like really pushes my GPU, while I can keep textures on high I have to set shadows and draw distance to medium and even then I get 30-40fps.
I'm more or less a poorfag, so when I make my next upgrade I want it to be a massive jump, I might have to wait 2-3 years, I plan on being able to play games at ultra 1440p or high/ultra at 4K by then. There's really nothing out there right now that I NEED to play at ultra settings and 60fps. If anything I might get a PS4 sometime in the next year if Horizon and other exclusives turn out to be good since right now the PS4 is just an anime simulator, CoD box, and Bloodborne machine.
>costs 25% more
>is faster
Shocking desu ne. I also predict the RX 490 will be faster than the RX 480. Screenshot this.
>>55607336
smoke weed every day xd
>>55607821
>is faster
because the 960 and 760 were such a good deal, amirite?
>>55607895
Some guy was hailing the 760 SLI a second coming of Christ the other day, so maybe.
Nice. the 1060 is a 20% cheaper card with 20-30% more performance than the 480.
>>55602590
>somehow believing the people who told us the 480 would beat the 980
promptly kill yourself
>>55605016
this is the dumbest fucking post I've seen today
>>55605174
AMD is dying, pic related
>>55602590
Awesome, nearly $100 more expensive after you pay the Nvidia tax and the most marginal "improvement" over the 480 imaginable - you'd have to be an idiot to waste money on a 1060. Sure if you're going high-end then buy Nvidia, but AMD's completely conquered the entry and mid level part of the market this time around.
>>55608122
>buying cheap AMD GPUs that break down easily
Yeah...really worth it...pic related
>>55603594
how in the flying fuck do you look at this >>55602622 and claim they have the same gains?
nvidia
>1080
6% gain
>1060
4% gain
amd
>fury
15% gain
>480
8% gain
see pic related. more results that show gains
nvidia
>1070
5%
amd
>fury nano
11%
amd receives stronger, HIGHER gains than nvidia with pascal. maxwell is so bad that nvidia has async disabled via the drivers on maxwell series cards.
how can anyone look at this and claim there the same???
also i own a 1080 and its not hard to look at these and see gcn being overall STRONGER at async than nvidia. all 3dmark just did was PROVE amd's lead with dx12 and vulkan over nvidia.
for fuck sakes do better amd shills. a nvidia shill shouldn't have to be defending you here with a win you guys finally achieved.
>>55602590
>wcctech
>videocardz
Into the fucking trash. Wait for the card to release goddammit. I'm sure it will perform better than a RX480, but you're not getting a 1060 for less than $280.
>>55604997
>d-delete this, the post
probably one of the same guys saying the 480 would beat the Fury X a month ago.
>>55608150
>for fuck sakes do better amd shills
AMD shills are literally mentally retarded. They were dumb enough to buy an AMD card so you know they aren't right in the head.
>>55605888
>14% OC on the core
>10% increase in fps
My old HD7950 had a 25% OC for about a 20% increase in fps so I don't see the problem.
>>55608113
soon nvidia and intel will halt progress and price drops, knowing theres literally no other options or no one willing to buy them
NVidia cybervolunteers, do you find it ironic calling AMD users drones when you spend hours shitting over the smaller company
>>55608113
Those numbers are pulled completely out of thin air, and even if it's real (it's not) the 480 has already highhandedly outsold both the 1070 and the 1080 this generation so very soon we're going to see those two lines start to converge. When consumers realize what a ripoff the $300 "founders edition" of the 1060 is, those lines will probably cross.
will the 1060 actually cost 250 when it comes out though?
>>55608175
>mfw subscribed to Adored
>>55608276
$300
>>55605476
>hurrr can't OC pascal
you don't even have one faggot
>>55608276
Absolutely not, if the 170 and the 1080 are any indication. They've already said the Founder's Edition is $299, and we STILL haven't seen 1070 or 1080 prices drop down to what Nvidia claimed the MSRP was supposed to be. In other words, Nvidia is just deceitfully using a bullshit price point to score cheap headlines about how reasonably priced their cards are and capture the hype before they're actually released.
Maybe a few months from now once people have realized how badly they were swindled prices will drop down to reasonable levels, but as it stands now the only choice an informed consumer has is AMD.
>>55602590
It cost $100 more for a reason
>>55607169
>It does employ async compute, just not in the manner many had hoped after AMD's marketing of the term.
>>55607317
>much misunderstood, but hyped async compute
Considering AMD built the async APIs, I think that we should be considering their definition the primary one. AMD has been pushing for these features to be included in mainstream APIs for years, which is why AMD cards from 2012 are seeing gains under Vulkan and DX12. This is how long AMD has been building async hardware.
Do you think they started developing Mantle just for fun? It was because Microsoft wouldn't get off their asses to build a new API. Once the API was done, AMD gave it away for anyone to use, as that is in their best interest.
https://en.wikipedia.org/wiki/Mantle_(API)
>Asynchronous DMA queue for data uploads independent from the graphics engine
>Asynchronous compute queue for overlapping of compute and graphics workloads
>overlapping of compute and graphics workloads
Asynchronous compute is EXPLICITY NOT the same as context switching.
>>55607527
>wants nice things
>doesn't want the information necessary to know what's nice and what isn't
Yo, dude, if you plug your nose and put on a blindfold, I'll feed you a spoonful of the most delicious food known to man. Promise.
>>55608538
Synthetic benchmarks don't tell you what's nice and what isn't. Games do. Have you ever played a synthetic benchmark, retard? I don't think you have.
I bet you unironically have 20 hours of "game time" in 3DMark Firestrike.
>>55608137
AMD GPUs get used a lot in bitcoin mining. Something Nvidia cards cant do.
I'm refunding my timespy upgrade. Sure it was only $5, but its the principle of the matter. If they are using context switching and calling it async it's a lie. Futuremark is going to be in a fair bit of trouble if people can prove that they created a biased benchmark for dx12.
>>55608605
>being this mad
Only an idiot would argue in favour of ignorance and then get mad when someone calls them on it.
Next time try using your brain.
>>55608666
You sure you aren't projecting there friend? Did I hit the mark, you actually do play benchmarks instead of video games?
Know what matters in video game? FPS. It doesn't matter how you get it.
>>55608673
>It doesn't matter how you get it.
look man, if you want to let black dudes fuck you in the ass so you can buy a computer for $3000 that gets 5fps more than a $1500 computer, be my guest, but not everyone likes getting diseases
>>55608712
I'm sorry to hear that you have to resort to that to make money. You should try a different job. I can't imagine it pays so well.
>>55608519
Didn't Intel licence some of AMD's technologies to for their newer integrated graphics?
Seems the only company not taking advantage of it properly is Nvidia.
>>55603622
>>55608626
Look at the failure testing rate, that is before the card is even put into production.
AMD cards fail WAY more often than Nvidia ones even just under basic testing.
>>55608254
Uhh those numbers are directly from AMD themselves
>>55603622
>>55602590
>Trusting leaks
Remember when retards were saying 480=980ti?
>>55608978
/thread
Probably some SMAA chicanery at play
Curry niggers are revolting and India should be nuked. I refuse to buy shitty AMD products because they're shit but also because of curry niggers.
>caring this much about performance
>when in 11 months the RX 580 will rape the GTX 1080 for a fraction of the price
why doesn't AMD just make a budget card that outperforms nVidia's flagship?
>>55609336
You can only waitfag so long.
>>55609592
>tfw you've been waiting for 7 years
3Dmark says 980Ti does not do async computer due to driver saying 'Nope!'. But.
http://cdn.arstechnica.net/wp-content/uploads/sites/3/2015/08/Review-chart-template-final-full-width-3.0011-980x720.png
Hmm...
>>55609612
Thats with Async compute disabled.
Nvidia has been disabling async compute.
>>55609221
>SMAA chicanery at play
what?
>>55609606
Too long, bro. I only waited six and I'm done, yet I still have to wait because supply of the 1080 is absolute dogshit.
>>55609606
>>55609684
You guys realize you could literally buy a modern $40 dollar normie card and it'd be better than whatever the fuck you're running, right?
are we at the peak of visual quality in games? how long before games reach life-like quality and the need of upgrading gpus is no longer needed?
>>55609745
its gonna be a long ass time. i dont think its any time soon actually. progress comes with competition and theres really not much competition going on right now. amd isnt trying to fight nvidia for the highest performer, theyre trying to hit the low to mid end. nvidias on top performance wise and they wont release anything groundbraking and new until amd starts doing the same. its at a stalemate, unfortunately. hopefully someone else steps into the mainstream desktop gpu market, possibly intel making discrete gpus?
>>55609745
>implying you can't exceed "life-like"
>>55609821
be kinda nice if intel could make a APU that would compete with the 1070/1080 and 480 etc etc while keeping the i5 6600k cpu performance
>>55609745
when we can hit 40tflops
>>55609606
>tfw you've been waiting for 8 years
>>55609821
>theres really not much competition going on right now.
AMD is banking on the longrun. The reason why the dumbasses showcased two RX480s against a 1080 was just to demonstrate the possibility of multiple GPU setups when proper APIs make good use of them. Having multiple smaller chips is better for all of us since they're cheaper to us and yield much more for the manufacturers, there's also Moore's law not being shit on smaller chips.
>>55609731
damn right I know, but now that I have the money im buying something i wont change in a long time
>>55609848
would be nice if amd could make an apu with competitive performance for a solid price, all theyve got is an fx6300 and r7 250x molded together to fit fm2. shit fucking sucks ass. if they could merge zen&rx480 then they might win the cheap shit market hard as fuck. normies will cream for that shit, not even 300 dollars for an amazing cool cheap integrated gpu cpu dealio. intel would shit bricks at that. assuming zen turns out alright. for the sake of the industry you better fuckin hope it does, intels got funky shit up its sleeve, what with mr. bazinga dicking around in the sciency rooms and shit.
>>55609918
i dont doubt it man. would love to see that become the case, i think theyve been hinting for the past few years they want smaller shit to work together. cough cough fx series cumstain
>>55609606
>>55609918
Multiple GPUs are shit though, for the same reason we don't have desktop computers with multiple CPUs
AMD is just struggling to even produce budget shit at this point, their products are failing massively and they'll be gone in 4 years guaranteed
>>55610045
>Multiple GPUs are shit though, for the same reason we don't have desktop computers with multiple CPUs
But anon, literally the only reason for that is that the cost doesn't justify the gains for typical desktop usage. There's effectively 0 return beyond quad core for desktops and you can reach that on a single chip easily.
>>55610045
>Multiple GPUs are shit though,
They are only shit because their implementation prior to these new APIs was awful.
>they'll be gone in 4 years guaranteed
You have no idea what you're talking about.
>>55609606
I still have my GTX 280
>>55609606
What exactly are you waiting for...?
>>55610191
a job
>>55607808
cool blog, may I suscribe famalam?
>>55602590
They said 480 is stronger than 980 and oced to 1600. We all know how that turned out. Stop posting that garbage site and it's garbage leaks.
>DX12
>Impying that DX12 not deprecated.
Install Gentoo musl with wayland.
>>55611244
>will not be open sourced
fuck off
>>55611294
>shall be open-sourced some day
Not fuck yourself in eyes
>>55602607
Fuck fury barely performs what's the point
So at the 480's price point, what's the best new card I can get right now? I don't want to wait for 1060s to come down from the artificially inflated original prices, but 480s are taking forever to stock and are gone in 5 minutes.
>>55612164
Sit in a fetal position and cry until prices come down
>>55612164
The NDA end 19 July, wait the test.
>>55612193
fug
>newer card is better then slighty older card
>ya but here look at these benchmarks of newer card with older cards
come the fuck on
man its too confussing
#Nvidia can't
Dx12
properly async
Stay under 51 C
Over priced.
Still too arse blasted to add 295x2 and raedon pro duos in game benchmarks that favor crossfire/sli
mfw: the 1080 is over hyped and will get ass creamed by vega hence the rush to get titan P out.
-Nvidoits didn't see this coming
#AMD can't
PCI-e keep inside properly the 75 watts
Drivers are too slow to come out
Can't get high clocks
-can't move on from 1 fetch and 1 decode GPU -hardware codex GPU and CPU side (doesn't realize this is the bottleneck on the cpu and gpu because of this one issue that held back bulldozer in the first place)
-too slow to get up to nvidia
-Wait is all they ever say.
-Vega is not even in existance
-the 295x2 and raedon pro duo still never get mentioned because software crossfire is STILL not embedded at driver level. (for fucks sake man do it you lazy ass)
all of this AYYYYYYYYYYY LMAO is too much for me.
This generation is over hyped I'm gutted the 1080 can't get over 30FPS on 4k
sticking with my R9 380 SOC until the price gouging stops and we have decent cards for 4K.
i'm not paying $1499 NZD for a card that can't even 4k at ultra settings with everything maxed.
Fucking /g/ hyperbole nonsence.
GET OUT with your shills
REEEEE~
>>55608978
Says very high preset...
thinking about finally upgrading from my 770. am i going to see a significant difference with a 1060?
>>55613369
>Can't get high clocks
You do realise how Nvidia does this and AMD doesn't right? A big one (while there are many other reasons) is Nvidia uses different clockspeeds internally for different parts of the chip. The clockspeed they tell people a card runs at (and what software reports) is only the shaders.
AMD uses a single clockspeed for everything so overclocking GCN bumps everything up to whatever value you set, which is why 1) power draw goes up so much and 2) you see much closer to linear scaling with GCN compared to maxwell and pascal. Equally this means you don't need absurdly high shader clocks because you are getting everything running faster, rather than just the shaders which now is creating internal bottlenecks (more so than any particular design has) depending on the workload.
This does mean - to a certain extent - you can throw any workload you like at GCN and adjust clocks as you need to get performance. Maxwell and pascal won't respond like that for anything that doesn't really touch the shaders.
Nvidia are just holding back on async (and holding back developers from using it) so when the next generation comes out they can obsolete all their previous cards and force "I only use nvidia cards" retards to buy new async enabled nvidia cards.
Hopefully AMD wins market share and people don't feel the need to fork out another $600 on the latest highend (but really midend after the ti comes out) gpu.
Then the Chinese steve jobs commits suicide and we all live happily ever after. The end.
>>55609745
when quantum computing becomes more accessible
>>55609094
>Look at the failure testing rate, that is before the card is even put into production.
do you even hear your self? how can a random polish site have such info? that has to be one of the most pathetic nvidia shill attempt i have ever seen
>>55608649
besides that they charge for unlocking resolution 5 bucks more...literally they made dlc for their bench...
bunch of hobos
>>55602590
I wonder what these scores will be when Async and the benchmarks in general don't favor Nvidia. When are these going to be fixed?
>>55607169
I think it's more about possible collusion and misinformation that people have a problem with not this us versus them mentality. I'm a consumer and I want the best product for this price and that's pretty much what it boils down to.
>>55607525
>>55607657
It's like this entire board works for nvidia and amd, it's like you actually get paid to advertise and argue.
>>55608228
>>55610205
>>55615646
Nope, just an average joe here who believes the GeForce® GTX line are the most advanced gaming GPUs ever created, driven by the high-performance, power-efficient new NVIDIA Pascal™ architecture.
>>55615610
How retarded would i be to upgrade to a 1060 from 960 when i use a 1440p
And i do know i was retarded for getting a 960 in the first place but poor
>>55615805
Why dont you just downgrade to 1080p?
>>55615865
I can never go back
>>55615886
Why dont you work hard and save money for a month or two for than Pascal Titan which is about to be released.
You probably wont be bothered to buy a new video card for like 6 years or when it naturally dies.
Nvidia is dead.
>>55615954
I have enough money for a 1070/80 but i was raised and grew up being as frugal as possible so spending more than i need do gives me lots of anxiety
But i completely understand that i'm being irrational
10XX is not capable of DX12 native.
Sorry fanbros
Well, Not sorry.
Do You fanboys like AMD, giant Juice, Big Black Baby Bo....
>>55616013
Just get the best warranty available. If you're not overclocking and it breaks down. You'll most likely get a next gen card If it breaks down as a replacement.
Mine was a reference gtx 480 broke down after few years and replaced with gtx 670 for free.
>>55602614
This. WCCFTech and Videocardz are far from reputable.
>>55617979
its like ken and whycry are on a battle to see who can spread more bullshit inside a year
but can it run crysis?
480 is such shit
>>55608113
This image just makes me sad. AMD's 5000, 6000, and 7000 series were so strong and competitive, but they lost market share any way. Fucking why?
>>55605477
>Huge gains
From Ogl 4.3 to Vulkan [async included] + (((AMD propietary shader intrinsecs)))
Woah, not expected
>>55609606
I always liked the idea of doing one generation behind. I did this for most of my cards.
I'll give an example:
When the Voodoo 2 came out I got a Trident 4MB 9750 AGP
When the Voodoo 3 came out I got a Voodoo 2.
When the Voodoo 5 came out, I got a Voodoo 3
When the Geforce 2 came out I got a TNT 2
When the Geforce 3 Ti500 came out I got a GeForce 2 MX 64MB.
When the Geforce 4Ti 4600 came out I got a Geforce 3 Ti200
When the Geforce fx5200 came out I got a Geforce Ti4200.
When the Geforce 6800 came out I got an ATi Radeon 9800.
When the Geforce 7800 came out, I got a 6800
When the Geforce 8800 GTX came out I got a Geforce 7800
When the Geforce 9800 came out I got an 8800
When the Geforce GTX 260 came out I got an ATi Radeon
When the GTX 480 came out, I got a Radeon 5870.
When the GTX 580 came out, I got a 480
When the 680 came out, I got the 570
When the 780 came out I got a 670.
When the 980 came out I got a 770
When the 1080 came out, I got a 970.
>>55622263
This is retarded. Some of those aren't even a generation difference.
Also:
>When the GTX 480 came out, I got a Radeon 5870.
>When the GTX 580 came out, I got a 480
>When the 680 came out, I got the 570
For what purpose? 5870 was within 10-15% of the 480 at worse. GTX 570 used the same chip the 480 did but at slightly lower power.
>>55622450
I said I did this for most of my cards. Key word being most.
The 5870 was cheaper for me at the time.
When the 580 came out, on the recommendation of the tripfag Dutchricer, I got a 480 as they were greatly reduced in price.
After a while, the 480 started artefacting and I got the 570.
>>55602590
>more expensive card has better performance
Wew lad
>>55622263
>When the Geforce fx5200 came out I got a Geforce Ti4200.
was there something special about the 5200? i got one in my first pc from my dad but i read a lot of people had that card. wasnt it absolut garbage tier as the number indicates?
I get the feeling that with async compute no one really knows how much of a performance benefit we should see in real world scenarios but everyone feels that AMD should be doing better because of their better/real implementation of async compute at the hardware level.
The result is chasing after phantoms.
>>55623270
It had really poor DX9 performance. This was a big deal as the FX series was touted as the definitive DX9 range.
>>55622823
barely better, and at a hugely inflated price
>>55620986
Marketing.
>>55603553
>10% better performance
>30% cheaper price in EUROPE
go fuck off USA
>>55609745
Youre looking at a minimum of 3x CPU single thread performance and GPUs, using todays hardware, with more then four times the ALU count and triple the frequency. You wont see whatever "life like" means to you for ten to fifteen years.
>>55609999
>>55602590
True tests tomorrow
So Nvidia jacked up the price of the 1080 by $100, called it a "founder's edition" and purposefully under-produced them so they'd sell out faster and make it seem like there was a huge demand for them.
Now that the non-reference designs are appearing they are pretty much the same price, if not more. Why is this "give us $100 as a pre-order bonus" thing still happening for the non-"founder's" cards?
>>55628003
>not understanding the concept of hoarding and jewing due to lack of supply
How are you even alive in 2016?
Is this your first time purchasing new consumer electronics?
>>55622263
>Not skipping multiple generations with the same number
>Buy C2Q6700 years ago upgrade with i7-6700k
>Buy 9800GX2 upgrade with 980Ti
Stop being wasteful man.
>>55607808
No one asked.
>>55628003
mfw founders editions are actually the cheapest GPUs u can get in Germany.
Any aftermarket card costs at least 60€ more
