The GTX 1070 will be a long step from the GTX 980 Ti to keep the pricing up on the older cards.
>>54635923
>980
>$349
Where
>>54636002
You have to wait for the 1070 to be released.
You would have to be crazy to buy a 980 instead of a 1070 for only $30 more.
>>54635923
>SEVEN-POINT-FIVE-GIGABYTES
>>54635923
This never happens. Nvidia cut production of the 900 series already, so the rest will be sold at the same price for SLI.
>>54636281
An absolute madman. Or you know, someone already has a 980ti and wants sli.
>>54635923
1. Your graph is lying you fucking retard
2. Nvidia has already sold all of their 900-cards. They don't benefit from "keeping the pricing up" you fucking retard
>>54637407
AMD has already announced that they got cucked again and polaris will only target low-mid end "value" segments
>>54637485
>980 is higher than mid-end
are you sober?
>>54637485
This, Polaris is shit.
Vega, on the other hand. 16GB of HMB2.
>>54637539
>making assumptions
How's that 10% performance gap and house fire Temps doing for ya?
>>54636683
>>54637404
I fixed it for you.
>>54637532
Yeah, that's the point retard. Both 1070 and 1080 are way faster than 980 and AMD can't keep up.
>>54637485
>This meme
The full Polaris 10 chip will be close to 1070 performance for $300-$350
The cut down version will be 10-20% slower for $250
Polaris 11 will be 960/380 tier for $120 or something like that
>>54637676
What's the point of comparing unplayable fps?
>>54637764
you're comparing cards. retard.
>>54637764
My thoughts exactly. 4K might be the biggest meme in tech history
>>54637793
>which of these cards is LEAST unplayable
It's unrealistic and basically useless
>>54637676
So even with your desperately cherrypicked example, the 1080 is still ~60% faster than the 980. This confirms what I said retard. ~70% was even what nvidia claimed in their promotional material
I got this from closed sources.
>>54637806
Just lower settings retard. Clean picture without jaggies/shimmering is way better than ambient oclusion and godrays and whatever else bullshit they tack on top
>>54637900
>wasting HBM2 on Pascal
lol
>>54637912
>wasting HBM2 on anything amd
lol
>>54637660
that's literally the low-mid "value" segment you dip
>>54637900
I don't believe that graph is authentic, but if Vega is going to be under 500 dollaridoodoos I'm going to get one.
>>54637764
We're still trying to get a solid 60 at 1440p, let alone 144.
>secret-document.png
>classified screenshot
Next time try to make it believable. Vega 10 should not be cheaper than 1080. Vega 10 is supposed to have 70% more shaders than Polaris 10 and should therefore be faster than a 1080.
>>54635923
Meh. I'm just waiting for a future proof 1080p 60fps card, or maybe the 980Ti if it drops in price. Yes the 980Ti is overkill for 1080p, but that's kinda the point.
>>54638049
>maybe the 980Ti if it drops in price.
it can dropped to €550 for me
>buying nvidia
ayyy
>>54638049
>future proof
massive meme
>980ti for 1080p
massive overkill
>>54638049
not even the gtx 1080 is really overkill for 1080p, current games with maxed settings can be like 60+ fps but not like a solid 120+ fps, not all of them at least
>>54638071
>massive overkill
not when it'll get gimped by driver updates
>hurr don't update
then you can't run games that use new API's/featurs
>>54636002
evga b-stock occasionally has them
>>54638077
At 60fps the 1080 is overkill and stupidly expensive. It's a current 1440p, slightly over 60fps card according to Hardware.fr's benches (yes I'm an actual dirty frog). Starts here.
http://www.hardware.fr/articles/948-15/benchmark-anno-2205.html
>>54638125
>slightly over 60fps
not overkill
in a couple of years it will not even do 60 fps in modern games
Does this mean buying 980 Ti right before the 1080 announcement was a good or bad move?
>>54638171
bad. price is dropping to 550 now
>>54638170
>slightly over 60fps
At 1440p for most greedy games with nearly everything maxed (less greedy games go at 90-110fps maxed), not at 1080p. Not dumping 790€ (no kidding, it's on the Intro page) for that. Again, I'd be looking for a comfortable 1080p/60fps card.
>>54637155
the ti is quite a bit more than the 1070, dinkleberry.
>>54638201
yeah the gtx 1070 will probably have better value for money for 1080p
>>54638028
That's because barely anyone has a 1440p monitor retard. Video game graphics isn't some thing that just magically gets better. The developers chose the level of graphics so that "ultra" with a top of the line card will run at ~60fps on the most common resolutions. Once 4k monitors become common, then games will be made to run at those resolutions, regardless of if we've had 2x or 20x performance increase by then.
>>54638049
Same goes for you, "future proof" is such a stupid phrase. The only way something gets "future proof" is if technology stops advancing.
>>54638077
>>54638125
Are you retarded?
You have to realize that review benchmarks are run with 4xAA at 4k and other silly settings that nobody in the real world would actually use. There are graphics settings on modern games that literally cut your fps in half but provide virtually no visual difference.
Cards as expensive as a 1080 are meant for 4k gaming. For 1080p it's extreme overkill.
>>54637676
>2 1080s make 2160
how much is this worth (when it drops in price) when 1080 comes out?
>>54638229
We'll see, considering >Europe the card will most likely be sold at 500€. Gotta wait to see if some will panic sell their 980Ti. Seems unlikely given the prices and the "meh" factor around the 1080 now that the benches are out.
>>54638384
>try witty post
>delete it
>>54638295
doesn't matter how expensive they are, they're not ready for 4k
>>54638361
>>54638171
You guys are so fucking retarded.
1. All you need to do is look at an aggregate benchmark and divide the score/fps of the card with the price. The one with the highest FPS/dollar is the best. It's literally that simple. If they're even, choose the newer card since it has more features that will be relevant for longer
2. If you had any experience, not just with gfx cards but technology in general, you would know that last gen pretty much never drops to competitive prices when a new gen is released.
>>54638435
Even a 390 can do 4k @ medium settings.
A 1080 can easily do 4k on max settings with the only exceptions being a few unoptimized games in which you have to maybe turn the shadows or grass detail down to high instead of ultra.
>>54638473
>competitive
There is no competition.
>>54638424
>wow someone made a mistake
>shit never happens, what a retard!
>>54638492
Competitive against the new generation from the same company retard
>>54638491
>unoptimized
Most people who use this have no fucking clue about anything.
Path-tracing is some of the most optimized rendering code there is, yet you will get about 0.1 FPS on consumer hardware
>>54638491
>A 1080 can easily do 4k on max settings with the only exceptions being a few unoptimized games in which you have to maybe turn the shadows or grass detail down to high instead of ultra.
sounds delusional as fuck
if you can do it 60+ fps on relevant games i'm sold
Which one?
>EVGA
>MSI
>GIGABYTE?
and why?
>>54638557
The one that performs the best once reviews are out retard.
Don't be a brand-cuck
>>54638557
gigabyte because evga is shit (only good for customer support and tradeups, not for the hardware itself) and msi is shit tier chink shit (people only think msi is good because they had semi-fanless gtx 970 but with coil whine for cheap)
>>54638577
not buying a 1080 meme card, I'm going to buy a 980Ti + price drop
>>54638592
1070/1080 is mandatory for vr and multi monitor at least
>>54638536
It can 4k@60fps every single game on the market on at least high settings. 97% of games on ultra.
People have been running 5760*1080 monitor setups on a single 680 or 7970 5 years ago.
>>54638617
>People have been running 5760*1080 monitor setups on a single 680 or 7970 5 years ago.
lmfao surely you're having a laugh m8
maybe on old games but not modern (even 680-7970 era) games
>>54638617
also
>It can 4k@60fps every single game on the market on at least high settings. 97% of games on ultra.
funny how nvidia demoed at 1080p then, you don't think they would have mentioned 4k gaming?
>>54638610
VR is meme, and so is multi monitor shit
>>54638615
I can get a 980Ti for less than 500 euros
>>54638536
I run 980 SLI and I run most of my games at 60 FPS. Most run at 40-50 on Ultra settings, so it doesn't take a huge bump. A 1080 will perform very similarly when overclocked.
The only time I have issues is shit like in tomb raider where geothermal valley runs like absolute ass.
A good comparison point would be a 380 running 1080p. If you overclock or dial down settings 60 FPS is very achievable in nearly every title on the market right now.
>>54638682
Which is still worse FPS/dollar than a third party 1080 retard
>>54638666
That's because Nvidia is demoing the card to retards.
On their actual product page they even listed the recommended resolution for the fucking 980 as 4K, let alone the 1080.
>>54638691
gtx 1080 is barely better than 980 SLI, 40-50 fps is not ok, that's what you'd expect after owning it for several years, not on launch day
>>54638698
are you sure?
>>54638642
Are you 12 years old? https://www.youtube.com/watch?v=ctOR-Bf-3Ps
>>54638666
Nvidia's marketing strategy is to convince poorfags with cheap ass 1080p monitors to spend all their money on a new GPU instead of buying a new monitor.
980
or
1070?
Performance wise, not price
>>54638725
The GTX 1080 is on average 5-10% slower than 980 SLI. 980 SLI also overclocks ~15-20% while it remains to be seen what kind of performance gains the 1080 will see from overclocking. (Due to lower IPC)
40-50 FPS on ultra settings is perfectly playable. 60 with minor reductions is absolutely playable.
In the history of PC gaming it has never been common place for high end videos to run perfect 60 FPS on brand new ultra settings titles. Maybe we had a couple years in the twilight years of the PS3/360, but go back a decade and that wasn't the case.
Here's a good example:
https://www.youtube.com/watch?v=CUt3z9Hr6RA
This is the kind of performance a high end 2006 PC would have got you in the most common resolution.
>>54638786
1070
>>54638786
>500€ for the 1070
If it doesn't floor the 980 (which shouldn't be hard), NVidia can put that card up their ass.
I can get a 980Ti for €475. should I?
fuck i hate /g/
bunch of poorfag kids that can only see the world from their own perspective
you don't think a rock solid 60 fps is nice? you don't think 120+ fps is nice? stay mad at those sour grapes
>>54638732
It's outperforming the Ti by almost 20% on average. And that's the reference design on a card that looks very overclockable
MRSP of 600$ -20% is 428€.
Did you drop out of high school?
>>54638829
MSRP for the 1070 """should""" be around 500€. If the 1070 is on par with the 980Ti...
>>54638850
>MRSP of 600$
>using the performance of a 699 card
>using the MSRP of 600 as example
Is 1060 a thing? Currently have a 960 and it's mostly good enough, probably will switch to 1070 though
>>54638857
>If the 1070 is on par with the 980Ti...
see >>54638732
the 980Ti is almost on par with the 1080
>>54638850
>20% on average
wrong. see: >>54638732
>>54638892
not yet, eventually
>>54638873
>a 699 card
That's the reference design I was talking about retard, the 3rd party ones will be both cheaper and faster because nvidia decided to cash in on early adopter faggots.
Let me guess, you haven't actually read anything but you still go on /g/ and call people stupid for knowing more than you
>>54638916
that was the post I was quoting retard. Do the numbers yourself
Can't you guys count? We know IPC of Pascal, we know core and bus of 1070 and we know number of cores.
It's trash.
>>54638894
>the 980Ti is almost on par with the 1080
According to HW.fr benches (yes I'm that filthy frog) with all the superfluous stuff at 1440p, it's 30% more performances than a reference 980Ti, and 18% over the EVGA SC+ version. Obviously if you live in Europe you get the short end of the stick pricing wise anyway, but 475€ new seems like a good deal regardless if the 1070 is in between the 980 and the 980Ti for 500€ (which wouldn't be a good value IMO).
>>54638077
Your eyes can't tell the difference between 60 and 120 fps.
>>54638986
but his autism can
>>54638986
for you
linus did a test on it too, he can see the difference
>>54638071
A 980ti still cant run modded skyrim at a solid 60 frames at 1080p. The 1080 will be a great skyrim card though due to the memory advantage, as skyrim can eat upwards of 16gb vram. Def go for the 1080
on youtube, i instantly notice if it's 60 fps video instead of 30 fps, the difference is dramatic even for video, and for games it's not only about the visuals, it's about latency too and the "feel" of the game
And I feel completely fine with my GTX 950 that cost me the equivalent of $75 in a store liquidation here in beanland.
>>54639021
>modded skyrim
>this meme in the current year
>>54637902
Jaggies haven't been a thing since maxed out AA at sensible 1080p screen sizes. 4K is a fucking scam. Enjoy your washed out medium detail textures and "playable" fps.
>>54638996
>>54639004
This reminds me of the audiophile madness. Many people claim they can distinguish 44 kHz and 96 kHz.
Soon you might need some gold-plated DP cables.
>>54639075
Not even memes. Truly that game eats videocards like nothing else, and is more intense than any other triple A release in the past five years.
>>54635923
And this all relies on the board partners being chill with taking $379 after watching people gobble up the shitters edition for $450. Realistically, it's going to be $425. Calling it.
>>54639125
You did that on purpose just trying to get some ``muh OCD'' reaction, didn't you?
>>54639118
jn.physiology.org/content/83/6/3548
stay mad at them grapes retard
>people will buy the 1080 for 599/699
>the 1080 Ti will cuck them all
>but will cost $1K
>>54639157
Well someone fell for it ;P
Why wouldn't I just wait for Polaris?
It will play all the games on very high/ultra settings for a bit less + I can pair it with a great zen for no bottleneck.
>>54639125
it's a shitty game and the mods are made by amateurs
>>54639182
>Why wouldn't I just wait for Polaris?
because you're a retard itching to waste his money and can't wait for the 27th
>>54639182
>Why wouldn't I just wait for Polaris?
if you're a poorfag the polaris might be ok but wait for the benchmarks
>I can pair it with a great zen for no bottleneck
zen is a meme and you should spend more on the gpu than the cpu for games
>>54639200
>$0.02 has been deposited to your shill account in bank of pakistan
>>54639084
>maxed out AA
The best AA method is super/downsampling, which is literally rendering at 4k (or even more) and shrinking to 1080p. That is the main thing that makes "bullshots" look so much better than the actual game. It also improves all the shaders, which traditional AA doesn't.
Here's a downsampled ingame screenshot. No AA method can achieve the same result.
>>54639126
I mean for one thing, just look at that gap disparity. Why the fuck would it be bigger between the 1070 and 980ti (which perform almost the same) than even the 980ti and 1080, which are leagues apart. Rescinding my initial claim of $425, good non reference coolers of the 1070 are probably going to end up going for at least $440. Especially when the vast majority of people prefer non reference.
Bold move by nvidia, generating hype with awesome pricing and then putting the responsibility on the partners to actually meet that price while they exorbitantly milk the teat.
I do hope I'm wrong, I want cheap performance. But let's be realistic.
>>54639084
Then why aren't you playing on 320x480 with everything including AA maxed on a 200$ computer, retard? Think of all the money you can save
>>54639283
Well thier sales will suffer if this is true, with people opting for used thingies.
>>54635923
>He thinks the 1070 will be only $379
I bet you believe in the Easter Bunny as well.
>>54639226
you can make a bloated game with any number of features to grind any system down to a halt, so what
>>54639238
>>54639084
In case it wasn't clear, you are fucking retarded if you think rendering at a high rez and shrinking it (1080p) looks just as good as not shrinking it at all (4k).
Do you think upscaled 720p and 1080i is awesome too?
>>54639335
wasn't the 970 around that price
>>54639313
Or the used prices will jump back up. Bargain hunting is become a dead market with too many bargain hunters out there. People are greedy as a whole, and we only seem to be perfecting it as time passes. Granted, with more subtlety and less direct cruelty, which is good, I suppose.
>>54639369
No retard, I do not think that. I understand how AA works. In case I wasn't clear, being able to play at a smooth 100+ fps is more important than your precious native 4K. Wake me up when we can actually enjoy 4K.
>>54638786
1070 has 75% of the CUDA cores of 1080 + slower memory and lower clockspeed. So at best it'll perform 25% slower than 1080. You do the math.
>>54639313
>>54639416
>>54639226
>cock-loving faggot posting gay shit
This isn't /g/ for /g/ays. Go back to >>>/y/
>>54638809
It's already confirmed to be faster than a 980ti
>>54639445
>smooth 100+ fps is more important
depends on the game, but I also prefer framerate over resolution in general
>Wake me up when we can actually enjoy 4K.
Why? You said 4k was a scam, have you changed your mind to agree with me now?
Also you already can "actually enjoy" it in pretty much every game pre 2008
>>54639416
the question is if it will have the raw horsepower to even do whatever the cancer is that you're doing with all that memory
not even modded skyrim looks all that great, does it?
>>54639447
nice reading there buddo
>>54638924
>the 3rd party ones will be both cheaper and faster
We do not have any actual indication of their pricing other than Nvidia's recommended price and the price of the reference card at $699. They might actually get priced at the MSRP, but it wouldn't surprise me in the slightest if the actual price turned out to be higher.
>>54639283
Samefag, by the way. To add to the nvidia 5 year plan, they're probably hoping the customers blame the board partners. Regardless, it seems like a win-lose-lose situation is inevitable, at best. I just hope we as consumers end up being the winners.
I really hope AMD brings some good specs and prices to the table. Even with first polaris being mid to low end, if it's priced cheaply enough they could shake up prices.
>>54639370
It was $349. These morons are talking shit as usual.
>>54639490
Waiting for the benches.
>>54639490
See >>54639447
It's going to be slower than 980ti, and Nvidia can shove their marketing slides.
>>54639487
I only like dicks when they're on women, faggot
Cute, thick feminine dicks that makes your mouth water from looking at them, not these homo-ass gay dicks
>>54639567
why don't you post some screenshots instead of that fag shit
and i don't think you realize just how much 4-16 GB is even when it comes to uncompressed textures
>>54639524
>Let me guess, you haven't actually read anything but you still go on /g/ and call people stupid for knowing more than you
Thank you for confirming that I was right and that you're a retard
>>54639509
4K is a scam now, that's what I'm saying. Obviously it's theoretically superior, I'm not denying that. Historically early adopters paid a steep price for new tech but got something tangible in return for it. I don't see that benefit yet for 4K when frame rates are still so low. The trade off isn't worth it, not to mention UI headaches.
>>54639524
Yeah, because graphics cards end up costing 30% more than the MSRP all the time. Is this your first card launch?
>>54639598
Are you genuinely incapable of comprehending that MSRP does not necessarily mean the price at which they'll be sold?
>>54639645
>what is the gtx 970
>>54639645
If you have any numbers from another launch where the reference cards were intentionally sold at above MSRP, by all means show them.
>>54639627
>4K is a scam now
I've been downsampling from 4k @120hz for 3 years retard. It's better in every way. If there was an actual 4k monitor that could run 120hz it would be even better.
>>54639645
They'll sell at whatever price the market dictates. Retailers don't give a fuck.
>>54639706
what are the mods even doing, just post a freaking screenshot, and you do you have any idea how much 4 GB is, like your png image, it's less than 1 MB, you could have THOUSANDS of those images on your graphics card, what are you even doing with them?
>>54639688
They specifically explained how "founders edition" is a new thing and priced above 3rd party designs
>>54639732
and the "titan cooler" was always more expensive than the typical plastic cooler so it's entirely plausible that 3rd party cards will be cheaper, they could use custom pcbs and different onboard components etc too
>>54639690
>It's better in every way except frame rate, the metric we actually judge playability by
No shit, fuccboi. I think we're done here.
>>54639709
Which obviously means it's not going to be priced so high that nobody buys it because the 980ti is a better deal
>>54639732
Yes, and if the retailers think that means they can ask more for non-reference designs, they will. Like when 290x prices inflated during bitcoin craze.
>>54639765
>I've been downsampling from 4k @120hz
>@120hz
can't you read?
>>54637900
I'm sketch as fuck, but this looks more legit than most shit. It even has a certain shitty boardroom meme feel to it without actually being shitty looking. Persistent rumors were 390x performance at half the wattage for under $300, so shit. Hope it's real, you mischief maker.
>>54639785
but they can't ask more for them because then people buy 900-series cards instead
>>54639787
144hz is the only way anon
>>54637593
>nvidja
>>54637912
>>54637931
I love you guys.
>>54639807
That's actually a GOOD thing for the retailers. They'll want to get rid of that inventory ANYWAY, now that it's being replaced with Pascal.
I have an Asus motherboard so I have to buy the Asus GTX 1070, I hope it's not too expensive in Euros
>>54639814
>not a multiple of 30
>>54638029
I agree, but they obviously wouldn't have called it that in company. Someone probably added it for context. Compare the red title to the rest of the picture.
I'm 85% sure it's not real. But you never know. Look's slidey and shitty enough.
>>54639860
Retailers don't actually own the merchandise. It's actually like ebay.
>>54639903
you wot
>>54639860
You're still stupid or underage, since at every card launch, the prices of the old gen never quite drop to competitive levels, but retards like you still buy them thinking WOW 30% OFF WHAT A DEAL
>>54639732
Holy kek, buzzwords must be like candy to you. You're the reason people buy "new and improved" snacks that taste worse.
>>54639926
Agreed
>>54639860
This. Like I said, bold move by nvidia. Let's see if it pays off, cotton.
>>54639899
he said
you're a homosexual.
>>54639927
Sound like you bought a card after clearance, when they jack it up for SLI/Xfire morons that buy out of producation cards.
>>54639927
>the prices of the old gen never quite drop to competitive levels
What does that have to do with anything? The retailers don't care about "competitive prices" they care about actually selling stuff. If they can achieve that without actually selling things for what they're worth, so much for the better.
>>54639929
You can't read properly. I put it in quotation marks to show I'm quoting nVidias marketing gibberish. Also you can try following the quotechain to understand context before posting your knee-jerk reaction
>>54639986
>What does that have to do with anything?
Because the entire argument started with retards thinking this is a golden opportunity to buy a 980ti. Read before you post
>>54638557
Not gigabyte, at least not fucking windforce. Their fans and redesigns are usually balls these days. EVGA is great, especially if you need what's basically a really fucking good reference blower. Solid, expensive. B stock used to be a viable option, pretty swamped these days. MSI is never a bad bet, red or green, unless you're talking instructional videos. ASUS is pretty good for green, not so much for red lately it seems.
Sapphire is undisputed champ on the solely AMD side. PowerColor apparently came out with some shit that was incredible for the price, but I was always wary. XFX used to be awesome, heard design is still decent to great, but customer service has slipped and and i think no more lifetime warranty.
Should I sell my 770 and 780 Ti SLI now? Or is it too late for that?
>>54640240
What about Visiontek
>>54640436
Love how they put the last guy who bought the 390x on the cover.
>>54640249
I think too late, 780Ti already drop on 200 €. I don't think they drop below 180 after the new cards release, they are overpriced and don't give more performance. I'm hoping to get 220 for my heavy OC Gigabyte model.
770 SLI is garbage if its 2 GB. And DONT buy 1080. Buy 1080Ti. And wait, there are rumors the Ti starts at 1000 €, but I think they will release the Titan first.
>>54640524
I'm buying two 1070s to replace my 970s when they're out. My 980 Tis will wait until the 1080 Ti inevitably launches, which they will to plow Vega out of the fucking water.
>>54640554
>970 SLI
There is your problem. 1080Ti is superior to 1070 SLI, look 970 SLI vs 980Ti. Even if they don't have that memory problem.
Why do you have 980Ti and 970 SLI?
>>54640577
I have 970s and 980 Tis in SLI, which replaced my 770s and 780 Tis in SLI.
>>54637660
It will be close to the 970, not 1070
>>54637900
>Vega (HBM2)
>under $500
Nvidia is fucked lol.
>>54640597
>970 performance for $300
>in the age of 16nm Pascal
Why does AMD even bother?
>>54640594
Why do you then have the old cards? Never have good cards lying around. Even 770 gave up to 180 € some weeks ago, but its over now, many people sell their old shit because of the Nvidia marketing. I'm still hoping to get a 980Ti cheap around 300 €.
I'm running 4K on the 780Ti, which is mostly limited by the VRAM.
>>54640631
>Why do you then have the old cards?
Because I replaced them once Maxwell came out
>>54640645
I think what he's asking is why you STILL have them and didn't sell them much sooner.
>>54640869
I don't normally sell my old PC parts, but having cards that old not even as back-ups is a waste of shelf space. I also need to make room for my old ASUS Z97 and i7-4770K once Kaby Lake comes out as well and I swap my Devil's Canyon into my secondary gaming PC.
>>54640621
Wut. Do you know what a price point is? Let alone price/perf.
>>54640597
>Close to the 970
Which basically means near 980, over or under it for less money which isn't hard since the 980 is highway robbery and AMD basically has no choice until next year.
>>54638259
>The only way something gets "future proof" is if technology stops advancing.
One word: consoles.
A 4GB 380 will take you all the way to the release of the PS5, which of course will be years after the PS4K.
>>54640597
>It will be close to the 970
Are you fucking retarded?
AMD already has a $300 card faster than the 970, you idiot. It's the R9 390.
Do you think they will release a new card that performs the same at the same price point, brainless worm?
>>54642121
>Do you think they will release a new card that performs the same at the same price point
Are you saying that AMD is actually competent and not run by blind, deaf, and lead-drinking buffoons?
Top jej
>>54642145
Compared to Nvidia, which brought us the 3.5 GB disaster, gimps older cards to force their customers to upgrade, and has every game their gameworks program touches become a pile of crap, AMD does seem like a pinacle of competence.
>>54635923
wow. they're really trying to milk it huh
amd will still beat it
>>54640645
Yes but you sell cards you don't use anymore.
Pro-tip: Donate me at least one of the 780Tis, so I can make SLI. The VRAM will kill it but at least I can try to run Battleborn above 30 FPS.
I just got this 970 for $200. Did I do good?
EVGA already announce a ricer version of GTX1080 for $699.
GTX 1080 price for Nvidia partners will most likely hover around $650 region.
>>54642738
>200
>970
I would.
>>54640554
sli is no good, especially these days, get a single good top end gpu
>>54642738
>$200
fucking good
>spending over $100 on a video card
Go watch movies, kids. Your graphic clusterfuck "games" aren't good and not worth paying a shitfuck of money to play.
>>54643446
>movies
Capefag
>>54643446
you forgot that they only last 8 hours and have no replay value or if they last more than 8 hours they're a broken mess that need the modding community to fix them.
![irreversible-movie-poster-2002-1020668432[1].jpg](https://i.imgur.com/y0dmgPv.jpg "irreversible-movie-poster-2002-1020668432[1]")
>>54643556
>only capefag movies exist
>there isn't a vast library of movies that span almost 100 years
weird.
>>54643446
shitty bait
going to the cinema is like $20+ each time and it's a lackluster experience that only lasts a couple of hours
>>54643557
>if they last more than 8 hours they're a broken mess that need the modding community to fix them.
what is final fantasy
>>54643581
fuck off tard, movies and games are apples and oranges, so what if you like watching movies, fuck off idiot
>>54643590
>final fantasy
>sequels
>to a game that shouldn't have had sequels
>that is shit
>that is just japanese weebshit that tries to hard
it's a shit series by the way, you know because it's absolute shit only played by weebs and tryhards
>>54639881
>not a multiple of 24
Sure is pleb in this thread
Can't wait for the 16 viewports in hardware.
I have a GTX 670, should I upgrade to 1070 when it comes out? Is 1080 worth the extra money?
>>54644731
For VR apps that support the new 16 viewports in hardware... definitely. The 10xx series will blow the fuck out of every other thing out there. Imagine no performance hits for 4 view ports used for pre-distortion and final mapping for stereoscopic vision. Nothing else will be able to touch this.
>>54644731
I say wait till the generation after this to upgrade unless you really want to do VR
>>54637569
u tell me fermi
>>54644776
>unless you really want to do VR
Why would you not want to? The 180/360 porn alone is worth it. Flying an F-15 from inside the cockpit? VR is definitely what you want.
>>54645239
>paying $40 for standing up and playing wii games
https://www.youtube.com/watch?v=fQ0JTkBZdyw
>>54636352
>>54636352
TWO HUNDRED AND FIFTY SIX BIT MEMORY BUS
>>54638557
Zotac
>>54645281
>Memory technology sees a major update with NVIDIA's adoption of the GDDR5X memory standard. The memory is clocked at a staggering 10 GHz effective, at which speed the GPU has 320 GB/s of memory bandwidth over a 256-bit wide memory interface. Such bandwidths were only possible with 384-bit or 512-bit GDDR5 memory interfaces. First-generation HBM being restricted to 4 GB, HBM2 not being readily available, and HBM-class stacked memory being more expensive to deploy (using on-chip silicon substrates) could have contributed to NVIDIA's decision to go with GDDR5X for the GeForce GTX 1080.
>>54645362
it's never at the full 10GHz, that's the effective memory clock.
That being said, Nvidia could just be talking shit out their ass again like they usually do (remember FERMI)
aoj
>retard
>>54645239
>being an early adopter
>>54637900
>classified screenshot
we /hackerman/ here
>>54643446
Mass Effect is. I watch movies too but the new ones are shit like the new games.
>check on amazon
>800$ with shipping include
>check local prices
>1100$
>300$ to buy locally
fuck the country and loyalty and all that stupid shit I'm not getting robbed
>>54643586
>going to the theater
The real bait here is you
>>54643606
>it's a shit series
>Stop liking what i don't like
>Memes
>weebs
This post is full of troll and it's fucking painful. (you)
![tumblr_m5o1u5nbog1rv4c4zo1_500[1].jpg](https://i.imgur.com/RavTzGS.jpg "tumblr m5o1u5nbog1rv4c4zo1 500[1]")
>>54645764
>being this butthurt about the cancer that is JRPGs
plz don't stab with gunsword
>>54635923
you guys think the 1060 will have 6gb vram? Or should I just buy a 960 now?
>>54646865
get the 4GB version of the 960, if not, get a 380 or even a 950 or 370x
>>54635923
look im just gonna get another 980 and sli them :^)
>>54647238
Kikevidya cucks
EVERY
FUCKING
TIME
>>54647238
async compute is not a crucial feature, it doesn't add anything, it's only performance-related. what matters is the end result.
amd gets a slight perf increase because of its shitty drivers that underutilize the gpu resources, while nvidia get little to no perf increase because the hardware is already being utilized quite well.
>>54645726
then the same argument about spending more than $100 on a graphics card could be made about spending more than $1000 on a tv. for a htpc you need decent hardware anyway, people said they were buying gtx 1080 to watch freaking anime
>>54647818
>async compute is not a crucial feature, it doesn't add anything
so you're saying something that distributes computing power between it's cores more efficiently doesn't do anything?
>amd gets a slight perf increase because of its shitty drivers that underutilize the gpu resources, while nvidia get little to no perf increase because the hardware is already being utilized quite well.
how come it looses frames and gains nothing in DX12 and vulkan then? (^:
>>54647869
it doesn't add anything tangible to the game, it's only a performance tweak that works for some cards (amd) and doesn't work for others (nvidia). nvidia cards already utilizes their cores more efficiently so async compute is unnecessary. and the perf increase on amd is overhyped anyway, what matters is the what performance you get out of the card in the end. and async compute fucks with the cache because you're juggling multiple things at once instead of doing one thing really efficiently. async compute is a total meme.
>>54647923
>doesn't add anything
>improves AMD performance
You're contradicting yourself. AMD even gave the Mantle source to Nvidia for free so they can adjust their drivers to work with it, they refused.
>Async compute fucks with cache
it also utilizes more of the PCI-E bus since all the cores can communicate with the GPU instead of just one core/thread doing all the work till it needs help from other threads
>>54647955
>improves AMD performance
it doesn't add anything to the game, all it is is a performance tweak, if nvidia performs better without it, it's no problem to turn it off, nothing of value was lost
>>54647984
>it doesn't add anything to the game
>it also utilizes more of the PCI-E bus since all the cores can communicate with the GPU instead of just one core/thread doing all the work till it needs help from other threads
>adds nothing to the game
>>54648017
kill yourself idiot
it doesn't add any more space ships or projectiles or anything, fuck's sake
>>54648035
>improves performance by utilizing both the CPU and GPU making sure they communicate and use the PCI-E bus
>doesn't add to the gameplay experience
>more frames
>a bad thing
>>54648035
>it doesn't add any more space ships or projectiles or anything
kek
>>54635923
>2016
>plan to buy pascal
>don't wait to polaris
>>54648045
fucking autist, if the gtx 1080 is more powerful than for example amd polaris, it doesn't matter if the amd card has async compute, the gtx 1080 will still provide a better gameplay experience
>>54638733
Lmfao this so true. I run 2 290's which cost me £450 combined and at 4k I can get 60 fps on any game if I turn down some settings. Naything more than 2x aa at 4K is completely pointless and some other useless shit. Looks a lot better than running at 1080p and I spent less on GPU's than nvidiots lol.
>>54648066
>Both SLi and CrossFire have an additional frame of input latency since they use AFR, Alternate Frame Rendering.
stay cucked
>>54648064
>1080
>not just a 20% improvement as usual
>has all that computing power behind an abstraction layer that only works with gameworks, CUDA and random Nvidia proprietary shit
>using a single thread to send info the to GPU till it gets overwhelmed and needing another thread
>inefficiently using the PCI-E lane
>being this mad that Nvidia can't do something AMD did 5 years ago
>>54648101
>1080
>outperforming any AMDshit by far, almost on par with the 2.5x as expensive, 2x as power consuming amd pro duo in frame rate with less latency and microstutter
ok idiot
>>54648063
Waiting for some fucking 120W card, expects it to beat pascal.
Fucking AMDrones everyone
>>54648132
I mean, that's what they said for Fermi, look how that worked out
>>54635923
>970 matched 780ti without a die shrink
>1070 with a die shrink will not match 980ti
AMDcucks actualy believe this
>>54637485
>>54637539
Polaris 11 is the laptop chip that runs on less than 50W.
Polaris 10 is a 390x on a 256bit bus for $250 with half the power usage.
>>54648156
>last years performance with low power consumption
That looks really attractive, time to throw some money at that sidegrade
>>54648131
I'm not saying the 1080 is shit because it can't outperform AMD
I'm saying their tech is shit because it's an inefficient use of it's resources.
When the 390x beats the 980ti in DX12 benchmarks when there's no gameworks cancer in it, something's got to give.
>>54648155
It's just going to be business as usual, I knew Nvidia was talking shit when they said the 1080 was more powerful than a titan X, it did the same shit the 900 series did to the 700 series.
>>54648101
>being this salty
>>54648195
>I knew Nvidia was talking shit when they said the
1080 was more powerful than a titan X
But they weren't talking shit
Why do you retards keep posting Ashes benchmarks?
The developer confirmed already that async compute is disabled on Nvidia cards and this includes the 1080, they don't have a 1080 to test the game and improve on.
>>54648224
>1080 was more powerful than a titan X
it wasn't, it can't even beat two 980s in SLI.
It's literally doing the same shit the 900 series did to the 700 series.
the 1080 is going to beat the 980ti, the 1070 when OC'd is going to also beat the 980ti.
and they'd gobble up all this high mid tier money till they release their $1100 titan card.
>>54648247
>it wasn't
Yes it is
see:
https://www.youtube.com/watch?v=-ZG54Da2pMM
>it can't even beat 980's in SLI
Wait what, it nearly beats it's predecessor in SLI configuration and that's not a feat?
Let's see AMD's 490x beat two 390x's in CF, do you honestly think this will happen?
>>54648238
it's the only straw they have left to grasp for
>>54648247
A 980Ti is only behind a Titan X by few frames in fact some 980Ti can get overclock better than most Titan X cards out there.
Is 1080 new architecture or is this just an overclocked/rebranded 980ti? Just like how 980ti is just a rebrand/overclocked 770
>>54648210
fucking rekt
>but m-muh async compute
>m-muh efficiency (still shit both in performance and in power consumption)
>>54648295
it's a completely new architecture, it's what we've been waiting for all these years since the 600 series
>>54648300
That's the 1080 rekking AMD's top card by 25% in literally the most favorable game you can find for AMD's architecture.
This is a game that is so heavily favored for AMD cards that a 390x matches a 980ti.
Yet the 1080 shits all over the fury x.
>>54637660
>r9 380 will be 200 shekels
>with lower consumption
>and will be as good as a r9 390x
What is wrong with this ?
>>54648238
Async was off for the Nvidia cards in the benchmarks and Gameworks specific shaders turned off for AMD Cards.
>>54648272
290x was pretty close to the 7990 when it came out.
it's a usual 20% increase as usual.
>OC'd
>obviously not going to beat previous gen SLI'd
I'm talking about stock clocks. I wouldn't even go into OC yet since the drivers aren't finalized for that yet. It may even beat the titan X after that but when watercooled it won't stay at the OC clock for over 10 minutes.
https://www.youtube.com/watch?v=w-3fi1ovAP0
they said it looks good in benchmarks but until the drivers are finalized to stop it from fucking up after 10 minutes I wouldn't call it a victory yet.
>>54648295
It's completely new and has nothing to do with the old maxwell arcitecture.
>>54648330
What's wrong with it in my case is that I'm already at that level of performance. I'm not going to upgrade just for better power efficiency.
AMD does not offer anything for people who are on last years high end cards, like Nvidia does with the 1070 and the 1080
>>54648345
the 1070 and the 1080 aren't their high end, it's their high midrange. the 1080ti would be their high end
>>54648336
It already beats the 980Ti and Titan X with beta drivers, it will only get better with new drivers ;)
The current driver(press only) is 368.16, 368.20 should be released the 27th.
>>54648336
>I'm talking about stock clocks
Those benchmarks in Digitalfoundry's movie ARE on stock clocks
If you want OC clocks look here
https://www.youtube.com/watch?v=Vm6kiLG1dTw
The difference with the titan x is even more massive here.
Really you're just spreading misinformation
>hurrdurr 1080 is not faster than the titanx
What fucking planet do you live on.
>>54648356
No that will be their enthusiast card, along with the new titan card
>>54648356
>>54648370
it's just a silly label, it'll be their best offering for a while, i'd call the gtx 1070 mid-high range and gtx 1080 high end/enthusiast
>>54648393 (Me)
gtx 1070 mid-high range
gtx 1080 high end
gtx 1080 ti enthusiast
titan enthusiast
going for polaris 10 just because freesync monitors are like $200 cheaper than gsync ones
>>54648359
>>54648361
Huh, mixed up the 980ti SLI and the titan Z for the Titan X. My bad.
They really should drop the Titan name to be honest, they were the x85s and the x90s before that.
Honestly forgot the X was the replacement for the black and the Z was the dual GPU version.
>>54648419
Nvidia doesn't even support Gsync anymore with the pascal architecture, they have their own framesync shit built into their GPUs that cap the framerate to whatever you want.
Which is what they should've done in the first place instead of Gsync shekel mining.
>>54648437
Yeah but you're blaming Nvidia for your mistake, they never said it would be faster than a titan Z.
As you can see here
>>54648272
it's almost as fast as
980sli which is very impressive
In some games it will beat it, in others it won't, and even then we know that the Single GPU option is the better one, because some games don't have proper scaling or no scaling at all.
And we aren't even talking about frametime differences, latencies will be higher with a dual gpu setup.
I'm curious to see if the new high bandwidth sli bridges of Nvidia will make a difference and increase scaling in SLI.
>This thread
>>54648462
>I'm curious to see if the new high bandwidth sli bridges of Nvidia will make a difference and increase scaling in SLI.
It does but I don't get why they don't just let the cards communicate between with the PCI-E bus. It's not like they even filled the bandwidth of the PCI-E 2.0 bus yet
>>54648325
the fury is still on 28nm, you retarded piece of dumb shit
and it still almost beats it
>>54648452
WrongNvidia's freesync is complementary to G-sync, they tackle different problems.
>GTX 970 will be a long step from the 780 Ti to keep the pricing up on the older cards
oh wait
>>54648482
>and it still almost beats it
Except it doesn't you mouthbreating mongeloid, the difference runs up to 60% in some benchmarks
>>54648484
it's not freesync, Nvidia has Fast Sync now so you don't need any Gsync monitors anymore.
>>54648452
>>54648484
wait... what?
>>54648495
>up to 60% in some benchmarks
>some
>up to
nvidia shill, please go and kill yourself
also say goodbye to 900 series drivers from now on
i'll be here still enjoying my 7950 since 2012
>>54648505
Fastsync, not Freesync sorry.
It's a way of prerendering frames, like consoles do.
https://www.youtube.com/watch?v=WpUX8ZNkn2U
>>54648512
I just gave you the ashes of the singularity benchmark literally handpicked the best benchmark in your favor and even there the fury shit x gets beaten by 25%
Yet you still grasp to this bullshit idea that the fury x 'almost beats' the 1080 in typical gaming loads
You are in full fucking denial and you're freaking retarded
>>54648495
>can't spell mongoloid
i think you're the mongoloid, you oligophrenic illiterate shitposter
>>54648531
so what if it doesn't beat it? it's not like the difference is big, and it's not like the fury is from this generation
>>54648515
so it's more GSync shit? Why is it marketed that they don't need more Gsync shit?
>>54648531
It depends on the shaders, Ashes developers said that specific shaders only benefit Nvidia cards and hit AMD cards hard.
So not all benchmarks are equal, Gameworks is fucking cancer.
When they released the benchmarks themselves they complained that Nvidia wanted them to turn on certain shaders to which they refused.
>>54648482
37% difference on average across 16 games.
>>54648505
gsync fags suicide status :
[X] watch
>>54648515
>>54648502
>>54648484
>>54648452
fastsync and gsync complement each other. gsync is for when your framerate dips below your monitor's refresh, fastsync is for when it's above to eliminate tearing
>>54648549
>Why is it marketed that they don't need more Gsync shit?
They don't said that, only this guy >>54648452
has been saying it.
Vsync only tackles screen tearing, g-sync does that already but makes games run smoother at low framerates by repeating frames.
G-sync is still better since the screen hz adjust to the framerate.
>>54648533
It's not my native language, I don't give a shit
>>54648541
>shifting the goalpost
it's AMD's top card from last generation with HBM
We are comparing it to Pascal's mid tier chip, the gp 104 with DDRX, a die that is almost half the size of that of the fury.
If anything we should be comparing it to gp 102, but that isn't out yet.
> it's not like the difference is big
The difference is roughly 30-40% in most benchmarks and the fury x is only 20% cheaper than a 1080, I'd say it's time to cut prices on the fury shit x
>>54648314
>>54648336
>It's completely new and has nothing to do with the old maxwell arcitecture.
>it's a completely new architecture
Nvidiots actually believe this
>>54648569
It's completely new, it mirrors the new AMD architecture just without Async Compute cores.
>>54648567
>it's AMD's top card from last generation with HBM
That's the thing, it's last generation, not current generation.
And to be honest it wasn't that far off from the 980ti
https://www.youtube.com/watch?v=lg9Pn6MLXjI
>>54648567
it's not even an english word, you retard
>>54648590
>It's completely new, it mirrors the new AMD architecture just without Async Compute cores.
It's just Maxwell on a smaller node, you fucking shill.
>>54648590
>That's the thing, it's last generation, not current generation.
That doesn't matter, we normaly don't see these kinds of leaps of the mid tier chip over lasts years big chip.
The 980 only beat the 780ti by about 10-15% at launch (the differences are slightly bigger now due to drivers).
The 1080, at launch already beats the 980ti at about 30% and even more so against the fury x (37%), and that's without optimized drives out yet.
You can try and downplay this all you like, if you are honest about the numbers you'll see it's a big leap.
>>54648569
Evidence to the contrary?
>>54648600
Again I don't care, go play with your lego
>>54648567
>fury shit x
why do you feel the need to shit all over amd, you retarded fanboy?
>>54648614
nice comeback, you immature little shit
go read a book, you will probably learn about basic spelling, you sperg
>>54648608
>The 1080, at launch already beats the 980ti at about 30%
you mean 20% on average, which is what the 900 series did to the 700 series.
the best improvement was in TDP
>>54648626
>"waaaaah give me some attention!"
Shoo shoo, off you go little boy
>>54648632
>you mean 20% on average
See >>54648553
37%.
>>54648615
>you mean 20% on average, which is what the 900 series did to the 700 series.
No I mean 30%, which isn't what the 900 series did to the 700 as I already explained
>>54648641
>100 - 73 = 37
>????
>Profit
>>54648132
enjoy your shitty memory bandswitch
enjoy your über power consumtion
>>54648680
Anon, do me a favor here and do the math. Show all your calculations.
How much larger is 100 than 73 in percentages?
>>54648680
Is this the level of math Americans are taught in school?
>>54648172
It's great for anyone unwilling to spend more than $300 still running 2012 med-tier cards
>>54648697
Agreed, but I'm not one of them
>>54648689
>>54648652
that's a 27% increase..... not 30 or 37.
common core really is screwing over america
>>54648680
>>54648652
I don't think common core is doing you any good....
>>54648680
Land of the Free educationTM
>>54648705
Here, I'll help you out.
http://www.percentagecalculator.net/
Use the third one. Type 73 in the first box, and 100 in the second one.
>>54648705
Holy fucking shit, I can't believe this
73 is the base, so its 100%
1% is 0,73
Now divide 100/0,73 = 136, 98
So it's 37%
>>54648735
anon...
>>54648689
73*1.37=100(.01)
>>54648696
>>54648713
>>54648715
Fuck your undeserved sense of superiority.
>>54648738
>>54648715
>>54648713
>>54648696
>>54648689
>what is relative
>using percentage increment in a relative chart
>>54648761
Keep going, this is awesome
>>54648738
>>54648741
>>54648761
Keep going this is a great weight loss program since I cant find my sides anymore
>>54648796
Explain why it's wrong and show the math
>>54648802
>lol I epik le trolled you gays xDxDxD
>>54648802
>needing proof why 37 != 100-73
AHAHAHAHAAHAHA
>>54648822
>>54648802
>>54648802
>subtraction
>your strength
it's how much the previous generation stacks up to the new cards, not the other way around.
>>54648835
You're right, they aren't absolute numbers, but already shown relative to the 1080 which is the base, I dun goofed
>>54648835
>>54648830
>>54648822
>>54648814
>>54648802
>>54648796
>>54648772
>>54648853
>4k GTX1080 851.3 (100%)
>4k GTX 980ti 621.5 (73.00599083754258%)
>851.3 - 621.5 = 229.8
>229.8/851.3 = 27%
>What is RELATIVE?
>>54648862
>>54648866
I guess you finally get it now
the funniest part about arguments on 4chan is how confident the retards are. this guy's vicious condescension while fucking up grade school math looooooool
>>54648887
Nvidia is still faster
>>54648877
>I'll post gore so that people can report and get mod to delete this thread to hide my embarrassment
>using a goddam fucking percentage increments
>in a fucking relative chart
>>54648887
>/g/
>technology
>asshats can't use calc to subtract
>divides instead
>>54648898
it's a leapfrogging game, AMD is two generations behind maybe they'll catch up, maybe they won't
>>54648911
>it's a leapfrogging game, AMD is two generations behind maybe they'll catch up, maybe they won't
I think it's unlikely they'll make up for this big of a difference in one generation.
What makes it even more unlikely is that all official leaked information about Polaris so far suggest small die sizes and low clock frequencies.
Vega will be a high end card but it will have to compete with 1080ti
>>54648939
they can be increased to alot, test dies usually have a low clock, I don't think they'll release those GPUs under 1GHz.
With a die shrink Polaris's 480 or 490 may be on par or beat the 1070. It's up in the air right now, but Vega is coming later this year, noone knows the specs of it however.
>>54648955
>With a die shrink Polaris's 480 or 490 may be on par or beat the 1070
I really don't think that will happen, the 1070 has released a big clock speed upgrade compared to the 970 so Polaris 10 would need the same if it wants to match it
https://en.wikipedia.org/wiki/Percentage_point
Important reading for some of the people in this thread.
>>54648980
>percentage point
Nigga, that shit does not apply to a fucking relative.
Did you actually read what you just fucking wrote?
Percentage is not a unit here for fucks sake since we have a fucking reference value for the calculation.
>>54648976
I think Vega will force the 1080ti out, seeing as I keep sing 15-18 million transistors.
Maybe the AMD cards will come out with a decent peak clock and can overclock more stably.
Nvidia has good results right now but it's the same shit, I wouldn't call OCing the 1080 to 2.1GHz stable yet since on a watercooled setup it starts being unstable after 10 minutes.
I say let's wait and see. AMD may be able to fit more transistors on the same sized dies as Nvidia. It's all up in the air as far as I'm concerned.
>>54648866
You just calculated how much smaller 621.5 is than 851.3. So you're right in that 980 Ti is 27% slower than 1080.
However, if you wanted to calculate how much faster 1080 was than 980 Ti, then you'd have to divide your 229.8 with 621.5 rather than 851.3, which would give you the 37%. This is the number we're looking for.
>>54648939
>What makes it even more unlikely is that all official leaked information about Polaris so far suggest small die sizes and low clock frequencies.
Yeah, because the 1080 is such a huge die. *rolls eyes*
Polaris is going to be 9% smaller anyway because it's 14nm, not 16 nm. And considering that Nvidia just put Maxwell on a smaller node with no improvements to their architecture, while AMD completely reworked their GCN architecture, it's not at all unlikely that Polaris will be more competitive than most Nvidia shills want to believe.
>>54649034
It's too early to say anything about vega or the 1080ti to be honest.
If we look at previous generations though, I don't see AMD's top card pushing out Nvidia's top card.
I think that guy with the watercooling setup fucked something up.
https://www.youtube.com/watch?annotation_id=annotation_3215094497&feature=iv&src_vid=bZrGImpdxy0&v=Vm6kiLG1dTw
This guy here gets the 1080 with a reference cooler stable to 2100 mhz without any issues.
Or he might have just received a better bin
>>54649061
>Yeah, because the 1080 is such a huge die. *rolls eyes*
A fair amount bigger than what's leaked so far about polaris 10
>>54649069
most benchmarking channels try to get their shit out 1st and don't do alot of longevity and stability testing.
I can OC my 390x to 1200mhz no problem, but will it last more than 10 minutes before crysis 3 crashes? no it won't.
or it could be that he got lucky in the silicon lottery. who knows.
>>54649061
>while AMD completely reworked their GCN architecture
>gcn 1.4 vs gcn 1.3
>completely reworked
Yeah right, I highly doubt they changed more about it than Nvidia did from Maxwell to Pascal, and even even if they did, let's be honest they needed to.
GCN 1.3 was ages behind Maxwell
>>54649086
Hardware unboxed is a respectable benchmarker, so I trust he did a bit more testing than that.
And he's not the only report I've seen about good oc's on the reference design, in fact all reviews I've read so far reported this.
Only your weird movie where this guy straps on watercooling seems to be the one where its not stable.
>>54649043
Why the fuck would you do that?
That calculation only apply if GTX980ti becomes the reference point for this relative chart, but that is not the case here.
>>54648901
Suddenly, I am the one who made the chart?
Anyways the difference is 31.5% instead of 37 in this case.
Everybody is wrong (by roughly the same amount), it sure as hell ain't 27%
>>54649115
>>54649043
Again, if GTX98ti becomes the reference numbers and others were calculated relative to it.
>851.3/621.5 = 137%
>137% to 100% = 27%
>>54649132
How did you arrive at that number
>>54649115
Because whatever is the reference point in the chart is irrelevant. It should not force you to state things in a silly way.
It is far easier to understand the difference between two cards if you state how much faster the better card is, rather than how much slower the worse card is. It makes no real sense to take the latter approach, regardless of what the reference point in the chart is.
>>54649142
Math and a brain. Try to keep up.
>>54649162
Suspiciously enough you aren't able to show the math though
>>54649101
Honestly I'm waiting for Jayztwocents to OC it and benchmark it since he usually keeps his shit OC'd when he gets them.
on another note, I'm perplexed about Vega keeping the same amount of cores as the Fury X, since the 1000 series dropped core count for more clock speed.
Hopefully you can OC HBM2 better than HBM1 since that seemed to hinder Fury X performance by alot.
I'm hoping AMD also drops the X from some of their cards and only brings them out later on like Nvidia has their TIs but that might be asking a bit much.
>>54649139
>>54649170
So assuming the 1080's score is the baseline (100%), because that's what everyone has been telling me,
Rel% = (851.3 - 621.5) / ((621.5 + 851.3)*2) * 100
= 31.206%
>>54649151
What the.
>Because whatever is the reference point in the chart is irrelevant.
Its not irrelevance because the percentage represent the reference unit.
The chart were made to to simplify thing as it get.
.Its 27% no matter how you calculate it.
>>54649194
It's 27% what?
Are you claiming that 1080 is 27% faster than 980 Ti?
>>54649190
That is the most stupidest shit I've seen all day.
Holy fucking shit, these chart were made for the fucking masses, for the generic audience who read the chart and can come up with an easy numerical number, or mental image for the comparison.
How in the fucking world did you end up with that shit?
>>54649214
No, I am claiming that the differences of power between these 2 gpu according to the chart is 27%.
GTX980ti is 73% slower compared GTX1080, which translate to 27% in differences.
GTX1080 is 137% faster compared to GTX980ti which also translate to 27% differences.
>>54649190
all of my what
>>54649218
What you're calling the most stupidest(sic) shit is a generally accepted forumlae for producing the relative difference between two numbers, in this case the performance scores between two graphics cards.
You can try going here for more information
http://lmgtfy.com/?q=calculate+relative+percentage
>>54649250
>GTX980ti is 73% slower
>GTX1080 is 137% faster
These are actually incorrectly stated. 137% faster would mean a 2.37x increase.
This quote chain started with >>54648632 , who claimed that 1080 beats 980 Ti by 20%. This was corrected to be 37% here >>54648641 . There is nothing wrong with the 37% value here. The 37% value in >>54648553 is ambiguous, as it could technically also mean the %-point difference, but anyone who isn't shit at math realizes that it means that 1080 is 37% faster than 980 Ti, because this is how the differences are stated 99% of the time.
But if you want to continue arguing semantics, go ahead.
>>54649317
>137% faster would mean a 2.37x increase.
IN RELATIVE.
For fuck sake stop missing the goddam fucking point.
>>54649332
You are literally making no sense.
>>54649337
Some of us are arguing with Nvidia shills. Nothing they say will make sense if it makes a product look better.
>>54648553
>27% over stock 980Ti
>10% over OC 980Ti
>700$ vs 540$
>>54649256
But that is wrong, since that calculation only apply to relative point changes.
>today GTX 1080 gets 850 points, but last week I only manage to get 650 points/
This is when your calculation apply, where the things that change is the reference point.
Were not talking about that, were talking about a % in relative to a reference point, where the thing that change is not the reference point, but a number in relative to it.
>>54649337
137% IN RELATIVE TO 100%
>>54649375
Dude just say 37% faster
>>54649190
What a weird calculation
If person A runs at 100 km an hour
and person B runs at 150 kms an hour
=(150 - 100) / ((100 + 150)*2) * 100 = 10
Performing your calculation would net me a relative performance of 10% for person B
While it's pretty clear that Person B runs 50% faster than Person A
>>54649375
>137% IN RELATIVE TO 100%
Ok?
That still doesn't make this statement true.
>GTX1080 is 137% faster compared to GTX980ti
Are you an ESL or what?
>High price for the Founders Edition
>Cooler runs into temperature limit
>Fans don't turn off in idle
>Overclocking more complicated than before
>Could be quieter
>Support for analog VGA outputs
>>54649332
The 980Ti appears to be 27% slower than a 1080
The 1080 appears to be 37% faster than a 980Ti
The words "relative" don't mean a fucking thing, the percentage is still based on hard concrete numbers, it's just there to make one single score a nice even 100 for the idiots.
You don't know math and your retardation makes us all worse people.
>>54649375
Maybe you should make English your first language, chinkboy, before shouting about shit you can't understand.
FUCKING RETARDS ARGUING ABOUT PERCENTAGE! QUOTE FROM THE REVIEW
> Compared to NVIDIA's GTX 980 Ti, the performance uplift is about 30%
SO ITS REALLY JUST 27% YOU FAGGOTS OVERCOMPLICATE IT>
>>54649409
>about 30%
Which is actually much closer to the number stated here
>>54649190
>>54649409
It depends on the resolution. 1080 is 32% faster at 1080p, 37% faster at 1440p, and only 28% faster at 1600x900 in the TPU review, hence the "about 30%".
>>54649390
I'm talking in the context of the chart anon.
If GTX980ti becomes the reference point, and other gpu were measured in relative to it, GTX1080 will be about 137% faster in the chart.
Also 137% to 100% is still 27% anon.
>>54649459
Let me repeat my question:
Are you an ESL?
Because you can't into English.
>>54649459
>137% faster in the chart.
That would mean a performance boost of 2,37 lol
>>54649476
In RELATIVE.
Pic fucking related.
>>54649469
...and?
How is that have anything to do with my argument?
Sure, my english is fucking shaky but at least I understood the fucking chart.
Defend this, nvidiots.
>>54649513
There's a reason why you've been told this several times >>54649476
Your problem is probably your English.
>>54649498
All that chart says is that the R9 295 X2 is 28% faster compared to the 980ti while the 780ti is 30% slower. Any other interpretation is rubbish
Absolute numbers ( probably average framerates) have been used to calculate performance to the 980ti but not between two alternative cards.
That's why you can substract and add between the 980ti and other cards, but not between two cards that do not include the 980ti.
>>54649527
What does performance per dollar get you if you can only get 20 fps in the newest games?
At that point will you be thinking "What an awesome experience, I can really see the performance per dollar in action!"
Or are you thinking "This is shit"
>>54649535
IN RELATIVE.
R E L A T I V E.
Holy shit, yeah my English suck but I at least undestood what RELATIVE is and how it was calculated in the goddam fucking chart.
137% in relative to GTX980ti.
STOP USING PERCENTAGE POINTS AND CALCULATION.
This is a goddam fucking relative chart.
To get to 2.34 points GTX1080 need to perform at about 247% in relative to GTX980ti
>>54649541
Which is exactly why the calculation that ended up with 37% differences is wrong.
Thanks for proving my point anon.
>>54649589
>Which is exactly why the calculation that ended up with 37% differences is wrong.
It is, I think you are mistaking me for some other guy
>>54649382
>no response to this
>>54649589
This is getting embarrassing.
>>54649589
Youre right. Dont bother with autists. Its 27%.
>>54649659
Care to refute anon.
>wow his English sucks
>*reaction image:*
>wow look at this guy embarrassing himself
>lol what a retard
Right, got anything else you want to say?
>>54649678
He simply doesn't understand a relative chart, he treats the percentages in there as absolute numbers and not as percentages relative to the 100% card
>>54649678
See >>54649396 or something.
>>54648605
proof or gtfo
>>54648735
not sure if trolling or just murriken
>>54638049
>Future Proofing
![13963343195_4ae6e76708_k[1].jpg](https://i.imgur.com/2k9wVir.jpg "13963343195 4ae6e76708 k[1]")
![aSENfYTGk8HVwDYNheVq4F-650-80[1].jpg](https://i.imgur.com/CgFCtWH.jpg "aSENfYTGk8HVwDYNheVq4F-650-80[1]")
![seal-gay[1].png](https://i.imgur.com/N8i2wPx.png "seal-gay[1]")
![1419266641149[1].webm](https://i.imgur.com/ZkUS86N.jpg "1419266641149[1]")
![ashes-dx12[1].png](https://i.imgur.com/wdWH4k3.png "ashes-dx12[1]")
![perfdollar_3840_2160[3].png](https://i.imgur.com/oU4Leoj.png "perfdollar 3840 2160[3]")
