So much for reaching 1070 performance with 1600Mhz OC. Review cards aren't making it past the 1400Mhz mark.
>>55256721
'bout what I figured.
>>55256721
AMDrones are completely delusional.
Once the reality check hits them next wednesday they may start killing themselves.
480 at 1266 MHz is about 95% of stock GTX 980 performance. 1400 Mhz is 10% overclock which is not much. That means a GTX 970 at ~1500 Mhz will be just as fast if not faster than a 480. A 980 at the same clocks will leave the 480 in dust.
I'm sorry, but you are just expecting too much from a 200$ card.
>>55256872
S U I C I D E
U
I
C
I
D
E
W A T C H
A
T
C
H
>>55256721
top fucking kek, looks like it's fiji all over again.
I hit 1900mhz on stock cooler 480.
>>55256914
>1400 Mhz is 10% overclock which is not much
It's actually a 30% overclock. The 480's stock clock is 1080MHz. It's why I figured these wouldn't be great overclockers. Not being able to label 1266MHz as the stock clock says they weren't sure they could even maintain that. Having a $200 card at their top also says it was the most they could push out of it. There was no way this was going to be capable of $400 performance with AMD just being kind enough to underclock it and sell it for half it's worth. They lose millions every quarter. They can't afford charity.
>>55257157
Still not bad for a 200$ card if you have GTX 960 or lower.
>>55257211
>The 480's stock clock is 1080MHz
No it isn't, the stock is 1266
>>55256721
i don't think anyone was expecting the stock reference cards to hit 1600mhz. It is the AIB cards with the 6+8 pin power connectors and better cooling fans/heatsinks that can potentially achieve the 1600mhz clock.
>>55257231
>he thinks a boost clock is a stock clock
>>55257276
AMD cards don't have a boost clock you moron
https://hardforum.com/threads/from-ati-to-amd-back-to-ati-a-journey-in-futility-h.1900681/
Kyle was right yet again.
>AMD Radeon RX 480 failed to OC to 1400 MHz
>Like I said 2 days ago, most review samples are having trouble reaching anything beyond 1350 MHz. The highest reported clock is 1379 MHz. The member of the PCGamesHardware team confirmed that their sample did not reach 1400 MHz.
IT'S OVER, AYYMD IS FINISHED & BANKRUPT
AYYMD HOUSEFIRES
AYYMD HOUSEFIRES EVERYWHERE
KYLE WAS RIGHT ALL ALONG
AYYMDPOORFAGS CONFIRMED ON SUICIDE WATCH
>>55257296
nvm, they do, but it's still shit
FINISHED AND BANKRUPT
>>55257304
We doubted him all along
AMD CONFIRMED FOR BANKRUPT BY TOMORROW
>>55257304
So, it's literally all over again.
NVIDIA won again.
Everyone go home.
>>55257298
It must be just a straight rebrand again, except for the die shrink. Meanwhile Nvidia cards are hitting 2000 mhz+
>>55257274
Pushing over 9000 niggawatts through a small die will make it impossible to cool no matter what because there simply isn't enough heat transfer area. It could also be unstable at high clocks because of its architecture. No cooling or voltage will fix that if the signals can't propagate fast enough at >1400 Mhz
>>55257304
No they won't, haven't you guessedo their pattern?
They will nitrogen cool & OC one, test in ashes of singularity and compare against stock 970 to claim victory.
Why does anyone even bother with these threads? What do people get out of this?
>>55256721
are you retarded? the 1070 isn't a mid range card
>>55257384
5 rupees per post :^)
>480 is a shrunk 390 which is just a rebranded 290
>>55257367
>They will nitrogen cool & OC one, test in ashes of singularity and compare against stock 970 and still somehow lose
>>55257421
No matter, you can get two to beat a 970!
>>55257384
Ask the AMD poorfags who were claiming the 480 would have 1080 performance. When you're barraged by that level of baseless shilling for weeks, it's satisfying to see them get their comeuppance.
I CAN'T BELIEVE I TRUSTED RAJEET
When will my 3570K become a bottleneck?
I got one which can't be overclocked...
Should I just stick with my 390 or should I try to sell it for the price of an 8gb 480
>>55257464
definately sell it for that sidegrade bruh
AMDPOORSHIT RX480 confirmed to loose against stock GTX970.
MY FUCKING SIDES ARE IN ORBIT
Told your motherfuckers Samshit's 14nm process is junk and there's a good reason why only AYYMD uses GloFo and the rest of the industry uses TSMC for mass production, but no, lets believe in the Pajeet and AYYMD
>>55257513
i'm crying irl right now
it isn't fucking fair bros.
>>55256721
>AMDrones thinking it would actualy be 1070 levels of performance
>kept telling them if that was true there would be no way AMD would price it at 200$
>actual benchmarks come out and it turns out it's just a rebranded 390 with better power effiency
>>55257541
why would you compare a 200 bucks card with a 450 card?
>>55257541
it's no a rebrand if it has
- different litography
- different shader count
- hdmi 2.0
>>55257560
Because AMD fanboys said it would be equal performance
>>55257452
These just keep getting pushed further and further away from reality, the 480 is supposed to be around about 390/390x performance
>mfw rx480 @ 1.3ghz still trounces the gtx980
>>55257599
Stock 980 maybe, not an oc'd one, Maxwell was one of the best overclocking architectures
>GCN
>8 stage pipeline
>expecting highclocks
How many times do you retards need to be told? The sweet spot for every GCN card ever made has been right around 800mhz to 850mhz, and process has nothing to do with it. Its a trait of the architecture itself, it is not a high clocking arch. Hitting 1150mhz-1250mhz is already incredibly high for a stock clock.
>>55257595
So it's going to compete against a 970 on the best case scenario?
Top fucking lel
>>55257573
well, if it is close, it would be great, if not, why do you care?
it only would be fair to compare it to a similar price bracket card, when and if there is a 1060/1050
>>55257630
Its going to replace the 970 since the 970 isn't in production anymore.
>>55257630
Thats the only point. The R9 390 and GTX 970 are considered to be the base line minimum recommended GPUs for VR. That is entry level VR performance.
That is why AMD stressed the RX 480 as being a cheap entry level VR card. Thats all it is. Everyone making these claims of competing with the GTX 1070 are just FUD spreading children.
>>55257636
It's not going to be close, are you fucking delusional?
only way AMD is gonna get ahead is if they invent the next gen technology for gfx cards
even then nvidia will just copy them and make better cards
>>55257630
If you care about performance any higher than that then I wouldn't expect you to understand why anyone would be interested in the 480
>>55257671
When will AMDrones stop moving the goalpost?
>>55257691
1/10
Rx480 prerelease BIOS are set to prevent more than 1380mhz OC.
After launch expect at least 1600mhz OC for reference cooler cards.
>>55257721
Lol, you shills never give up
>>55257721
>Rx480 prerelease BIOS are set to prevent more than 1380mhz OC.
Nigger what the fuck is the point of that? Especially on the reviewer samples?
>>55257332
Overclocked krippler 3.0 with less ipc than a 980 ti somehow
>1380 mhz drawing 137 watts, memory alone will eat about ~30 watts
>max wattage is 150 with a 6 pin
>undoubtably hitting the board limit with just the 6 pin
wait til the 8+6 pins come out you silly geese
>>55257763
What are you trying to say
>>55257768
TDP is not power consumption, sherlock.
Based on early benchmarks RX 480 seems like a solid card for $200. AMD seems to be focusing on mid-range which would make sense since they are making a lot of mid-range chips for locked down gaming PCs. Also that segment is probably bigger profit-wise than the high end.
>people actually believed currytech
>>55257421
At least in dx11 mode.
>>55256721
>omg, some post on a forum says so
novidias straw man arguments are getting worse with time
>mfw ncucks will reply to this post with amdrone shit
>expecting the 6 pin reference to hit its full potential
what are you OP, retarded?
>>55257755
Prevent leaks
>>55257768
>he thinks they're current regulated
75W + 75W is all they're allowed to draw and still call themselves "PCI-E." It's not all that can be supplied.
>>55257843
>moving the goalpost
We all know where this is going to end
>wait for vega
In b4 "drivers will fix it."
>>55257859
Wait for navi.
>>55257879
If AMD survives that long
>>55257868
They will actually
>>55257894
When will drivers fix 300~ series shitty overclocking?
>>55257894
Please stop, my sides can only take so much.
>>55257666
Like async or HBM? Doesn't seem to matter much
>>55256721
go back to >>>/v/, don't you have a console war to be fighting?
>>55257891
Vega doesn't exist. It's just a shrunken Fiji. Polaris is a shrunken Hawaii.
Nothing but rebranded housefires.
>>55257967
So what is Navi?
>>55256721
What the hell is that guy trying to tell us? Did you translate that with babelfish?
>>55257975
Two shrunken Fijis glued together
>>55257967
Now this is what I call desperation.
>>55257993
So a shrunken pro duo?
>>55257859
>moving the goalpost
>amd sets original goalpost 390x performance
>false flagging nvidiots set new goalpost at 1080 performance
>card reviewers say it has 390x performance
>nvidiots proclaim amd finished and bankrupt
>>55258011
>>false flagging nvidiots
You mean AMD fanboys, who are now backpedalling hard when actual benchmarks show what fucking idiots they were
>>55257461
With a 480?
Nah, not really, depending just how the game is optimized. I got a 390, and the performance levels between my 3570k, and 390 are snuggly paired. Over clocked or not, 3570k is still mighty powerful.
>>55257452
No one has ever said that unironically. Leave that straw man be.
>>55258042
>taking anything posted on 4chan seriously
>>55257612
And at only 200% the price!
>>55258093
at 200% the price you get a 1070 retard
>>55257859
Not him, but I've been waiting for a good cooler and 8+6 pin since day one as well. Maybe you're just dumb?
>>55258140
That would be the case if the problem was thermal throttling, it isn't.
You're just tech illiterate
>>55256721
PCGH are known Nvidia shills. I wouldn't trust anything they say.
>>55258164
No, it's power throttling. Which will be solved by extra power connectors and more generous power targets on non-reference cards, like always. Back to /v/, kid.
>>55258185
>like always
You mean like with fiji and Hawaii?
Oh wait
i expect all your apologies when the custom cooler 8+6 cards come out in 3 weeks
>>55258185
Crashing is not "power throttling," and a 6 pin literally does not care about how much power goes through it. It will keep pumping power until the PSU pops or it melts.
>>55258258
why do 8 + 6 pin cards exist then? apparently they can draw all the power they need through a single 6 pin
>>55256721
He said their card couldn't reach 1400 MHz. Also that their office is air conditioned and that noise tests are done in a sound proof chamber.
They're putting an Arctic Accelero Mono Plus on it next to see if they can reach higher clocks with a better cooler.
>>55258589
PCI-E standards
>>55258258
Nope... The spec for 6 pin connectors literally doesn't allow pushing trough anything higher than 75W similarly to how the spec for the 8 pin connector isn't allowed to push trough anything higher than 150W.
God you n/v/idiots are idiots... Of course a card with a single 6 pin connector and stock cooler isn't going to be pushing 1400 MHz when it's stock clock is at 1080 Mhz. I don't think anybody really expected the reference RX 480 to beat the 1070 regardless of how much it could be overclocked. We already know that a non-overclocked reference 480 beats a GTX 970 and comes worryingly close to a GTX 980. As for how much the overclocking will boost the 480, that's something we'll see once the embargo lifts, but it's obviously going to beat a non-overclocked 980.
>>55256721
But the only rumour I recall about the RX 480 hitting an overclock of 1600 MHz was about "beast mode" AIB cards that would be sold at $299.
1400 MHz with the reference model is pretty much what was expected, isn't it?
>>55258773
It doesn't allow pulling more than 75W and still be called "PCI-E," moron. That has literally nothing with how much current can go through a 6 pin. They're wires. The only regulation they have is "melting."
>>55258827
Sure, you can *technically* pull trough as much as you want until the wires get so hot they burn the insulation and the whole thing shorts, but companies don't go over the limits in the specifications. If 75W is what the spec says is the maximum, then 75 is the highest a card will pull trough a 6 pin connector. It's that simple.
>>55258880
>AMDfags imagining circuitry that doesn't exist to explain not being able to OC over the point where anybody with a brain could tell you the limit would be.
>>55257573
You mean AMD shitposters were saying that to get you to sperg out.
I wanted to try out the red team and get something that performs just a bit better than a 970. That's all it is, and all I wanted.
>>55258880
It's not that simple. Cards will pull whatever they need to reach the clocks and voltage they're trying to hit, even if that's at the boost clock.
The only hard limits are (or at least should be) temperature targets. Nvidia sets Power Targets too but you can bump those up to 20% over the max anyways.
>>55258968
>n/v/idiot doesn't understand official specifications for things and companies building things to follow said specifications
You know whats worse than console war faggots? People that want AMD to fail. For some reason that want Intel and Nvidia's cock rammed down their throat with high prices and no innovation because they have no competition.
Seriously go kill yourselves
We'll have to wait and see if it's a temperature or a power limit. Or maybe Polaris just can't clock higher than ~1400 MHz and we'll all be disappointed again.
The last one seems most likely.
>>55257595
That chart is fake.
Here's the original.
>>55259042
That's the clocks and voltage, moron. At 1266MHz and stock volts the card won't pull more than 75W+75W. This has nothing to do with what it will pull if overclocked and/or overvolted. You can be limited by shitty onboard power delivery, but that's not the 6 pin. The 6 pin doesn't care. Adding an 8 pin will not magically unlock another 1266MHz of overclocking headroom. The 6 pin could deliver that if the card was capable. Nothing cares that it's out of spec.
>Buying reference
Are you stupid? Wait for the AIB's they will demonstrate whether it is really capable of good overclocks or not. I will take the reference card results with a pinch of salt but they should still be competitive for a midrange card. Nvidia are still trying hard to make out AMD fans think it is some uber 1080 killer when it's clearly not aimed at that market segment.
>>55259068
power, temperature, or bios limitation.
>muh voltage limitation
>muh BIOS restrictions
Nothing but excuses
If you had two brain cells to rub together, you'd understand why it's unlikely for 95% of RX 480s to break 1400MHz after reading this slide
>>55259132
Is that crossfire or single GPU?
Either way nvidia is finished
>>55259376
Single gpu. Nvidia just can't compete anymore
>>55259390
Just wait for Vega, there will be literally no point in nvidia's continued existence then
What ever happened to WAIT FOR THE BECHMARKS?
>>55259443
Benchmarks?
Who cares, I want to see performance.
>>55259463
>I want to see performance.
Better buy a 1080 then
>>>pcpartpicker 1070 with a baseclock of 1670mhz
I have predicted this even before the new cards were announced.
AMD a shit
>>55257773
Typical AMDshit blabber when he realizes the thing he's been holding out for for years is literally a lukewarm week old turd
>>55259410
>Just wait for Vega
Just wait for Polaris, they said. They also said the same about Fury. Looks like the goalpost has been moved once more. What next? Wait for Navi because Vega won't deliver either when they finally release it next year.
There are no brakes in the wait train.
>>55257849
Leaks that show the card is not capable hitting 1400? Sure, that makes sense.
>>55256914
>AMD retards still think % of clock increase = % gained as performance.
>>55262109
>he can't recognize obvious shitposting
'just wait' -- t.amdfag
'I want AMD to fail' -- t.stupidgoym
'buy the founder's edition: its worth it' --retarded aborted fetus with a pipe through its frontal lobe, AKA, the smartest Nvidiot alive
>>55258780
Those cards are expected to be 1500, not 1600... some asian got it to 1600 on a stream and we are wetting out panties thinking stock will all do this, i'm personally assuming everything above 1400 is a golden chip till shown otherwise.
stupid nvidiots
rx480 will reign supreme
OCd it will be on par with a 1070.
I want some game benches instead of firestrike and shit.
>>55264233
just look at some 980 benches and imagine them a little bit worse
>>55264277
That's not how it works.
>>55256721
>24C
I wish I knew summer temperatures like that. It's 37c outside right now where I live, and it's midnight.
>>55264310
Nice equator, faggot.
>>55257773
Clock to clock Pascal is slower than Maxwell, kind of retarded top compare since they are different architectures
Still, performance due to higher clocks will always scare anyone due to Netburst and Bulldozer
>>55258773
>I don't think anybody really expected the reference RX 480 to beat the 1070 regardless of how much it could be overclocked.
The technically illiterate saw that the 480 allegedly runs very cool and draws little power, and assumed that meant there is enough headroom for overclocking to get close to 1070 performance. There are a lot of people that post things like "why pay twice the price for 10% more performance" in regards to the rx 480 vs gtx 1070.
>>55257755
To avoid retard reviewers claiming a gorrillion TDP since they were able to clock the cards really high
>>55264333
But I'm pretty far from the equator. I'm at 33.4484° N, 112.0740° W.
>>55259207
I think a lot of poorfags will buy reference with the rx 480/470s. AMD used a really small heatsink to get the price down, so the AIB cards with decent cooling will probably cost more relative to the reference cards than in previous generations where the reference cards had more expensive heatsinks.
Here's what I think will happen a lot.
>4gb rx 480 with a decent cooler costs $230 or more
>8gb reference rx 480 costs $230
>poorfag thinks 8gb means better performance on their 900p monitor and buys the reference 8gb card
or
>reference 4gb 480 doesn't cost much more or even cost more at all compared to an rx 470 with a decent cooler
>poorfag buys the reference 480
>>55264477
reference 8GB 480 is $240 actually
AIB will probably add another $50 for cooler, custom PCB, phases, etc
a high end 480 will be close to $300
>>55264846
>a high end 480 will be close to $300
That price creep... Makes me wonder how 2x reference 4GB RX 470 at 150 USD each would compare to a top end 300USD 8GB RX 480 for 280 USD.
I still don't think most aftermarket overclocked 480's will be over $250-260, maybe those super special limited editions ASUS shits out.
According to videocardz, which Nvidiafags insist is a reliable source, overclocking the reference cards nets you a 14-15% performance increase.
Which seems pretty nice, aftermarkets should be hitting over 20% if indeed the 6pin is a limiting factor.
>>55265153
It's only a reliable source when it doesn't favor AMD.
>>55265118
8GB reference is $240 already, no way youre getting a custom cooler and all of the other shit for $10-20.
I have a GTX 670 currently, and will more than likely just go with the RX 480 because I still have a 1920 x 1080 60hz monitor. Should I wait for non-reference cards, or is the difference marginal?
>>55265370
i would wait, the leaks show the 480's reference cooler being trash
even if aftermarket doesnt OC much further they will be quieter and cooler
>>55265382
I'm not in a rush, so I might as well wait. Is there a general timeframe for the non-reference releases?
noVidya sure is in all defence today, scared of tomorrow?
>>55265430
all of the people planning on buying a 1060 sure are flabbergasted!
>>55264846
>a high end 480 will be close to $300
Well, from what we've seen from the leaks/speculation about the 1060/ti, it'll be around $230-$280 and have almost 980 performance on the reference cards, and probably exactly 980 performance from AIB cards. http://wccftech.com/nvidia-geforce-gtx-1060-confirmed/
My speculation is that because there was a large performance and price gap between the gtx 960 and 970 that gave the $200-$300 bracket to AMD, Nvidia does not want to repeat that and will release a 1060 closer to $200 and a 1060 ti closer to $300, or the 1050 will be around $200, the 1060 will be around $240-$260, and the 1060ti will be around $310-$320. In the first scenario, the 1060 would be slightly slower than a 480, but slightly cheaper as well because Nvidia is trying to undercut AMD (probably why they haven't announced anything below the 1070 yet, they're waiting for the NDA to be lifted on the 480 to finalize pricing), and the 1060 ti would cost a bit more than an rx 480 8gb with good cooling, but would have a performance edge over it as well, or it would be equivalent to it. In the second scenario, the 1050 would be in a weird gap between the rx 470 and 480 4gb, the 1060 would be on par with rx 480s with good aftermarket coolers in terms of price and performance, and the 1060 ti would be around Fury X to 980 ti performance with better price and power consumption.
So much shilling,
THIS IS THE REFERENCE CARD YOU FUCKS
>>55265459
If Nvidia releases a better than typical __60 this cycle at a price that rivals 480 it would put AMD in a tight spot until Vega which would need be ready to compete with 1080ti
either way competition is good for us all, the 480 looks like it will be an awesome card for people wanting to upgrade from something slower than a 980 or who are building a new PC, if Nvidia also makes an awesome cheap card i dont see why thats not a win win for people who are actually buying video cards
>>55265459
>nvidia having wafers for another GP104 bin
Hilarious
>>55265459
How will the 1060 ti be past 980 ti performance when the 1070 is just ahead of the stock 980 ti?
>>55265562
I thought the meme rumor was that Nvidia has a low yield rate, explaining the high price and low output of 1080s/1070s (personally, I think the 1070 having more stock means that the low availability of the 1080 is caused by the low supply of GDDR5X and the high price is because Nvidia won't have any competition in this bracket for a while). So wouldn't that mean that they have a lot of chips that they can cut down for the 1060 ti?
>>55265615
There's a 10% difference in performance between a 980 ti and a 1070, and I used the 980 ti as the high end. But yeah, that does seem a little high on second thought. It'll probably be 20-25% slower than the 1070, because the 1070 is 20% slower than a 1080.
>>55264382
>the technically illiterate
Cool story bro, you can't even understand the grammatical difference between "technologically" and "technically" and I'm supposed to value your opinion on other people's knowledge?
>>55265640
There's no way in hell any company would sell fully functioning chips that are already in short supply so they can cut them down and sell them for a lower end card.
>>55256721
The review models sure, maybe we'll see more overclock with third party manufacturers and better power smoothing etc?
eh maybe. still good card for price.
>>55265660
Who pissed in your cheerios?
>>55265615
>>55265640
An overclocked 1060ti could probably reach stock 980ti speeds,
stock clocks would be just slightly above the 980 but below the 980Ti, an OC 980 can't reach stock 980Ti speeds so the 1060ti could fill that gap replacing the 980, which leaves another gap for a cheaper 1060 card to fill in between the 970 and the 980.
https://www.youtube.com/watch?v=8qVLapLW2es
Now that actual gameplay benchmarks are starting to come out it has become pretty clear that Nvidia has everything figured out and they are competing with themselves trying to not cannibalize their own sales by releasing GPUs that are too strong or too weak.
>>55265712
Someone trying to flaunt superiority but failing to grasp a basic concept such as spelling and grammar relations?
>>55265736
But I'm not trying to flaunt superiority? My only comment in that reply chain is >>55264382
I was literally just trying to explain how that misconception came to be. Sorry I offended you so much.
When did anyone claim it made 1070 performance?
It's a dieshrunk 390X for 200 fucking dollars.
Nvidia has no cards that compete with this.
>>55265832
you said, again,
>the technically illiterate
The technically illiterate can't even read. This is by and far the wrong place for the technically illiterate.
>>55265875
>dieshrunk 390X
Are you retarded? Can you stop using words you don't understand?
>>55265915
I'm incredibly sorry I used the wrong word. Will you ever forgive me? Please.
>>55266021
Just admit you're a /v/tard idiot with as little knowledge of what you're saying as anyone else here.
>>55266083
I'm sorry, I'm genuinely confused as to why you're so hostile to me. I just explained, I've seen a lot of comments from people expecting the rx 480 to be able to overclock quite a lot because of the low temperatures and power draw on the reference card. I didn't even really say much about what I thought of the card. I was just dealing with how that misconception came to be. Can you tell me why you're so offended by that?
>>55258880
Bullshit. AMD does not adhere to pin spec, havent in generations
See: 295x2 and others
>>55266119
>>55267168
>>55256721
Oh no!
So a $200 card with OC won't be as good as a $400 card?
>>55264420
Hope you like pizza
>>55267188
i hope i can see a RX480 Lightning
>>55256721
Regards,
Dolan
>>55256872
>this is what Team Green faggots actually believe
>Shills: The thread
Stock RX 480 beats or equals the Msi 390x gaming at 1100 mhz which is about as far at the 390x will go. 12000+ firestrike vs 11600.
Gtx 980 will do 14000 firestrike at average oc of 1440 mhz.
Stock RX 480 OC 10% 1380 mhz puts it at supposed 13700 firestrike.
Ref vs custom pcb 1070 10 power phases, 8+6 pins gains +150 mhz of OC, 1785 mhz average to 1950 average core clock. An additional 10% increase in clock with a subsequent 6.5% FS score increase.
Assuming the same gains on Polaris puts it at 1500ish mhz, and a theoretical 14600ish firestrike score beating the OC gtx 980.
>>55258011
>>55258042
I think it is sad that everybody seems to care about the trolls
>>55265729
so much fantasy
nvidiots are delusional
>>55267419
I can't follow - does it mean I should try to get a used 980 for a similiar price?
>>55267594
>used 980
>200$
only in lala land
>>55267607
I take this as a yes then
>>55267635
you take it up the ass faggo
>>55267594
A 980 becomes better than a rx 480 at about $260.
>>55257967
>2016
>Being this delusional
Getting really tired of these threads.
Polaris is supposed to carry the next iteration of GCN stream processors targeting a mainstream audience at a range of $100 to $300.
It was never meant to compete in the higher performance range.
The only real, and I type it loosely, benchmark we have of any Polaris card is the RX480 with Raja at Computex. Doing some numbers here and there, we can expect it to compete or maybe even surpass the 390 and the 970 for $200.
Recent events at E3 show the RX470 and the RX460. The former being cut down Polaris 10 judging from the PCB and the RX460 being Polaris 11. In other words, the RX480 is the highest performing card in the Polaris lineup.
Since they mentioned a range of $100-$300, it puts the RX460 at a $100 and the RX470 at about $150. The latter one is very interesting. If you remember, the 7850 was a pretty cute overclocker and it was cut Pitcarn(Pitcarn PRO). If the RX470 follows, then we might have a nice card that can match the RX480 with a little OC and $50 less, making it a killer at its price.
The RX480 on the other hand is definitely within the last generation of lower high end cards, possibly surpassing them. Benchmarks point to above 390x performance, but I'm personally skeptical about it. It will be another amazing price to performance card.
If you're running anything on or above the 970 or the 390, then I would be waiting for something like Vega or fatter Pascal. The truth is that while the 1070 is an amazing card for its price, it doesn't deliver all of the high frames at 1440p or 60fps on 4K. This lineup of Pascal is very disappointing and it feels like nothing new. Look at the true Pascal line, they're monstrous chips with HBM and all that good shit. We are stuck with these chips that aren't really anything new except a little better than the 980Ti.
Also, reminder that you should WAIT FOR BENCHMARKS. 3 days left, fuck.
>>55268222
But the wait is unbearable!
simple question; I want to be able to play games @1080120 fps min fps. Which card should I buy?
>>55268451
2x 1080 unless you want sub-120FPS drops
>Nvidiots think this is about people upgrading from last years cards so shit the place up about 'muh $200 card for last years performance" when in fact it's about making a $200 card for people wanting to upgrade from 3+yo cards or new game players.
Seriously. You make yourselves look foolish with this nonsense. We all know it is not aimed at 1070 territory. We all know it's not going to reach anywhere near 1070 levels without a AIB and a price hike.
The claims of overclocking to ridiculous levels were put out by trolls and Nvidia shills. Also DX12 although it will help with AMD performance is something we must look forward to. Not look at in the here and now since there is hardly anything using it yet. But again Nvidia bozo's will use it as an excuse for shit flinging. Nobody can predict the future but one thing is for sure. DX12 will eventually become the standard, Right now though DX11 is what we have to deal with in terms of performance analysis and ANYONE saying this card is going to compete with 'enthusiast' Nvidia cards are morons and trolls.
AMD are not lying about the RX480's performance to dollar ratio. The specs are what we expected for a reference card. Expecting crazy clocks for low power and temps was a pipedream.
I swear Nvidia twats have been posting half the AMD hype bullshit just to cause a shit storm.
>>55268608
The tl;dr is that it's a badass midrange card and nvidia fanboys think it's a high end card
I guess they would after being the loos for nvidia's shit for years.
>>55268634
The tl;dr is that it will trade blows with the 970. At $200 that's not too bad, but not what AMD needs to secure a bigger market share.
>>55268608
>it's about making a $200 card for people wanting to upgrade from 3+yo cards
I'm still using a hd 6970. Which is 6 years old by now?
Worth an upgrade?
>>55268634
It's a cuck card, plain and simple.
If you can't afford a 1080 you need to seriously consider whether PC gaming is for you. I would buy a PS4 or an Xbox One if you're considering a cucked out RX480.
Come back to PC gaming when you can afford it.
>>55269107
Probably
Mid range cards might sell more than high end, but have much smaller profits. It's development of a new architecture that costs, not the physical silicon itself. PCBs, coolers and other shit are also the same for both 200$ and 800$ cards.
Wouldn't be surprised if Nvidia still made more money even if the 480 sold more than the 1080 and 1070 combined.
>>55269206
>don't buy a midrange amd gpu with 2306 SP
>buy a console powered by a shitty amd gpu with 768 SP
hmm
When people were posting "leaks" hyping up amd
>hurr wait for official benchmarks
When people are posting "leaks" bringing down amd
>screw waiting for benchmarks, this must be true
>>55270019
Given that nvidia spent ten times as much on r&d they'll need ten times the profit
canada fag here
someone in my city is selling a 980 classified for 400. i'm torn between waiting for a 480 or just throwing down. i might be able to bring it down to 375 who knows
>>55257768
>>max wattage is 150 with a 6 pin
max SUSTAINED, sure, but peaks can easily be 170 to 180W
>>55270166
Poor little cucky with his cucky card.
Getting a 1070 or a 480? Might as well put your cock in a cage and let the bull fuck your wife. Same difference.
>>55271724
>>55269206
how long per day do these shitposters spend on this? do they actually enjoy it?
yesterday was shitty leaks for the 480
today is good leaks
http://semiaccurate.com/forums/showthread.php?p=267069#post267069
>>55271914
Note that they get mad when consoles are mentioned. They are basically console users. Specifically ps4 users
>>55256721
we know that already..
amd has in plan 2 versions of the card
one with only 1 6 pin connector
and one with 1 6+ 1 8 connectors
you can already see on some gpu z leaks that the card on 1380 is drawing 149 watts pretty much is being holded up by the power it needs
>>55256914
gonna ask for proof but then i remeber no one has leaked any review yet...
If It would get 1070 performance for that price it would singlehandedly end Nvidia as a company we know today.
But again for this price, I'm still getting it.
>>55272051
Spoken like a true cuck.
Well the bull only gets to fuck my wife on the weekends, I'm sold.
Go back to console gaming kid, or ask your parents for more allowance money.
>>55272092
The fuck does that even mean? Fuck off /v/edditor.
>>55272164
Is wittle cucky wucky conofoozed???
Poor wittle cucky wucky with his wittle cucky card.
NOW GO PREP THE BULL
>>55272199
>delusional nvidiot sees 70% increase on perf on dx11
>never talks about dx12
>the card is 8% faster than a mere fury x
>for 200 bucks more
>comes on 4chan to shill
>thinks he is superior
>oh the irony
>>55258052
He's right you know.
Okay kids, settle down. The adults are talking here.
I want a single card for 1440p/at least 100hz, is the 1080 enough?
>>55272453
>at least 100hz
Actually yes. Vega and 1080Ti will do 120fps however.
AMD can't even put out a competitive offering as per usual. Guess I'll just get a FE 1080 before the excessive demand drives the price up even further.
>>55272532
>vega
Vega will be lucky to get to 1070 levels. Total cuck card.
>>55272532
Is 1080 the minimum, or could a 1070 or something like that handle it?
>>55272594
1080 is the absolute MINIMUM for real pc gaming.
