![rotr-overall[1].jpg](https://i.imgur.com/3Cpd8ZC.jpg "rotr-overall[1].jpg")
Nvidia shills on suicide watch
http://www.legitreviews.com/amd-radeon-rx-480-undervolting-performance_183699
![division-overall[1].jpg](https://i.imgur.com/HeC7PGa.jpg "division-overall[1].jpg")
![bf4-overall[1].jpg](https://i.imgur.com/D1id4L2.jpg "bf4-overall[1].jpg")
>It appears that AMD was pretty easy going with regards to the higher P-state voltages as we were able to reduce ours from 1.15V to 1.05V with the new AMD WattMan software and still had full gaming stability. By lowering the power level down it increases the efficiency of the GPU and for gamers that means you get to hold the boost clock speed longer and get better performance and use less power.
NO DRIVERS
>>55414985
I'll buy one and /g/ cannot stop me!
>>55415019
this card keeps getting more interesting for me
![[Laughter Intensifies].jpg](https://i.imgur.com/8WnjBIQ.jpg "[Laughter Intensifies].jpg")
>>55415073
>GTX 1060 has a 192 bit memory bus
so being the inept company amd is, they released an unstable overvolted gpu that needs to be manually undervolted.
>>55415174
its going to be undervolted to the same levels as this article tomorrow morning with a software update
>>55415205
More to it than that. The memory is getting tweaked as well, over all the total board's power consumption will drop, and variability in power draw will be markedly lowered in cases where GPUs were effected.
Not all of them had this issue with power draw to begin with, it looks like it was tied to the ASIC quality rating, and Wattman's handling of adjustable TDP between the die and memory.
Where are the nvidia shills?
>>55415514
Crying.
While it is nice to see AMD fixing the issue, it should have never been one. This type of problem should have been seen/foreseen by the development team. Or the people who test the GPU. And if it was tested, why was it pushed out the door? Did they honestly think that NO ONE would find out? With people doing extremely intricate benches/reviews on the card?
I sincerely hope that them "undervolting" the PCIE power draw fixes the card. I hope that the 3rd party coolers and designs include the 8-pin/ 8+6 pins setups. This card hitting 1500mhz+ without blowing something up would be nice to see.
>>55415129
>192 is more than 64
>DO THE MATH
>>55415618
The problem came from the Wattman software and how it interacted with the GPU through the drivers.
Software bug, not a hardware issue.
Nothing is going to make the cards clock any higher though, they're already OC'd. Hitting 1400mhz or even slightly beyond is pure silicon lottery.
>>55415514
Monitoring this thread since there's nothing to say. You've gone from, "1600MHZ GTX 1080 KILLER," to, "If you undervolt it it can run at its stock frequency a little longer before throttling."
>>55415618
>And if it was tested, why was it pushed out the door? Did they honestly think that NO ONE would find out?
This. This is the most interesting question here.
>>55415656
Didn't affect 100% of GPUs
Internal testing and validation of dies, and completed cards didn't include the use of their new overclocking software
It wasn't that they expected this issue to magically fly under the radar. Its just an unforseen software bug. Its not even directly related to the default voltage for each pstate.
The voltages are higher than necessary because AMD did no real clock binning for reference consumer cards, but that isn't whats causing the aberrant power draw from the slot.
test
>rx 480 can't hit 1500mhz even with water cooling and 1.3v
I'm fucking done. I'm buying a 1060.
>>55415743
https://www.youtube.com/watch?v=Jq47qmwcus8
>>55414985
Oh hey AMD. Thanks for the extra 4GB of vram. You're so geneous.
>>55415743
>waaaah this 8 stage pipeline arch can't clock to the moon!
GCN has never been, and never will be an exceptionally high clocking arch. Some of the best overclocking Southern Islands cards only hit around 1450mhz, and 4th gen GCN has the same pipeline.
The RX 480 is a $200~ mid range GPU, not an enthusiast part.
>as a based consumer tech bro allow me to explain what the electrical engineers at AMD should have done...
>>55415618
least its fixable, unlike 3.5gb vram
people forget too fast
>>55415733
Makes sense. Thanks.
>>55415814
No telling how many cards can actually run stable with a full 0.1V undervolt. Worse than PCI-E being out of spec would be, "unstable at stock voltage."
>>55415129
width x rate is what matters, but yes 192 is unfortunate, forces us to get 3gb cards or 6 gb cards.
Post all and every meme regarding gpu's from the last decade.
I'll start with woodscrews and 1.7 yields, also "no drivers"
>>55415997
The card internally adjusts ultimate voltages because its capable of monitoring the quality of its own transistors through AVFS sensing paths. If a card needs a variable 5%~ more voltage at a given time then it will sense this and adjust accordingly.
>>55415646
After like 5 minutes of googling, Wattman software seems to be part of the AMD Radeon pro software suite yea?So how was AMD software having problems talking to AMD made drivers for an AMD card?
>>55415656
Exactly. Software or hardware is irrelevant. Something like this should have never been in question in regards to basic quality control. I mean I get it. AMD feels the pressure to release a competing product.But for fuck's sake didn't they say in the past they "want to stop being seen as the cheaper option?" How the fuck do you expect to break that public opinion if you can't get bare basic stuff like this down?
>>55415878
I never forgot about the 3.5/4GB issue. That was a slap in the face and the fact that there was no punishment to speak of (to this day) is mind boggling. How are they still getting away with putting "4GB GDDR5" on their boxes? I was afraid that would become the new trend for both major GPU makers."16GB GDDR5!!." Only 10 usable.
>>55414985
how come ?
>>55416025
Housefire of course.
>>55414985
>RX 480 launches a few months before zen
>AMD desperate for sales decides to make cheap gpu
>Implement plan with said gpu to kill mobos and force users to upgrade their base platform
>Poor fags decide to wait for zen to buy new mobo/ cpu
>AMD laughs all the way to bankruptcy
>>55415129
GTX 1060 has more bandwidth per core than the 1070 or 1080. It'll be fine bro.
>>55415129
>only 33% lower than 480
nvidia's DCC is more than 33% better, so effectively it will have more bandwidth
what's your point?
>some german guy managed to push his to 1500 mhz
I'm pretty optimistic about the nitro, the chip can handle it with the right power
>>55416189
I'm poor
>>55416207
sounds about right
>>55416205
1,3V
WATERCOOLED
DIDN'T EVEN HIT 1500MHZ
Stop being delusional. I also wanted to believe, but AIB cards will OC to 1400 at max.
>>55416234
Also remember, watercooling is expensive and kills the point of this card.
I'm still excited for the AIB cards tho.
>>55416205
> 260 watts for an 18% overclock
my sides
>>55414985
>just discovering this after release
wow, sure they are experts on the subject. you have been pajeet'd.
>>55414985
well i love getting shills into the wall
but there isnt a card out there that doesnt gain from undervolting
>>55416044
>That was a slap in the face and the fact that there was no punishment to speak of (to this day) is mind boggling.
You know what another slap in the face? The fact that so called "3.5GB card" beats what AMD has to offer with 8GB...
The only "advantage" is that some games at unplayable settings are slightly less unplayable.
>>55416234
gpuz says 1500 mhz
didnt even hit 1500
>>55416299
Turning up textures has zero impact on performance when you have the VRAM to handle them. You could still play something like Rise of the Tomb Raider on the same playable settings but with the highest texture setting if you had more VRAM. The idea that having a pitiful amount of VRAM doesn't matter because the game would be unplayable at higher settings is retarded. You don't have to turn any other settings up, just textures. The game then looks better with zero impact on performance. It's not rocket science.
>>55416299
If you ignore DX12
>>55416309
https://www.youtube.com/watch?v=Jq47qmwcus8
Watch the video, faggot.
>>55415090
/thread
>>55416299
2 games is hardly "beats" the 970 is ahead only on fallout and witcher
but on total average 480 is smoking that half breed of a gpu
>>55416376
Winning out on average with only half the ROPs of Hawaii. It would have been a total monster if it were a lower clocked 64ROP die.
>>55414985
You know what I do to my 970 to make it perform? Plug it in.
>>55416430
And then spend a few hours carefully adjusting settings to make sure you don't trip the 3.5GB VRAM limit and tank your performance. :^)
>>55416037
>AMD cards are now aware of their poo in the loo-tier build quality
>legitreviews.com
Sounds like a trustworthy website OP.
>>55414985
>AMD shills on /g/ happy that undervolting a card increases performance in a single test
AMD is so retarded that giving a card more power reduces performance.
You also understand that not every card is going to be stable at lowered volts right?
Then AMD will have to lower base clocks or boost clocks and that basically becomes false advertisement because of poor testing before sale.
This also has the potential to fuck OCing in the ass as well right?
like I'm fine with shills existing, but can you not be retarded at the same time?
Jesus, how do people consider /g/ to be a legit tech board.
>>55416530
I do love a good rambling post from a complete tech-illiterate.
>>55416582
feel free to retort me instead of adding nothing of relevance faggot
>>55414985
great
now instead of your system crashing because of over current, it crashes because the gpu is now overclocked and has even less voltage than stock.
>>55416623
How tech illiterate are you?
>>55416582
he's right for the most part though
So should I just keep my reference RX 480 that comes in a couple days? Or should I just return it and wait for the aib variants? I'm honestly fine with either, I just want something recent that isn't my 7950.
>>55416650
how tech literate are you?
when you verclock you need more voltage than normal (usually a lot more) so when you take voltage away then overclock on top of that you make your system more likely to crash from lack of power.
>>55416689
aibs are generally better, reference coolers are meh.
>>55416689
You already bought it so might as well stick with the reference 480, but a general rule of thumb is to wait for the custom cards, so you fucked up
>>55416707
My thing is I don't care much about the noise, for thermals later down the line I could just get an aftermarket cooler that I could reuse.
What do you think? I'm honestly 50/50 right now.
>>55416719
Well if I refuse the delivery I get a flat refund that's why I'm so wishy washy right now.
>>55416736
*Full refund
I've never bought a nvidia card. Not because I'm a shill but because the amd card always seemed to be a better value for the price. No longer. Nvidia cards command a price premium for a reason now.
>>55416608
I don't argue with retards.
>>>/v/
>>55416689
>buying reference cards
Wait for the Sapphire Nitro.
>>55416703
you always want to have VCC as low as possible, no matter what you do.
if you OC if you DC, you want to have your VCC is low as it goes, because you wont have as much Watts and thus will create less heat.
This is pretty much basics of EE.
>>55416530
>AMD is so retarded that giving a card more power reduces performance.
There is nothing "retarded" about that. Lowering heat output and power lets any die run in peak boost state for longer.
AMD did no clock binning with consumer Polaris10 dies, voltages are more than generous just so they could move volume.
>You also understand that not every card is going to be stable at lowered volts right?
Incorrect as AVFS accounts for this. You're utterly clueless.
>Then AMD will have to lower base clocks or boost clocks and that basically becomes false advertisement because of poor testing before sale.
No they won't. All of your speculation is using a laughably flawed premise.
>This also has the potential to fuck OCing in the ass as well right?
Lowering voltages at stock clocks has literally nothing to do with OC headroom at all. Jesus Christ, you are legitimately stupid.
>like I'm fine with shills existing, but can you not be retarded at the same time?
The irony here is palpable.
>>55416684
Not even remotely.
>>55416703
only if you use amd. on my 980 and all of my friends 970's we can get to 1400/1500 mhz without any extra voltage. on my 980 i don't even have to adjust the power limit slider to get 1450 mhz.
>AMD shills says that their driver issues have been fixed with the latest update
>the whole over voltage thing is because of AMD's incompetence with software
Really makes you think.
>>55416725
if you really don't care, then keep reference.
I rock GTX780 reference and I'm fine.
>>55416752
I am forced to be agreeing with you fellow unbiased consumer. The AMD are just a poor cards only bought by poor people. I and all my friends have buying superior Nvidia GeForce GTX graphics from this time on.
hi guys
>>55416782
I'm so very conflicted right now.
>>55416763
well no shit.
but if you put it too low it's going to crash.
>>55415654
>"If you undervolt it it can run at its stock frequency a little longer before throttling."
If you've actually read the thread then you'd know that undervolting it means it can holds its BOOST clocks better, not its base clocks.
>>55416755
So you don't have anything relevant to say, gotcha
I mean, you could have just admitted defeat instead of cowering like a bitch
>>55416765
>Look mum, I can use big words to make me seem smart
Do you work for wccftech?
they could use some brain dead retards like you
>>55416817
>get blown the fuck out for being a tech illiterate retard
>desperately shitposts in retaliation as he realizes all of his idiotic speculation looks childish
Every single time.
>>55416234
>Stop being delusional. I also wanted to believe, but AIB cards will OC to 1400 at max.
HardOCP, the biggest NVidia shill anti-AMD site on the internet, has reported that AIBs are saying that the cards hit 1600MHz but it depends on silicon lottery.
>>55416765
he's right about the voltage. it can cause instability with the lower voltage. even adoredtv admitted it.
>>55416811
i.e. 1266MHz.
>>55416879
Hes not right about anything, you can't make a blanket vague statement about a subject that is explicitly granular.
Lower voltages for any given pstate, and how it affects stability, if at all, depends on the given ASIC, and how far they're being lowered. Reducing a boost state voltage 10% when it has a 15% stability margin, and the chip itself regulates per transistor voltage in real time, would never adversely impact stability.
>>55416940
according to some sites the asic quality of a lot of the cards are below par and hence why they had to increase the voltage to get as many units in stores as possible. it will affect some cards
>>55414985
>>55414994
>>55415019
HOOOOOOOLYSHIIIIIIT
FUCKING 0.5 to 2.5 FRAMES PER SECONDS IMPROVEMENT
and
AND
IT DOESNT MELT YOUR MOTHERBOARD NOW!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
FUCKING 1070 BTFO, 970 BTFO, 390 BTFO, NVIDIA BANKRUPT BY TOMMORoW
>>55415073
>holidays
>decide to work for a month on a company to build "gaming computers"
>I have to install windows 10, and drivers of the mobo, gpu etc. so they can work out of the box
>Just finished to build a computer with a 1080
>put the CD provided by nvidia to install the drivers
>it install nvidia geforce experience bullshit
>no drivers
We don't connect those computers to the internet, it's too much troubles, it's faster to just install the drivers using the CDs provided by the manufacturers you pay 700€ for a GPU and they can't even ship it with its drivers.
>>55414985
This shit is just getting weirder and weirder
>>55417091
750ti and 960 were far worse on the PCIe problems and nobody complained that their motherboards were fried from those cards, what makes the 480 so special
>>55417230
this is bullshit. i worked in a company which built pc's too and all windows verification and drivers will be done on the internet because it was way faster.
everyone disregard this flaseflag post.
>>55416789
hello best gpu software ever made. you are lovely. i love you.
>>55417299
Lookz nice
>>55417246
Not the same thing at all.
https://www.pcper.com/reviews/Graphics-Cards/Power-Consumption-Concerns-Radeon-RX-480/Evaluating-ASUS-GTX-960-Strix
Plus even if it was the same thing happening, two wrongs don't make a right.
>>55417246
no, no they weren't they pulled over spec for milliseconds at a time. never a solid minute like the 480. stop shilling and get your facts straight.
>>55414985
>>55414994
>>55415019
DELETE THIS
>>55417463
Oh right these constant peaks are far better to AMDs problem that can be fixed via software
>>55417724
You're an idiot. Look at the lows, it's not "constant peaks". Average is perfectly acceptable every single second, unlike the 480.
>>55416689
never buy reference cards
never ever
there's literally not a single reason
I'm getting really really tired of saying this
>>55417246
You mean a 960 produced by only one manufacturer. Anybody who didn't buy an Asus Strix apparently didn't have problems.
>wattage
>>55418120
Just like how not every 480 goes beyond PCIe standards when pushed to the limit?
>>55416044
Because all 4GB of it is usable? Yea your performance goes to shit, but it's not any more shit than having just 3GB of memory.
I mean everyone is going to meme about not having 4GB of VRAM so I guess it's a trade between being memed for 3GB of VRAM or "3.5". Which is better? I don't know.
>>55417237
It's not weird at all. If you undervolt it, you lower its power consumption, and thus it can stay at higher clocks for longer.
The only weird thing is why AMD didn't set it to 1.05 V to begin with, but that could very well be the silicon lottery.
>>55417724
That's correct, since those peaks are expected. It's just how transistors work. Even with the "software fix" which under volts the card, RX 480 will still have those peaks. They are going nowhere. Might become smaller though.
>>55418189
Its pretty simple.
ASIC quality of chips seen by reviewers and by consumers now have ranged from 70% to mid 80s% I'm willing to bet Apple got all 90% and up, Voltages are generous enough that shitloads of chips passed validation so AMD could have a huge volume of consumer cards at launch.
The voltages of all the pstates wasn't the cause of the power draw over the slot though. Separate issues, though they have a lot in common.
>>55414985
>Return
wat?
>>55415770
Yeah too much work for the average overclocker.
So pretty much undervolting a GPU running on stock clocks made enough space to unlock its boost clock - which made no difference in performance even with 150mV drop. I'm happy for ya AMD owners but I still won't buy it as I want at least OCed 980 performance for my 2560x1080 monitor. I still need to ask two questions:
- How exactly such an easy to repair fuck up could slip through
- How exactly are they going to fix it via drivers update if it probably means changing voltage in bios via something like flashed usb.
I'm cancelling my Nitro 480 preorder and going to get a MSI 1070 Gaming X when I can grab one cheaply enough. May as well push the boat out a bit. At least I will be able to play every game at 1080 fully maxxed out @60+FPS with headroom for a 1440p monitor should I decide to get one later on.
>>55416025
3.5 gb
>>55417230
How hard to is insert a cable or a dongle for wifi? Sounds like you work in a retad gaming computer building shop.
>>55416037
I kind of wonder if that's the problem with it drawing too much power, if the calibration for that is out of whack and overvolting like mad
Polaris had good start then a bad few days but now it's back to #1
>>55419543
I think that's the main reason why Nvidia rushed 1060. People don't trust 970 since it's outdated and has only 3.5gb vram so AMD has no concurrency in mid price range.
>>55416037
What does AVFS stand for?
>>55419628
Adaptive Voltage Frequency Scaling
>>55419646
Thank you
>>55417246
>far worse
The problem with the PCI-E slot was a thermal issue caused by over the spec current. Learn what overcurrenting is. It's basically plugging too many toasters into a wall socket. It will cause the wiring to burn because it's pulling over the rated current for too long.
>>55419413
AMD obviously didn't take into account that tech review sites like PCPer and Tom's Hardware had riser cards to measure PCI-E slot current and voltage.
>>55416207
lol
>>55417724
>constant peaks
The peaks don't hurt the mobo, the heat does you stupid fuckface
As long as the average is within specifications it can go up and down like a roller coaster and it wouldn't mean much. The problem occurs when you go over the rated current and it creates too much heat that the slot can't get rid of and you start damaging the solder and contacts.
>>55419574
>People don't trust the 970
>Everyone and their dad has a 970 on steam
Why do AMDerps create their own reality in their head?
>>55416205
I think when the rumors came out about this chip hitting 1600 on air. It might have been the AIB(AfterMarket) cards.
Pic; RX-480/w EK Water Block
>>55419726
>As long as the average is within specifications it can go up and down like a roller coaster and it wouldn't mean much
Have you heard of electromigration?
>>55419738
>buy $250 card
>slap a $120 waterblock onto it
>1070 cost $379
>>55419735
But anon I own 970 myself and I never experienced the 3.5 vram meme. Still I can tell that 480x is a bestseller on every major UK website.
>>55419493
They made a quadro too, so 7GB
>>55419574
>3.5gb
Except it has 4GB and can use all 4GB and no one cares except for brain damaged AMD shills because they're desperate for something to latch on to.
>>55419760
You mean $450, there is no 1070 selling for MSRP
>>55420036
http://www.newegg.com/Product/Product.aspx?Item=N82E16814125875
Hi.
>>55414985
WTF is going on at AMD that they shipped a card with questionable stability at stock power and questionably BETTER performance with slightly less than stock power?
A lot of consumers when they get a budget card they try to overclock it to get the best performance without loss of stability. With the 480 they have to underclock it to ensure stability.
Wait, I don't understand, AMD fans said this was all a lie and stuff, after they told me i should wait for it to release based on expectations that didn't quite came to reality and stuff...
>>55420049
Pretty sure that $400 is higher than MSRP.
>>55420144
Pretty sure that's less than $400 unless you're retarded enough to confuse 3 with 4, plus $389 was only for ref cards, that's a non ref.
>>55415618
If you've ever been curious about why R&D budgets are sky high and things take a long time (for example 7970 respin/revisions took months, and the final XT/XTL revision was more than nine months)...
AMD has had a bear clawing at their tent for some years now and they are in the middle of taking multiple risky paths to build any footholds they possibly can. Someone decided that skipping QA here and there was ultimately the more effective path rather than being later to market and/or having considerably less volume to sell. And regardless of this current power issue I assure you that person (rather the people responsible) will be getting nice bonuses every fiscal quarter until rx480/70/60 sales are saturated.
>>55420049
Btfo by another anon before i even got a chance to reply
>>55419757
it is an inevitability with all hardware.
>>55420152
Oh wow it's a whole penny less! You fanfags are getting dumber by the day.
>>55420187
You're being plain retarded, you've been btfo twice now, just stop posting any time.
>Also implying anyone in EU will even buy 480 when its shitty ref version costs 280EUR if you want something decent like Sapphire or MSI.
>>55420171
Large repeated swings in flux will do considerably more damage than a constant or minimally-varying current will. We're talking on the order of bringing a component with a nominal 10 year lifespan down to 2 years. Additionally, this is only looking at electromigration: as ripple current variation increases, electrolytic capacitors on the motherboard will slowly degrade, resulting in decreased filtering. There's a whole host of reasons you don't want to constantly vary the load.
I mean, maybe you're planning to replace your motherboard in two years anyway, but to act as though no meaningful damage is being done is simply incorrect.
>>55420152
>enabling shitty sales tactics like not rounding up on a number ending in four consecutive nines
>saying it's less than $400 even though sales tax exists
>saying it's less than $400 even though the page says right up front there's a $5 shipping charge
Holy fucking shit. Also you're wrong, reference cards were $450 MSRP, partner boards starting MSRP was supposed to be $380.
>>55420215
I'm not the guy you were arguing with, Einstein. Keep stroking your microcock to a fucking ($400) video card though.
>>55414985
I don't understand what the big deal is.
People have been tweaking hardware for better performance since hardware was invented. This is nothing new.
Pnr
>>55420252
nope. in the end 480 with a $120 block is retarded when you can buy a 1070 for the same price. even when you factor in the founder edition it's still only $70 more for twice the performance.
>>55416448
GeForce experience. 1click
>>55418120
960 actually didn't have any problems at all, neither did the 750
The 480 is literally the only card that has this problem, AMD fucked up bad
>>55416376
A 2 year old card. I have a 970 and even though I am pretty stoked about the 1080 and 1070, I'll have a look at the 490 first. The 480 however wouldn't even be worth it if it cost 80, because I already have a very similar performing card (since 2 years)
>>55420319
In the end these are fancy toys. People will do whatever they want with their toys and usually regardless of value.
Nobody questions the guy restoring an old tbird and blowing it (supercharger, not destroying the block) instead of spending that money on a better performing porsche. If I want to watercool a mid range card for kicks even if the final cost is around the level of a better card that's my fun and my decision.
>>55414985
Then why make it a motherboard killer to begin with?
>>55417230
Gee if only there was some form of removable storage you could use to transfer GPU drivers from an internet connected computer to the one you're bhilding
>>55420953
got to kill all the intel boards to make room for zen
It's a software bug in Wattman. The boards passed PCIe compliance because it wasn't used to test them. It was developed for the OEM cards that can handle higher Vcore because of their better coolers.
Simply put, the boards are fine, and a small change to Wattmans default profile is all that's required. And because they run faster at lower voltage there should be no effect on performance even with lowered voltage. They'll just run at the same stock speed whilst saving enough power to comply with PCIe.
Storm in a teacup. It hasn't blown up thousands of PCs, and was never likely to. Most electrical connections, including PCIe slots, are rated far lower than they are actually capable of. The 'margin of safety' built in is usually at least 50%.
Being 10% over will have no effect whatsoever. And that guy whose card blew up ? Well, it was probably going to anyway.
>>55420651
I do question that guy, and think he's pretty weird. Not a bad mean kind of weird, but just sort of backwards strange.
>>55419543
The 480 is sold out literally everywhere
https://www.nowinstock.net/computers/videocards/amd/rx480/
>>55421590
That is because its the new 970 as far as regular buyers go. Fast and cheap.
>>55420423
Okay faggot, Geforce Experience gave me the worst settings on GTA V. Shit was unplayable at 34 fps.
Sorry for the offtopic but i didn't want to open a new thread just for this:
Does someone knows where can i find the crimson drivers privacy policy?
It's trying to connect to google from some time now and i don't remember reading about it during installation.
>>55415770
It pulled 260w . What the actual fuck.
So pretty much the way to pull 1500mhz is to solder bits and do a shit ton to it. Yep because everyone at home can do that.
That just shows how much butchering you need to do on it.
A proper overclock with watercooling
https://youtu.be/ysnHdXqG7Gw
>>55422204
>260w
When will we see decent passive cooled cards?
>>55422255
Its stupid. Why the hell did they release a card with serious faults which how on earth they managed to miss is beyond me.
So they sold to the public before release as a mid level card such as the 390 but itll use less power.
Well thats well and truly a flying bucket of shit isnt it
>>55422291
There are no serious faults here, no hardware issues at all.
>>55422317
So the possibility of motherboards onboard sound chips melting on cheap boards isnt serious
>>55422348
>look at me regurgitating stupid shit like a child
There is no risk of motherboards "melting" from 80w~ being pulled through the slot. If that was a problem then every overclocker on the planet would be melting boards left and right since overclocked cards routinely suck power from the mobo 12v. The only adverse affect on board audio sees is noise from mild electrical interference.
The cards drawing power more heavily through the slot is controlled by by firmware behavior set in software. That is why a driver fix is being pushed today.
I look forward to your inane childish shitpost detailing how this is literally the worst thing to ever happen ever.
>AMD shills pretending that undervolting improves performance
Next they'll be saying that using cheaper parts and child labor improves performance
>>55422382
"regurgitating shit". What posting informtaition I have read from multiple sources.
You are just being a cunt as precious AMD have released a gpu which has failed in expectations and to its promise.
Also that video says it pulled 130w from the PCie
>>55422382
You are very wrong, it has already killed people's pcie slots and even shuts down some motherboards completely.
>>55422450
haha keep telling yourself that loser, AMD has beat woodscrews hands down every time they go head to head/
You need to wake the fuck up and leave your moms basement.
>>55422471
You are mentally ill
>>55422467
>that one miner without a powered riser
Gee willikers, putting more cards than the board has pcie slots sure is AMD's fault! Fuck AMD for my own stupidity.
>>55419574
>People don't trust 970 since it's outdated and has only 3.5gb vram
people aren't aware of that. They trust nvidia for the expertise.
>AMD has no concurrency in mid price range.
R9 390 was better for your bucks than the 970 in general.
>>55422450
>childish shitposting right on time as expected
The mentally ill are always predictable.
80w~ is the average draw from the 12v through the slot, and this issue isn't even affecting 100% of cards.
>>55416189
Nvidias DCC has to be 50% better in order to make up for it. Math genious.
Lads, should I buy a reference 480 and throw an ARCTIC aftermarket cooler on there, or just grab a Sapphire model?
>>55422518
Sapphire model obviously
Surely Vega will solve all of the 480's problems.
>>55422496
Its you having to do all the work here, attempting t justify why you bought one of these toasters.
End of they are a low/mid card which cannot beat a 970 in a lot of games.
Why get shot of a 970 or 290x for these? theres no reason to.
>>55422518
No, anyone with half a brain is waiting to see how widespread the burning out of mobos is before touching this shit.
>>55416817
kek got rekt screencapped for future nvidia shill
>watercooling the 480
why the fuck would you do that
How does anybody think this is an representative application of the card
Eh, I'll just go 1080TI when they come out in august.
I'm running R290's in Cfx right now and They can't even max out Doom. Not to mention the lack of ANY comparable to shadowplay.
>>55422700
i want a water-cooled r7 240 4gb (ddr3)
>>55422730
.. you should be able to max out doom... are you using latest drivers?
>>55422700
EK make a block for the 750. Yes, the 750 - the card down from a 750ti.
>>55422751
16.5.2 drivers
Overclocked to 1019C/1328M
Have to dumb it down to high, can't even ultra it. Even then I'm getting 40 in open areas. Fuck it even bars me from nightmare graphics
>>55422777
give 16.6.2 a try.
>>55415814
This
Jesus fucking Christ
>>55422730
>>55422777
>1080ti
>august
keep dreaming
>shadowplay
keep dreaming of that youtube career
>OpenGL DOOM
shitty benchmark. Vulkan patch will give a massive performance boost.
>>55422777
Nightmare is locked behind a vram wall iirc.
>>55422826
I taste the salt through the screeeen
>>55422826
>keep dreaming on 1080ti
>implies vulkan patch will get released this century
>>55416747
Go for the refund, get the Sapphire nitro.
>>55422884
>ark DX12 patch
>>55423191
Can we turn "keep dreaming" into a new meme?
>keep dreaming
skyrim vulkan patch
>>55419757
Electronmigration is a problem in miniaturized circuits like CPUs and GPUs. Some fat copper trace isn't gonna give a fuck.
Even the relatively thin traces that go up to the 5 12v power pins in the PCI-E slot are colossal compared to the kind of circuits where you'll have to worry about it.
>>55423215
>Nvidia's promised async driver
>>55423339
Well they need to, at least, not loose any performance while using dx12 (right now 900 series or older are slower on dx12 than dx11)
>>55423361
Their twitter said they were woking on a driver about 6 months ago. In fact they made this claim when the AoTS dramas kicked off.
>can't even drive native vga analog without conversion
enjoy your LatencyCD panels
>>55416205
The guy is fucking derbauer, you can be sure he binned the card or even got a cherry picked sample from AMD itself.
>>55416205
Still loses to 980.
>>55423422
480 only performs as good as a 980 when async compute is involved.
That's scenario that unlocks the GPU's potential. Good thing that's the future where we are headed.
Nvidia only retains the edge in outdated stuff.
>>55423440
DX12 is still a few years out from becoming widespread. It took more than 4 years until DX11 became standard and DX12 is locked to an OS that a lot of people do not want even though it's "free".
Buying a card that's on par with the two year old 980 for future games that may or may not use a particular feature of DX12 seems very silly.
>>55423485
DX12 has a faster adoption rate than DX11. Fun fact: over the first year of both apis Mantle had a faster adoption rate than DX11 did.
Naturally though Mantle became obsolete once DX12 and opengl next (i.e vulkan) were announced.
>>55422512
192 is 25% lower than 256 moron shill
>>55419799
I must say, I'm impressed with nVidia engineers and CI, they somehow able to utilize 120% of GPU. If only AMD was able to do something like this.
>>55414985
>>55414994
>>55415019
No fucking way
>>55421454
>>55424213
Disgusting
>>55419543
i just feel 1070 is just overpriced, its good price but cheapeast is basically 480euro and thats most shit edition. 1070 should cost 370euro, then its good card. Im waiting for custom 480. To check out there performance.
Why does 480 trigger so much shitposting
It's pretty insane, I wish we had IDs now..
>stable at lower voltages
>put high voltage on it anyway
>let it be thermally and power throttled
amd wat r u doin
Strange that it has not filtered to rabbit first.
>>55415104
get a 4gb and unlock to 8gb you dip
>>55424560
Holy fuck!
Your English is so broken you don't deserve a computer.
>>55424759
I understood it easily. Then again, I don't speak English natively and learned all of it through GTA 1-3, Bold And The Beautiful, Die Hard, the Shining and Terminator 2
>>55424759
You're so intolerant that you don't deserve an Internet connection.
>>55424213
>filth
>damage
>corrosion
Jesus, what does he do, stand on the case and blow loads into it?
No wonder it fried.
>>55424759
I fuck your mother to broken every night, she also said she is ashamed of her son since hes such a huge cuck and doesnt have no girlfiend, she even thinks you are gay cause your room is full pictures of fatman and pajeet man
soo the rx 480 is worth buying now?
>>55425143
If literally a max of 2fps perf increase by undervolting is enough to sway your opinion, then sure.
>>55425164
im just asking you rancid swine
>>55422826
nvidia wont let vulkan on nvidia sponsored games till they find a way to gimp amd with it
Has AMD released the update or detailed it yet?
>>55423485
dx12 is already getting a lot of traction mainly because there are 7 big companies that is pushing it
and only 1 that isnt
>>55425186
And I answered, dipshit.
If you already have seen all of the 480 benches out there, imagine them being half an fps to 2 fps better.
>>55425214
>there are 7 big companies that is pushing it
and only 1 that isnt
What the fuck does this even mean? There are 8 companies involved in PC gaming? Care to name them just for funsies?
Some 4GB models can be flashed to 8GB?
It's like some kind of hilarious inversion of the 3.5 shenanigans.
>>55425241
pc gaming? no not at all
remember this
its ms/amd/intel/samsung/ibm/sony/quallcomm
(i was gonna say apple too but they are dicks in general)
they are all pushing their variant of mantle
amd currently has 100% of the consoles ms is also going to use a cucked dx12 on xbox too...
eventually getting a game from consoles and porting it on dx12 will be much easier and way less time consuming than to create a dx11 path for it
that is what amd on their event was saying about a homogenous ecosystem
I literally cant find a fucking 4gb to snap up in this shit country, i dont even think we got sent a single one
>>55425274
well who didnt saw this coming? but people said the 4gb cards are using a "slower" samsung memory module whatever that means
its like all those 9600 ati that could become 9800xt lol
>>55425312
The 4 to 8 thing explains why there was a reported shortage of 4's in various countries. They were putting out a bare minimum of them to meet demands and contracts, since they were really losing out on selling them.
At least until true 4GBs are readily produced.
>>55423638
DCC only works on texture data and certain types of it at that. The GPU processes a lot more maths than just shuffling textures around, and so the actual benefit from compression is limited.
50% =/= 50% real bandwidth saving, hell probably not even 25%
>>55425373
In the UK Overclockers.co.uk got sole rights to sell the 4gb cards I believe. (selling in the UK)
