![maxresdefault[1].jpg](https://i.imgur.com/vyRJveY.jpg "maxresdefault[1].jpg")
>Nvidia ships the 970 with a "supposed" 3.5 gigs of vram.
>Shitstorms of people going crazy over themselves angry at Nvidia despite the fact that the 970 DOES come with 4 gigs of vram.
>AMD ships the 480 and draws more power from the motherboard than is recommended and is able to completely fry motherboards
>Everyone is defending AMD
Why the double standard?
only /v/ is defending amd because /v/ is contrarian
>>343644441
Because the problem only affects some of the users, and they can easily avoid it by either undervolting the card or not overclocking.
970 on the other hand will always have that 3.5 problem.
>>343644441
The extra 500 MB wasn't vRAM
Still not as bad as purposefully frying your mobo tho
>>343645607
So the "fix" that is acceptable for you is to lower the performance of the card yourself so that it doesn't fry your hardware. AMD defense force going all out.
>>343645607
>they can easily avoid it by either undervolting the card
Damn, so this is the power of AMDogs.
>>343644441
Article says the problem is with the single 6-pin connector. So why the fuck didn't they use a 8-pin or 2x6?
i have a 290x 4gb, would upgrading to the rx 480 8gb be worth it?
>>343645607
>Easily avoided
>Literally no way of knowing if your motherboard can deal with it or not
>Literally no way of knowing for sure until it's already been at full load
Man you sure am retarded
>inb4 gimp yourself just to be safe then
>>343645795
No, the actual fix will arrive later with a driver update. Everyone who bought 970 had the 3.5 problem, only a few with 480.
This is a non-issue blown out of proportion.
i have a 970 and i really like it
>>343645886
its only the start of the problem.
The REAL problem is they designed it from the ground up to run on the 6pin and will have to deseign the entire board to get it to stop how it draws power. this isnt so much of a "its not getting the power from the 6 pin so its drawing it elsewhere" problem, its a "this was designed to draw this much power from the pin and this much power from the rail".
IE, the entire first run is fucked and will need a 2.0
>>343646081
Do you like your comp?
Because the 480 is frying comps and could start a real, legit fire.
>>343644441
>>Everyone is defending AMD
You what? Everyone thinks this is a massive fuck up, only shills defend the 970 and 480.
>>343646081
>upgrading
>>343646292
AMD has no shortage of shills.
Some fun reading
https://community.amd.com/thread/202410
>>343644441
Well one was true and the 3.5 GB problem is there no matter what. The other will get fixed in a driver update on tuesday.
I'm pretty sure there's been equal amounts of shitposting for both at this point.
>>343646553
>this coming from another rabid nvidia shill
>>343644441
the amd problem is actually not an issue if your motherboard is fairly new and not a piece of shit.
it's been completely overblown
>>343645607
damn i sure hope this shitty card that can't even compete with a 970 doesn't fry my system!
>>343644441
because AMDs problem is a bug/mistake that can easily be fixed with a software update?
how is this even a question?
>>343644441
Serves them right. I like AMD and all but I always thought the whole obsession with relying on PCIe as much as possible (removing SLI bridges, using fewer/smaller PCIe power connectors) was fucking retarded. It's the same kind of retarded as trying to build a powerful gaming and productivity PC in a MicroATX case just because.
>>343646081
Don't listen to sensational fucks like >>343646247 and just get the Sapphire Nitro for the RX480, it has an 8-pin connector.
Yeah the power draw thing is a problem but it's already "fixed" by aftermarket cards. Meanwhile every version of the 970 has the 3.5 problem.
>>343646553
>https://community.amd.com/thread/202410
im enjoying this thread
>nvidia shill has been said like 30 times
>I LOOKED UP THE SAME SCREEN NAME AS YOU AND FOUND THE WRONG PROFILE BUT IT DOESNT HAVE THE SAME WITCHER HOURS AS YOU SO YOUR AN NVIDIASHILL
>I see you own a 750. That broke your system, not the AMD card.
>that one mother fucker analyzing the images in photoshop and other programs to 'determine if the OP is an nvidia troller'
>I DONT SEE A HDD, SO YOU MUST BE AN NVIDIA FANBOY
>>343645607
>prayer is easier fix to a problem than lowering graphics settings a little
>>343648086
>I have no idea how 8pin works
>I have no idea how GPU's are designed and work
>I have no idea what the term 'aftermarket' actually means and just parrot AMDfags
Because everyone loves rooting for the underdog.
>>343647094
More like, hardware update. They used cheap materials to reduce production costs.
>>343646553
>Your post is appreciated, but frankly speaking, is clearly trolling. You either used a custom input display for the error code in the given photos, or you physically damaged your motherboard earlier and now expect AMD to pick up the damages. You're going to need a lot more evidence, Chief. We can't have guys coming on here and fabricating stories when they're hitting their motherboard with a hammer or spilling a little piss in the motherboard to try and get their big AMD payday.
this is some next level fanboyism right here.
>big payday
>96$ motherboard
>op already said he wants nothing from AMD
god damn these people are sad
>>343646220
No. Buildzoid did a livestream of overclocking and modding, and found the PCB is designed in a way that the VRMs are supplied exclusively by the PCI-E slot. It's a hardware design fault which software can't fix.
https://www.youtube.com/watch?v=plC7tOYIqBw
>>343648394
Instead of greentexting like a meme-loving fuck maybe you can tell me why an 8-pin power connector wouldn't alleviate the issue.
>>343647094
I can tell you right now that you're getting hoodwinked. This cannot be fixed without PCB redesign. It can be circumvented to an extent using software by lowering the TDP of the card. Problem is that this means performance also suffers.
>>343645886
From what I heard and take this as a grain of salt, AMD wanted to appear more efficient to the technologically illiterate by using a single 6 pin.
8 pin and 2x6pin are scary and signal that your graphics card is super inefficient. Even tho 1070, 1080, 980, 970, etc all have 2x6 pin or 2x8 pin configurations despite being pretty efficient cards. The GTX 960 has a single 6 pin, but that's way below the 480 in performance and TDP.
>>343648636
they literally made a press release saying theyre going to fix it via software update. stop spreading reddit memes, friend.
>>343648086
>dont buy the nvidia card thats posting higher scores than the 480
>buy my shitty poo in the loo backed GPU thats made in a factory that looks like they made shitty toys in instead
>if your lucky, you might EVEN get one with the antistatic bag!
>>343648963
>This cannot be fixed without PCB redesign.
welp, thats sad if true.
>>343645607
lmao this is what AMD fags consider an appropriate fix.
>>343648978
>they literally made a press release saying theyre going to gimp your card via software update
fixed that for you friendo
>>343649105
>dont buy the nvidia card thats posting higher scores than the 480
Which one does that at $200?
>>343648636
This is false. The components used are not cheap. It ultimately comes down to an engineering design fail.
>"supposed"
>>343644441
Because AMD is the scrappy underdog keeping NVIDIA honest.
If they go out of business we are all beyond fucked. NVIDIA will engage full Jew mode. Enjoy your 18 month midrange revisions for $600
>>343649118
If you want the card, just don't buy the shitty reference design. Any board partner can build a better board. Hopefully whatever driver fix they do to it won't affect properly configured cards.
You're really good at generalizing, OP
Both sides of this bullshit argument defend their side and attack the opposing side. /v/ has not been a hivemind for years, and the fact that you haven't noticed this tells me you're young and new to this website. Please die and stop baiting people into arguing with your shitposts
>>343649173
no they are and have been notorious for using cheap parts, lets not lie to ourselves and everyone here, remember in 2009 some of their cards melted their fans.... While running WoW lol
>>343648885
because 8pins, while put out more power, are only 2 extra grounds for safety (which is why they can put out a little more power). The card itself is designed to draw said power from the rail/PSU no matter the circumstance. Its how it was designed. card manufacturer's dont change the actual design of the cards, they just tweek it here and there.
What your sugesting is basically like saying EVGA or MSI's 970's dont use the 3.5/.5 ram set up and instead use straight 4gigs. That doesnt exist because its how the card was designed. Its built into the actual fucking card's DNA. No one except AMD redesigning their card can fix this.
>>343649238
>keeping NVIDIA honest.
But they don't.
>>343649528
RX 480
https://www.youtube.com/watch?v=qG2e-v94L4M
GTX 480
https://www.youtube.com/watch?v=fz8o3OiHYt0
>>343648963
My Strix 970 has a single 6 pin connector.
>>343649840
>starts his analysis with begging for money to buy an AMD card
>puts a fucking time limit on it and the beggar tells his fans to hurry up and donate so he can buy it TODAY
>uses paint
Im having a hard time even getting past this part. Hell, he hasnt even done a single piece of the analysis and we're 2m in.
>>343650232
What a piece of shit.
>>343649803
>a little more power
You mean twice as much (150W vs. 75W)
>>343650414
in my mind, thats a little, but you are right, it is double. Doesnt change that its still designed to take it from the rail as it is
>>343650327
Actually it's a single 8 pin.
>>343650285
>please donate
>somebody needs to make up the extra cost
this cant be real
>>343646147
>inb4 gimp yourself just to be safe then
I cut off both my legs to avoid falling.
>>343649803
There is a difference in something that is designed into the GPU core and something that is designed into the board. Power delivery is on the board. Memory controllers are on the chip. Nvidia could, by the way, enable the full 256bit interface on the 970 if they wanted to but board partners couldn't. In this case, board partners can change the way the power delivery circuit works.
>>343650795
That's hell of a different from a 6 pin innit?
>>343650958
I guess. I'm not an electrician but it's overclocked out of the box so I guess they know what they're doing.
Good thing I have no interest in overclocking.
>>343649165
1060
Have the GTX 750 TI boost. Should I wait for the 480 fix or just drop it and move to a 970?
>>343651357
Wait for 1060.
AMD already fixed it with a driver update and it's not even a big issue to begin with.
>>343645607
>undervolting and gimping performance a-okay because AMD
>reducing VRAM usage or capping it not okay because nVidia
AMD drones, ladies and gents.
Upgrade every gen or every other gen?
Got a 970. Wondering if I should go for the 1080 (which will be a few months before it becomes available for MSRP) or if I should wait eternity for an 1180.
>>343652072
undervolting does not gimp performance. Less heat and less power draw results in higher boost clocks.
As long as it remains stable that's a perfect fix for the problem.
>>343652309
if you want small increments then every gen, more expensive. I would probably recommend every 2 or 3 gen.
>>343652610
>undervolting does not gimp performance
>>343652772
Do you even know what voltage is
>>343652865
Do you?
>>343652772
It's true if clocks are not changed. I don't see why they wouldn't be changed, though. Unless the card is massively overvolted to begin with.
>>343652865
You obviously don't.
Yes, let's undervolt stock cards and hope to god they'll run stable at current clcokspeeds.
Fucking dolt.
>>343652610
>>343652610
>undervolting does not gimp performance
AMD and nvidia both suck you fucking children
>>343646220
Literally a fucking hired shill or lunatic on /v.
Congrats. You're fucking retarded.
>>343653562
>>343653449
Are you the elusive Sony Ger?
>>343653536
remember
people who have different viewpoints from your own are always shills
>>Everyone is defending AMD
Are we on the same board?
>>343646220
>3.5
To be honest so was the 3.5 thing and yet here we are.
>>343652954
>>343653083
>>343653338
If the lower voltage is still within the components' specifications there won't be an impact on performance so long as clock speed is maintained.
This isn't even a GPU thing this is a basic concept of digital circuits. You are still running at the same clock speed. You still have the same FLOPS. But you are drawing less power. The danger comes when your voltage becomes too low, but you don't get performance drops you just BSOD because 1's start looking like 0's.
Undervolting can conceivably give you a performance increase because it can allow you to maintain higher clock speeds while not overheating. The ONLY reason you think undervolting sacrifices performance is because you think MOAR POWER = MOAR PERFORMANCE.
>>343644441
welcome to /v/ where everyone is a cheapskate and pirates everything on the cheapest computer components money can buy >>> AMD
>>343652772
>>343653185
>>343653338
>/v/ is this tech illiterate
>>343644441
Because it doesn't do anything to your motherboard. A simple overclock achieves the same effect. No one will tell you overclocking is a bad thing even nvidiadrones. Only shills are attacking amd for it.
>>343654124
>This isn't even a GPU thing this is a basic concept of digital circuits
And yet this poor AMD drone can't even grasp that.
But do keep on spouting buzzwords and terms you don't even understand, it's very amusing to see someone struggle so hard with their cognitive dissonance that they just start making shit up.
>>343644441
It hasn't fried anyones motherboard. Just like people memeing about crimson drivers, AMD will reveal the obvious: That there are failsafes that prevent that from happening, and nvidiots will be silent.
>>343654720
How about you prove him wrong?
>>343654475
A massively overclocked 390X takes like 50W from the PCI-E Slot. That's half of a 480.
>>343654124
Have you seen motors running? They run faster the more voltage they have while drawing current as required. Even if the circuitry works, its fans will go slower and dissipate less heat, so there will be worse performance due to heat at least. It will probably run worse too due to other things inside that I don't know about
>>343654124
>If the lower voltage is still within the components' specifications
That's a mighty big *if* there, pal.
The entire process of binning depends on how well each card tolerates voltage, heat and performance. You can't just undervolt every card and expect them all to run at the same clocks.
>>343655175
>Even if the circuitry works, its fans will go slower
Why would fans be affected by core voltage?
>>343652610
>undervolting does not gimp performance
I thought summer /v/ was just a rumor.
>>343655175
Fans will still be running off a PWM controller that does 12V. The voltage will not be changing on anything other than GPU core and/or memory.
Undervolting reduces chip temperatures. My i5 750 is undervolted by 0.1V and it has a significantly lower TDP at the same performance as stock.
>>343644441
AMD fags are reactionary and a vocal minority
who knew?
>Why the double standard?
Because most people here are poor. Simple as that. It's not like they have the choice to buy nvidia.
>>343655568
Ironically the reason AMD marketshare is so low is because their cards were unaffordable to gamers for a long time in the past.
>>343655175
>when /v/ tries to /g/
Doesn't the 750Ti draw too much power through PCI-E as well?
>>343655858
No. You're probably thinking of the 960 Strix.
>>343655858
Nope, only the 6 pin versions are capable of drawing more than about 70W overclocked.
>>343655661
Their quality went to shit after the AMD aquisition, which correlates with marketshare.
>>343644441
/v/ is too poor to buy nvidia and 480 was their last hope
>>343656305
Nah, AMD cards haven't changed in quality. They were just fucking ridiculously expensive back when 290X was still the hot shit.
The problem with the 480 affects >5% of users and is due to a manufacturing issue.
The problem with the 970 affects 100% of users and is due to Nvidia designers being incompetent.
This is basically the end of AMD, motherboards will be dying left and right, there are already confirmed reports that this card has fried systems
The ensuing PR hurricane will effectively end them as anything but a ghetto option for people who can't afford an nvidia part at all
>>343654475
It's a well known fact that overclocking reduces the lifespan of your components and increases the chance of failure. Simply increasing the voltage regardless of cooling increases degradation of components from the increased electron flow.
>>343656725
>>343656867
>The problem with the 480 affects >5% of users and is due to a manufacturing issue.
The problem with the 480 affects 100% of users and is an intentional design issue. A manufacturing issue could not produce this result
>>343644441
>Why the double standard?
Retarded brand loyalist consumers.
>>343656867
>>343645607
This AMD shill is out of control
>>343656867
480 has a chance to fry your mobo because of the problem.
970's problem means you get lower performance in some games at high resolution/settings.
>>343645795
>>343645827
>>343646147
ey can easily avoid it by either undervolting the card or not overclocking
by EITHER
>EITHER
do you know what this word means? It means one of the options provided after that word. So by not overclocking, you already avoid it.
I'm not on anyone's bandwagon, just pointing out your failing reading comprehension and retarded straw man
>>343657495
the power draw issues have nothing to do with overclocking and were confirmed by 5 different reputable hardware review sites using stock clock
>>343657495
>by not overclocking you still avoid it
no you don't
the 480 still pulls significantly more current from the PCI power on the 10th and 11th pin of the 12v rail by more than 50%
>>343655175
I'm just gonna be nice and assume you're not trolling, and really don't know anything about electrical anything.
Motors are not digital components. Motors take electrical current and turn it into rotational force using the Lorentz force, which increases with electric field intensity (E) using the equation: F' = qE' + qv' x B'. So increasing voltage will increase its force, which generally increases its speed (exceptions being synchronous motors which run at a fixed frequency).
The reason why undervolting does not impact performance of a processor is because, once the voltage is at a point where everything functions as intended (e.g. transistors switching on and off properly), higher voltages don't do anything. This is why you never hear about "overvolting" to gain performance. It's not because higher voltages can damage components (they can, but you're probably safe going a few tenths of a volt higher), it's because it literally doesn't do anything.
In the same way, undervolting doesn't decrease performance. So long as you don't hit a point where the voltage is so low your card starts operating differently, you can indeed use less power and get the same performance.
Also your fan's voltage and your card's voltage are not related.
>>343644441
>able to completely fry motherboards
if you have a shit mobo, sure. what people fail to realize is that most cards overdraw at some point in time. ever overclock? you probably went beyond spec.
>>343644441
ayymd
>>343658316
>if you have a shit motherboard you deserve it
>even though your motherboard worked with every other GPU, this faulty one isn't at fault
If the card literally expelled noxious gas that knocked you out every time you turned your pc on AMD drones would still defend it.
Can't fucking make this shit up. Its hilarious.
>>343644441
Because people love underdog stories.
>>343658316
Even the power monster known as 295X2 didn't draw over 75W from the mobo.
Properly designed cards don't overdraw from the mobo, regardless of how much power they draw total.
>bought nvidia because muh cuda muh compute
>they turn out to be the biggest jews
>>343658574
>If the card literally expelled noxious gas that knocked you out every time you turned your pc on AMD drones would still defend it.
i leffed
>>343644441
Because people on /v/ are largely poor fags and anything over 200 bucks is a personal insult to their way of life.
>>343658638
Heh you could run 3 of those before you'd reach the same levels as a single 480.
>>343658316
>you probably went beyond spec.
This doesn't make any sense. The "beyond spec" is rated for +/- 9%, no other GPU's go out of spec like that and overclocking has nothing to do with that
>>343658645
>bought amd because cheap poorfag
>can't play games normally because drivers crash all the time
>>343658224
>In the same way, undervolting doesn't decrease performance.
In order to run your hardware at a certain clock, it requires a certain amount of voltage. Nobody here is arguing that voltage controls performance directly, it does so indirectly. You're just arguing technicality because you're an AMD fanboy and you're buttblasted about the 480 having a major issue.
GTX 960 for $160. Get it? I have no GPU right now.
https://www.amazon.com/gp/product/B013LDXPFO/
>>343659065
it's probably the best choice for only $160
Hold up. The 6 pin is made as extra power. Unless im wrong, doesnt the card default to using the pci slot power before it pulls from the 6 pin? In that case, it means underclocking or undervolting doesnt matter unless you cut the power usage by over half.
>>343659065
Just wait for the 1060.
>>343659265
its largely about current, the 480 overdraws current from the PCI slot that puts unneeded stress on the motherboard
Im looking to run a 1080p TV, 1080p monitor, and 720p monitor. I like modded skyrim, emulators like dolphin, lag of legends, and grand strategies. I will not accept under 60fps, im fucking picky. I have a 270x atm. What would be a noticeable upgrade for under 300?
Where my GTX 680 bros at?
Still holding out strong, runs pretty much everything just fine still.
>>343659501
>What would be a noticeable upgrade for under 300?
the 390 I think is under 300 now
>>343659572
Would i get like double performance?
>>343659648
You would get a significant jump in performance over your 270x
>>343659705
thanks mate.
>>343659696
eventually i wanna use free sync, monitor for them is 150 dollars cheaper. I have never had problems with nvidia, but g sync is too expensive.
will aftermarket 480s suffer this same problem? I was going to purchase a 480 but now I'll probably just wait for the 1060.
>GTX760 releases
>no one bitches about it drawing more power from the motherboard than recommended
>RX480 releases
>everyone bitches about it drawing more power from the mother than recommended
Why the double standard?
Why is the 290x so amazing?
I want to upgrade but every damned card that is better is like 500+ bucks. Its ridiculous.
>>343660051
>will aftermarket 480s suffer this same problem?
unless they undervolt, yes
>>343659974
gsync doesn't cost anything extra, you're just looking at cheaper monitors
>>343659974
wait forprobablyfixed aftermarket 480s then
>>343660067
>>GTX760 releases
>>no one bitches about it drawing more power from the motherboard than recommended
the 760 doesn't go over spec
>>343660114
cheapest g sync 1080p is 370. Cheapest free sync is 230. Thats a lot of money for the same effect.
>>343657815
Not him. That burnt 24 pin ATX connector looked like it got burnt on the 3.3 volt pin, intead of the -12v pin.
The card currently has too much voltage for the clocks its running, I've heard people have been actually gaining performance after undervolting, thanks to it not hitting the power limit. Dont know what effect it has on the boost clocks.
Im still waiting for AIB cards, got me an old PCIE 2.0 board from 2008 with an i7 920, dont want to take any chances
>>343660051
Just wait and see, they might fix it they might not. AMD is also going to release a driver update on Tuesday which may or may not fix the issue without gimping the card.
Ultimately you should wait for the 1060 to come out anyway.
>>343644441
970 also draws more power and no one cared
>>343660114
What is the cheapest 1440 gsync/freesync?
Thinking about getting a good 1440 card and a 1440 monitor to get with it.
>>343660172
I dont think aftermarket fixes the problem. The traces are set in stone by amd. The card is built the way it is. The manufacturers can add shit, but if i recall correctly, they cant really change the power it pulls from pci vs 6 pin. Adding an 8 pin doesnt change anything i dont think.
>>343660379
970 doesn't draw more current than the PCI allows
>>343660207
>cheapest g sync 1080p is 370. Cheapest free sync is 230.
......you have to compare the monitors bro.
gsync doesnt cost anything extra. freesync, like amd products, is for cheaper products. understand?
>>343660070
290x was way ahead of its time.
it is just overshadowed by r9 390 atm because it's more accessible. However r9 290x is actually more hardware contained than the r9 390 so in most cases it can actually keep up.
I have r9 290x and I am amazed how beast this card is despite because 3 years old.
>>343659265
Apparently people have been able to under volt without lowering clockspeed.
That should reign the wattage in a bit.
>>343660484
g sync does cost extra dude. It has a card that had to be put in the monitor that nvidia charges for.
Who gives a shit? Get a evga 1070 sc and be happy as fuck
>>343660205
I meant the 960 lad
Just a typo
>>343660770
the 960 didn't either
ASUS's 960 strix edition was extremely overclocked and overvolted but it didn't fly motherboards and it only went over spec current by 15% not 50%
>>343660484
My benq 2411z was something like 100 bucks less than the gsync version. Stop horseshitting.
>>343660770
http://www.pcper.com/reviews/Graphics-Cards/Power-Consumption-Concerns-Radeon-RX-480/Evaluating-ASUS-GTX-960-Strix
>>343644441
one of the two has a driver fix.
>>343660460
actually it does draw more than 75watts through PCIe. And: How much power the mainboard provides through the PCIe is the choice of the mainboard producer that can set a multiplier depending on what he thinks the mainboard can and can't do. You can set the multiplier yourself even if you want to do it.
>>343655175
Different circuits on the board man
>>343661224
>actually it does draw more than 75watts through PCIe
I said current
you can have the same wattage with different voltages and currents
>>343657495
>So by not overclocking, you already avoid it.
Except that even at stock settings it still overdraws.
>>343660379
No it doesn't. It also has 2x6 pin which gives it plenty of overhead.
These figures are from stress test:
Motherboard 12 V 41.07 W 52.83 W
48.36 W
Min max avg.
So yeah, I guess nobody gives a shit if it goes up to 52W when the spec is 66W.
>>343644441
Because it took months to find out about the 3.5GB issue and by then everyone had bought one.
>>343661224
You don't know what you're talking about and are spouting gibberish in an attempt to defend a corporation that doesn't know who you are.
Just stop.
This is a real, serious issue with the 480. People aren't trying to hurt your feelings.
>>343653901
Its Nvidiots and AMD apologists clamoring for the biggest retard award.
>mfw the sapphire and partner boards will fix everything
Nothing of value was lost.
Doesn't the 970 actually have 4 gigs of ram, just that .5 of the ram is slower?
>>343651118
Didn't know the 1060 was out or that people had samples of it yet to prove that
>>343661472
Yes.
Actually having 3.5 GB vram would be worse, because even though the 0.5 GB portion is slow, it's still much faster than system ram.
>>343661398
They can't.
Partners don't have access to change the actual PCB, AMD has to revise the base card itself. This is not a thing that goes away overnight
>>343661768
Partners do have access to change the PCB unless AMD goes 1 step beyond full retard. MSI has fully custom PCBs for their nvidia cards.
>>343661472
The legitimate worry with the 970 is that Nvidia will not so much downgrade, but neglect the drivers/per game optimization on the card as it gets older.
You need the drivers to tell games that the last .55 gb basically doesn't exist, so the game can budget itself.
It doesn't lag as bad as if you totally ran out of Vram, but its still something that they'd need to implement on a per game basis, extending on into the future.
Happened to my 760, meanwhile competitive AMD cards are outpacing my old card aincd they're still supported
>>343662072
>but neglect the drivers/per game optimization on the card as it gets older.
this is categorically what they do not do, however. it's literally a made up thing from 4chan.
>You need the drivers to tell games that the last .55 gb basically doesn't exist, so the game can budget itself.
Yes, anon, that's why they fixed it in drivers like 2 months after the card came out. You do not have to make these changes for each game, you fix it at the driver level and games are agnostic to the changes.
>Happened to my 760, meanwhile competitive AMD cards are outpacing my old card aincd they're still supported
You have to do some serious lying to yourself to make this scenario true.
>>343662072
That depends on how similar future architectures are. Both AMD and nvidia neglect architectures that are not similar to their current one.
For example 5770 has absolute trash tier support because it's the previous architecture.
At the same time something like 290 still has good support, because AMD just rebranded the same architecture for 390. Same goes for something like GTX 480 back in the day. It still had good support while GTX 580 was the top dog, because it shares architecture.
Kepler is very dissimilar from Maxwell, which is why support for it has gone poof.
>>343660887
>nvidia shills just straight up fucking lying
I'm seriously getting tired of this shit, nvidia. Just fuck off already
http://www.tomshardware.com/reviews/nvidia-geforce-gtx-960,4038-8.html
>The very frequent spikes beyond the motherboard slot’s supposed limit won’t cause immediate damage to the hardware, but there might well be long-term repercussions that are hard to judge now. The same goes for how the system might otherwise be impacted with problems such as “chirping” on-board sound when the mouse is moved. The Asus GTX 960 Strix should do a much better job smoothing these spikes out.
>>343662524
The reviewer properly pointed out what he felt was a problem, but since the ASUS board failed to burn any motherboards and it's isolated to 1 vendor it never became a big deal.
Also, I know for a fact that some other benchmarkers have tested that ASUS board and found nothing anomalous with the PCI-E power consumption.
>>343662887
>Also, I know for a fact that some other benchmarkers have tested that ASUS board and found nothing anomalous with the PCI-E power consumption.
Explanation on that stuff:
>One interesting note on our data compared to what Tom’s Hardware presents – we are using a second order low pass filter to smooth out the data to make it more readable and more indicative of how power draw is handled by the components on the PCB. Tom’s story reported “maximum” power draw at 300 watts for the RX 480 and while that is technically accurate, those figures represent instantaneous power draw. That is interesting data in some circumstances, and may actually indicate other potential issues with excessively noisy power circuitry, but to us, it makes more sense to sample data at a high rate (10 kHz) but to filter it and present it more readable way that better meshes with the continuous power delivery capabilities of the system.
>Some gamers have expressed concern over that “maximum” power draw of 300 watts on the RX 480 that Tom’s Hardware reported. While that power measurement is technically accurate, it doesn’t represent the continuous power draw of the hardware. Instead, that measure is a result of a high frequency data acquisition system that may take a reading at the exact moment that a power phase on the card switches. Any DC switching power supply that is riding close to a certain power level is going to exceed that on the leading edges of phase switches for some minute amount of time. This is another reason why our low pass filter on power data can help represent real-world power consumption accurately. That doesn’t mean the spikes they measure are not a potential cause for concern, that’s just not what we are focused on with our testing.
>>343663568
>That doesn’t mean the spikes they measure are not a potential cause for concern, that’s just not what we are focused on with our testing.
only important line out of two paragraphs of someone typing for the sake of it
>>343663919
>potential cause
You didn't see the key word. Potential. I doubt the 300W spikes that RX 480 produces are relevant. I don't think there is anything on the mobo power plane that can be damaged by spikes of high current.
>the exact same people who posted the 3.5GB meme daily are now desperately and frantically trying to find bad things to post about nVidia and coming up with excuses after the blunder that was the 480
>>343660067
750ti also draws more than 75w through the PCIe
I looked into it a bit and its a non issue, just shills trying to downplay AMD as usual. The guy that got his motherboard burnt was running 3x RX 480s plus a FX power hungry CPU, and his motherboard looked old as fuck. As long as your motherboard aint from 2005 this isnt a problem.
>>343665017
CPU power draw doesn't matter. You're undermining your own credibility on the issue by including irrelevant pieces of information. This suggests you don't understand what is going on and are just talking about shit you don't understand even on the most basic level.
>>343659065
>960
>ever
that card was bad when it was released. Nvidia is only good for the high end
>>343665230
The guy that got his motherboard burnt was running 3x RX 480s on a decade old motherboard
is that better?
>>343665390
No, because his motherboard was not a decade old - AMD FX processors aren't that old. They were released in 2011 or so, which is about 5 years. Eg less than a decade.
In 2011 we had graphics cards like Nvidia GTX 480 which took double the power of RX 480s and supported 3 way SLI. It is safe to assume that motherboards from back then are more than adequate for running 3 cards.
>>343664672
>4GB is the standard even for midrange cards
>Nvidia still ships 3.5GB and 3GB cards
>>343666454
AMD's flagship can't even play ME on hyper settings because it doesn't have enough VRAM
L M A O
>>343665259
i bought a 960 and im enjoying it fine. will get a 1060 next :)
>>343667736
Hyper settings were intentionally implemented by Nvidia to fuck over any card with less than 4GB, even their own.
Pointless insane view distance.
>>343668061
>Pointless insane view distance.
pick one. insane view distance is pretty much a holy grail type thing, having it available at any cost is worth it to some people
>>343668061
>it's only pointless when AMD can't do it
m-muh muh-m-muh nvidia boogeyman!
L M A O
Both companies suck at making reference cards and only fools buy them OP.The 970 with 3.5rm cant be fixed it's by design but the AMD overvolting can with ease from other manufacturers who were going to do it anyway with there PCB skills.
>>343668916
>tech illiteracy
>spelling mistakes
Sounds about right.
