GTX480 / GTX470 Reviews and Discussion

Page 35 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.


No, that would never happen, they would never actively try to get reviewers to change their reviews or influence the public;
http://www.legitreviews.com/article/1264/1/
http://www.fudzilla.com/content/view/18329/1/
http://www.hardocp.com/article/2010/04/01/geforce_gtx_480_real_world_temperatures_sound

Sure and they had nothing to say/do about the noise comments either. :evil:

as if ati's decision makers wouldn't do exactly the same things if they were to trade shoes.

Sure but they also caught flak for it as well, or did you miss the HD2900/R600 series like I mentioned. Heck they caught as much flak for the X800 series being hot initially too with the X800PE being very hot, however when the GF6800UE never made it to public due to those same heat issue, that noise suddenly died down too.
So how is this different? :heink:

IMHO, that's a bit unfair considering they are taking the risk of taking flak knowing that it's the part of the pains of innovation with brand new technologies. When you start from scratch and reinvent a massively huge and complex amalgamation of the newest and completely unproven scraps of engineering, you're bound to take the risk of ending up in a carfire on your test flight.

And if you do you get criticized for it, and rightly so. And when you do it right (like the R9700, GF8800, etc) then you deserve praise. I don't understand your issues other than having some kind of feelings for a corporation, which really doesn't care about you or the user (none do).

The thing's hot, everybody knows that yet people still feel the need to colorfully remind us in the most pejorative way I've seen on any anglophone hardware board. Doesn't any of their huge list of contributions to the world of graphics technology earn them the right to expect a little more respect or at least the benefit of the doubt when it comes to the next tolerating one miserable generation of products that are merely decent as opposed to amazing?

WTF are you on about? nVidia is not alone in that respect, and what they did they didn't do for the good of all mankind, but for their own profit. What of the negatives they have done (as have others)? They should get equally good/bad treatment as anyone else, just like AMD's sucky CPUs when they're crap, and intel's when they're crap, perfect example similar to this would be if everyone hushed up about the very warm prescotts. Should that have been hidden for all the 'good that intel has done... etc.?'

And forgive my saying but moderators are supposed to be broad-mindedly fostering...

No, we're here to keep the discussion in bounds; and framing your statement like that is the opposite of trying to foster discussion. :mmmfff:

As for the gtx 480 not being supported,...

It's not, and your posts in this thread would've lead others to believe it is right now. You can explain it however else you want, it doesn't change that, and the fact that support is a future prospect. Similar to the AE comments and then the excuse of it not being needed if it's not already in the cards or else coming too. :heink:

And from the sound of it, I don't think you know what the opengl enhancements are in their current state if you're comparing them to the mercury playback engine

I didn't compare them, nor compare the base OGL acceleration to the dedicated plug-ins either; however I also don't falsely attribute them to one camp or the other to give them a boost for all that they've done for mankind. :pouah:
And I don't think you're in a position to say what other people know and don't know especially in reply to a post showing your statement to be erroneous. Seriously dude. [:thegreatgrapeape:5]

Now start giving Matrox and 3Dlabs some love for all they have done for the Graphics world !! [:jaydeejohn:5]
 
For my $0.02 on the OC'ing discussion;

HD5850 hits 80c at 25% fan, on average. This isn't a fermi-suitable chassis, just your average-joe (not OEM though~) build. Fermi, in this situation, would probably hit around 95c @ 80%fan (95c is the agreeable temp the fan kicks in). This is when we know the card gets obnoxious.

Now, let's say we overvolt. The 5850 would go to 85c @ ~55% fan, Fermi? I'd guess 100c @ 100% fan. And IIRC, running fans at 100% will kill them. Then you have the issue of heat in the rest of your case, and noise. But if you don't care about that, you have to consider that gaming for a few hours a day on a 100% fan is not a good idea, unless fan technologies have improved (which I don't pay attention to!).
 


Oh boy, what am I getting myself into?
Alright, I guess it would have been wise to disclaim that I have no interest in this polarized high school parody of a political debate between red and green. But that aside, are you seriously suggesting that the links you provided point towards irrefutable evidence of shady pay offs? I'm honestly confused here. I'm sure I deserve the sarcastic attitude by someone's standard, but please, elaborate! Are they wrong to say that the massive exageration is unjustified? And I'd love to know what's going on with these crude photoshops of ceos on review sites that's worth losing all credibility for! Is that how low discussions on this subject have fallen?



Fair enough. That's all good to be unbiased in who we shake our fists in anger at, but what's the aim of our outrage in the first place?
Do you really think that it's the thousands of negative comments on bulletin boards that pressured engineers to solve all the technical shortcomings of these respective designs? Apparently that's one thing that hasn't become any different in all this time and that's really contributing to the problem.



Wait a second here. Why is that rightly so? Can't you see that investing in a new architecture, with a new potential ceiling, new feature set and new horizon of possibilities is good for all of us? I don't think there's anything in it for nvidia to release lesser products in a competitive market especially at a time like this and therefore it shouldn't be farfetched to say the shortcomings weren't intentional; so what's so important about chewing out at a team of engineers who failed at impressing an audience oblivious of the efforts they put in on every aspect, even though they tried their best within their limits? Would you rather they had simply made the gtx400 a declination of their older design just smaller, faster and with stamp on additions of the bare-minimum of new features to earn the dx11 sticker so years after years, we're looking at the same things repackaged?



This is getting less and less civil. Don't get my tone wrong here, I'm in no way trying to be confrontational.
To answer your question, I'm on about the fact people seem to think nvidia's failure stems from laziness or "negatives", whatever that's referring to. It's as if they're guilty of something horrible when all they've been trying to do will benefit you, the end user in the long term.
And why does it matter that they're in it for profit? Everyone's in it for the profits; the morally-bankrupt corporations you and I and everyone pay month for internet or electricity are in it for the profits. It's like we keep on arguing on how to judge these faceless companies for the product they produce as if they owe it us all to put out something breath-taking so much it's their fault we are disappointed when we set too high expectations for ourselves. Of course, I think we should have the right to hold an opinion on whatever flavor of the month poster boy easy target but look at the stuff not just nvidia but all those other companies have done before, hardware t&l, multiple pass bump-mapping, parallel shader architecture, n-patch tesselation all of these techs sputtered when they first came to market, due to being closed source and hardware-exclusive or being called pointless cosmetic fluff attempting to lock you in a brand. Look at where these things are now, it's all evolved and become ubiquitous in almost every gpu on the market!
Do you think the people back in the days giving nvidia *** for promoting pixel shaders instead of coming up with a solution for the memory bandwidth bottleneck made them work any faster on the crossbar controller architecture?
No, I think if any nvidia engineer were to read some of the comments made at them today and realize how much more vulgar it's becoming, they'd be so pissed off they'd delay the release of an innovating technology and file it as a submarine patent just to give into the general cacophony of name calling and pejorative attitudes just to spite us. (yes this is a joke, put the safety back on)



I disagree. My statement was made as candidly as possible and within the context of an outsider's perspective, this place is filled with childish bickering and arguments aimed at some sempiternal illusion that someone can "win" and argument if they just have the last word or if they obtain more corroboration from equally unqualified authority. I fail to see how a call for a little more moderation to tone it down a tad is the "opposite of trying to foster discussion", nor do I see how repeatedly ignoring calls for these flame wars to go outside is keeping "the discussion in bounds". Furthermore, judging by how much efforts you took in responding to me rather than cleaning up the place, I detect that you might be more interested in discrediting my presence here due to my pointing this out and for a moderator, I think you're teetering on the wrong side of a very thin line by taking it negatively and responding this way.



Look, I really didn't want to argue but now you sound like you're trying to keep scores on what I'm getting right and wrong because I'm somehow being adversarial for being excited about a brand that's taboo to support at this time. Sure, I got that one wrong; you may quote me on that. But what the hell have I done to deserve being addressed in this sort of judgmental tone? Was I out to dupe poor innocent graphic designers and effects artists to shell out for an nvidia card only to be disappointed that only one of the two programs doesn't support mercury? and as opposed to what? an ati board which supports neither?
Is my previous statement about after-effect completely disjunct that stance? It was an honest mistake; I remembered reading rumors that adobe was working on a cuda enabled video engine and I assumed that they'd apply it to both their video oriented products; what's the date tag on that post disproving that you found? I might not even have had the time to catch up with the explosion of new information that occurred around midnight yesterday.



That's good but right back at you there, you would have lead others to think it were a fair comparison by bringing it up as an argument where it was irrelevant. Opengl acceleration so far has been mostly for cosmetic and interface responsiveness enhancements while the cuda implementations are rumored to increase performance by more than a full order of magnitude. Allowe me to remain unimpressed that opengl enhancements were even mentioned.

I'm pretty sure I'm not pretending to have an accurate map of what others know, nor have I shown the pretense of knowingly offering erroneous statements for the sole reason of shabbily defend some sort of black or white stance but I'm also allowed to use such a low level of cynicism every now and then without being accused of being a witch and confronted with a piece of google/wikipedia phd refutal. Also, I don't have brand loyalty per se, I simply happen to favor whoever supports innovation and all attempts to further branch off today's boring status quo and meander somewhere that'll end up much more interesting. In that way, I do give matrox much credit for pioneering in making cards with multiple ramdacs for better multi-screen support way ahead of the game and 3Dlabs for their hardware frustrum culling engine, oh and don't forget S3 for bringing hardware texture compression, and every company whose engineers have made what computers are today. Yes, seriously dude; you may mock me for that too I guess.

When I replied to you, I honestly didn't mean to start a conversation like this. I was hoping for something more candid, I guess. I've only moderated one very early board in my days and the unwritten rule to me is that people do see in your tag of authority a beacon to indicate the general tone and spirit in which you shepherd your community with. In short, you should be setting an example of civility and order. Being amicable, or neutrally-stanced is generally not a requirement but it tends to attract like-minded and objective people. Here, several people have expressed discontent at the pace, meaningless content and general amount of belligerence in posts looking at the board through ati-colored glasses in this one thread alone yet, not a word was said about it. I don't mean to question the way you do your job here but why are you ignoring them? Then again, I'd be delusional to think I could change anyone's mind here so I guess I'll just drop it...
Responding to posts like this in the same fashion is actually really easy but in the end it fails to make things better due to being an irresistible taunt for the opposite party to respond with even more grandeur under the illusion of merit that is a handful of locals nodding in approval. So, sorry about making it worse for what it's worth. I can take a hint so I'll just go lurk elsewhere.
 
The 5850 has a stock voltage of 1.08v, so yes its a massive increase.

Also, the GTX 470 wasn't overvolted because there is no supporting software yet.

We are still comparing a 140 MHz overclock to the standard 125-150 MHz overclock on the stock 5850. That isn't a low overclock. Also the GTX 470 is clock for clock faster than the 5850, so it may gain more performance though we simply don't know at this point.
 


Wtf? Our fans on our ballasts at work spin at 3600 RPM (100%) for half a decade before they break down, and they are exposed to the elements which is why they corrode. Non-shitty fans will not die faster just by spinning at 100%...
 
First of all you cant compare 100Mhz on 470 with 100Mhz on 5850.

You need to compare percentage, as 100Mhz on 470 is MORE than 100 Mhz on 5850

140Mhz on 470 = 23%
150Mhz on 5850 = 20%

However I agree that 5850 is much more overclockable than 470.

I dont really think you should overvolt the 470, you will end up hitting that heat wall almost imediately.

My 5850 with a mild overvolt (to 5870 voltage level 1.16v) will do 950/1200 with fan on AUTO! and on auto it gets to about 35% and 80ish degrees C, on a normal case (no extra fans just regular big tower)

I am sure I can go over 1Ghz if I overvolt more and increase fan speed. This is reference Sapphire card.

1050/1200 is 45% overclock, to reach these levels a 470 would have to run at near 900Mhz. I really cant see that on a reference card. 800ish may be but 900 impossible.

Not to mention power draw at 900 could be probably around 400W may be more 😀

 
Overclocking a video card is on a card by card basis. PERIOD. I can point to forum posts and newegg reviews for 5770,5850,5870's that were not able to O/C at all. Not one bit, or they got what seemed like memory flickering/artifacts. Whether this is related to the huge reports of grey screening, flickering and vertical stripe crashing I don't know. But tons of cards were RMA'd because of those problems. So now to paint a picture that every card out there can o/c because it runs cooler than card X is b/s.
edit : all IMO
 
 


.... no you don't.... if you double the voltage, you double the voltage.

2E = 2I x R

but it can also be

2E = I x 2R

See the difference (conceptually not mathematically)? The current may not necessarily double at all, and if a graphics card isn't a resistive load (which I'm guessing it isn't) then its impedance can change. An exponential equation needs an exponent, otherwise it isn't exponential.

Assuming power draw on the 400 series cards remains at 300W steady, the "limit" of its connectors draw (I realize it could potentially go above this but we'll stick with the 300W limit for the sake of argument), if the Voltage was increased, you would actually need LESS Current to provide the 300W of power.
 
guys its not that simple.

Same chip on same voltage but ran on 1v will draw LESS power at 2GHz than 3Ghz

driving more Mhz = more power.

Also all reviews investigating best power/performance ratios are NOT the highest overclock. Because for 10% increase in Mhz you get MORE THAN 10% power needed. Thats why he said exponential, its may not be exponential but it increases with higher curve.

So say if you have 470 running at 607 and drawing near 250W peak, then it would draw more than 375W running at 900, may be even 500W.

Just try it with your CPU if you want or find some review and make some analysis.

I have power meter and I can do some analysis also may be this weekend
 
http://www.evga.com/forums/tm.aspx?&m=288625&high=chizow&mpage=1

This guy OC'd his 480 to 800/1600/1848 @ 100% Fan = 83C Max.

This was tested with Furmark.

He said the only reason he couldn't go higher was due to the lack of overvolting ability currently.

With the fan at 80% he got 750/1500/1848 @ 80% Fan = 89C Max.

He also noted this was with his GTX 480 in the bottom of his case at a very bad position for airflow and said he would've likely had much lower temps if his card was actually placed near the side ventallation of his Antec 1200 with air directed at the card.

During gaming sessions he reports that with the fan set to 70-75% his card never goes above 65 Celsius, again, with the card in a non-ideal spot.

He also reports that up until about 80% fan speed, the 480 is the same noise level as a GTX 280. Above 80% the 480 is noticably louder. So he said below 80% isn't really noticable noise wise in a case, but above 80% you will hear it; however, he also states that even above 80%, while noticable, it isn't an obtrusive noise.

Overall you can see, the concerns about temps and fan noise are vastly overstated if you simply manually adjust fan speed.

So without overvolting the card, he can OC 100-120MHz higher than stock he figures with the fan only at 80% while remaining at roughly 90C (less if you have good airflow) during Furmark.

70-75% fan speed seems to be the sweat spot, 80-85% if you want to OC and overvolt and stay under 90-95C.

Trying to find more user reviews >.<
 


Although resistance does change relative to voltage (and heat), and it is theoretically possible that in the case of the GTX480, the resistance could scale PERFECTLY proportionately to voltage, i highly doubt this is the case. Furthermore, any resistance curve other than that perfect one would result in an exponential increase, as the exponent does not need to be 2.

2E = 1.333I x 1.5R is exponential, just not with an exponent of 2.

So only in one, theoretical and highly improbable circumstance (the GPU's resistance scales 1:1 with voltage through its entire curve) does power scale 1:1 with voltage. Although I don't have the tests to prove that this is not the case, i think it is important to note that in every other case, power scales exponentially with voltage. (I am currently ignoring cases where resistance increases MORE than the voltage, which would result in power use increasing more slowly than voltage -- i think hundreds of articles of power-usage in overclocking articles have shown that this is not the case in reality.)
 



Isnt that the way we compare Fermis ability to OC?

Did I use bad words like you ? NO. I only give facts and not emotions and personal notes like you and the other guy lmao childish

You still cant admit the reality that a 5850 OC higher on Air than a 5850? Overvolted or not. At least I have a 5850 to proof that, how about get a 470 and OC it to 900 overvolted on air? Well guess what? You cant right now! I am not responsible for killing the card due to overheat. :lol:

I am sure people here care about 470 Max OC Vs 5850 Max OC on Air stable. 😉

Why not free performance boost? Who doesn't want some? 😀
 
Isnt that the way we compare Fermis ability to OC?

Did I use bad words like you ? NO. I only give facts and not emotions and personal notes like you and the other guy lmao

No all you were doing is spamming about your 5850 and babbling about its o/c performance vs its heat output at % fan speed blah blah blah blah oh and that was with 10.2 not 10.3 and more random spamming how no one understands these things but you etc. lmao.
5850 is slower than a GTX 470 is that emotionless enough for you.
 


lol? This thread is about Fermis, hence we can talk about OC abitlity of Fermi on Air , right?
I made a comparasion of 2 cards in the same league, 5850 and 470. Hence, with enough knowledge, a 5850 OCd will surpass a 470 OCd while being cheaper and has been out for 8 months 😉

Thank you for calling proofs as spamming :lol:

My point is overvolting 470 & 480 is dangerous on Air. End Of Story
 



wow you really love to define the rules dont you. I mean when something is not in your likes you ban it so it cant hurt your feelings 😀

Relax mate, nobody is questioning that stock 470 beats stock 5850. You can sleep tonight
 
Open you own thread up, and tell the world about your exploits there ? Justify your 5850 purchase in there. Then you can make your endless psu and microwave jokes. Instead of here where people are trying to learn about the New Nvidia cards.
 


I'm sorry , but take a look at this thread. Official ATI HD 5970 reviews and discussion . http://www.tomshardware.com/forum/274507-33-official-5970-reviews-discussion Its even pointed out , how that thread is free of trolling. Eli got a big kick out of it. He personally has turned this thread in to a mess.
 
lol its true, 480 SLI will use near the energy of a microwave, isnt that talking about Fermi reality? I just learned that Fermi ability to OC on air within safe margin is bad.
Tesslation & GPGPU &3D for deep pockets is its main selling point.

Why a mess? Because I stated True facts ? I dont see why it is bad?
 
Isnt that the way we compare Fermis ability to OC?

Did I use bad words like you ? NO. I only give facts and not emotions and personal notes like you and the other guy lmao childish

You still cant admit the reality that a 5850 OC higher on Air than a 5850? Overvolted or not. At least I have a 5850 to proof that, how about get a 470 and OC it to 900 overvolted on air? Well guess what? You cant right now! I am not responsible for killing the card due to overheat. :lol:

I am sure people here care about 470 Max OC Vs 5850 Max OC on Air stable. 😉

Why not free performance boost? Who doesn't want some? 😀

I just linked the 480 OCing fine on air 😛

Not to mention that a 400 series does more work per clock than a 5800 series, so just because it might not overclock as much, doesn't mean a 5850 will beat a 470 in performance when comparing overclocked stock cards.

It is all very subjective as well. Are we talking purely overclocking here as high as it can go, or are we saying an OC that is stable for everyday use and doesn't annoy the user with noise/heat. Because a 470 can probably make it to 900MHz or past on air when they finally release an overvolt tool, but then you'd likely have the fan at 100% to keep the temps below 100C. Or you might even make 1GHz on air but you'll push 100C on 100% fan speed.

In my opinion comparing OCing results isn't very indicative of the card's performance anyways, because there are so many varying degrees of OCing depending on how much effort/time the owner wants to sink into OCing that card. Comparing stock is really what is important. At stock, a 470 beats a 5850, and if you wanted to hardcore OC both of them, I'm sure at both of their MAXIMUM cable OC's (we're talking the point where you need ln2 and the card dies still) a 470 would still beat a 5850. Most consumers that buy a 470 or 5850 will probably not hardcore OC them is my guess, I could be wrong on that one though, so stock is really in the grand scheme of things what matters.

If you really want to have a contest though, when I get my two 470's (finally decided on them 470's, just gotta order them now) and SLI em and OC them to hell and back, you can do the same with a couple of 5850s and see who wins 😛

The 400 series are not BAD cards like everyone wants to make them sound, they aren't everything they were hyped to be, obviously, but they aren't garbage either. The 5800's are also very good cards, I was happy to get to test a couple of them as well. Each fit a different user segment.
 
I am talking about what you can reach right now not hoping in the future, not on another planet.

I am not saying their bad, if you get a good deal on them why not, or have water why not? Air is what most users use 90%+.

1Ghz Air on a 470 and on SLI? Thats overstretched lmao Wanna try that if over voltage programs pops out? Not responsible for damage.

Wake me up when a 470 reachs 21K Vantage GPU score on Air.
 
My personal honest opinion is that 470 is probably suitable for some people. Its more future proof, Drivers are not mature yet so more improvements can be expected (while on ATI I doubt). Dx11 features are much better so the card will probably hold you more than a 5850.

Its only slightly more expensive currently (if you can find one)

The more heat/power is no issue for lot of people so as I said depending on preferences there will be lot of buyers :)
 
I am talking about what you can reach right now not in the future, not on another planet.

I am not saying their bad, if you get a good deal on them why not, or have water why not? Air is what most users use 90%+.

1Ghz Air on a 470 and on SLI? Thats overstretched lmao Wanna try that if over voltage programs pops out? Not responsible for damage.

Wake me up when a 470 reachs 21K Vantage GPU score on Air.

You keep bringing up another planet ,and water cooling. When you know the person your quoting is not referencing that. Your the one flashing your bios and trying IMO crazy tricks to get more performance out of your card, your the one heading for a shiny bricked gpu.

LOL, I'm not going to debate your ruined card with you, you did that song and dance with izzy a page back. Where you babble about stock o/c cards and their measly 25mhz overstock difference. When you bake your card , tell them that you o/c and raised the voltage with their approved software, and tell us the results.
 



Wanna bet? My OCd GPU will last longer than your stock 470 😀 --> Heat generated by electronic circuitry must be dissipated to prevent immediate failure and improve long term reliability.

So you are thinking that AMD is providing tools to brick our cards without setting safe limits? Again we are on earth wher Companys doesnt want their cards bricked. To brick the card notty, you have to hardmod it to go above safe voltages. So Gigabyte are crazy to provide a factory overvolted card?

Back to topic: When will we be seeing OverVotlage program for Fermi? EVGA precision?