The GeForce GTX 480 Update: 3-Way SLI, 3D Vision, And Noise

Page 6 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.

spoofedpacket

Distinguished
Jun 13, 2009
201
0
18,690
[citation][nom]Mottamort[/nom]I can't help myself but say this : try taking your own advice much?Oh and something maxing out at 92C which ISN'T causing alarm (if not for your card, then for the other half-melted components in your system) is rather strange...I don't care what NVidia say about it, 90C+ scares the crap out of me. I wouldn't put that card NEAR my system...[/citation]

The heat goes right out the back. I guess you also will not be running a 5970, or think they are scarey, since they are a whopping 4C cooler under load.

I've got a GTX295 that has been peaking at around 90C in my system for almost a year and it has absolutely no impact on the temps of other devices since all that heat blows right out of the case.

In total, I have two case fans, Noctuas, which push/pull air through the system and my CPU is overclocked from 2.66 to 4ghz. It never gets over 70C even with prime95 running.

I'm not sure why so many completely misinformed people manage to still be on this site, happily spewing their opinions in the face of facts and common sense. It is just incredible. Then again, I don't exactly see any Tom's articles resolving myths of heat and how to cool a PC the proper way. People just tend to form some group opinion, then it gets cycled through these comments as "facts" over and over. Really sad, actually. Since you'd think someone with enough brains to overclock a CPU without a hand-holding should understand things like airflow on some level.
 

derek85

Distinguished
Feb 4, 2008
44
0
18,530
This is not a fair comparison.... obviously the HD5870s you have were running out frame buffer in certain cases, while you calculate the dollar to fps ratio based on a resolution that is particularly disadvantageous to the 1GB ram on the HD5870s...

Can't you compare at 1920? Or use the HD5870 E6 edition?
 

lsorense

Distinguished
Jan 9, 2007
14
0
18,510
Looking at just the cost of the video cards for performance is false economy. After all if one set of cards is $800 and the other is $1000 but you need another $1000 of hardware to run them, then the cost to compare is in fact $1800 versus $2000. You can't just look at the performance and cost at the individual component level if you want the best performance for your money.
 
Good article! I'm currently happy with a 5850 since I only run at 1920x1080. Might Xfire them later... or upgrade to something else... who knows! But for the money, I think I'd prefer to save a lot of cash on the 5870s than the GTX480, unless of course I'm running several monitors for the super high resolution, in which case I'd probably have that extra cash to blow.
 
[citation][nom]cangelini[/nom]The Bokeh filter and GPU Water were both turned on for the GTX 480s (which is why the analysis says we can't compare them). I'll clarify on the setup page.[/citation]

Cool yeah, it was just confusing, especially since without the GPU-accelerated water others were getting similar results so the mismatch confused me as to the comments.

ATI doesn't have a problem in a case or on an open-air bench. That's why they weren't contacted. Nvidia does have a problem on an open-air bench. That's why they contacted *me.* It's notable that running these cards in one of the "approved" cases helps alleviate the acoustic issue. And I do believe that someone interested in buying one of these cards would want that information, no? I'm not sure whether you're suggesting I altered my methodology to favor Nvidia, but that's not the case at all.

NAh, Chris it's not a jab at you it's a jab at nV requiring this. I'm sure that if ATi made the equal effort on the 'case design' end of the equation instead of the core design end, they might request a specific case to improve their sound and cooling results, but it just makes me laugh at the certified case. I have no doubt people truly interested in the GTX480 will make efforts and pay expenses to get the experience they want, it's just that these retests and considerations were not made for other such loud & hot cards in the past, nor for things like the intel Prescotts which trends were an impetus for the BTX design prior to launch curing the devlopment phase IIRC.

This is not a comment on your testing, just that we have to have so many reviewers return to testing heat and sound in a case to show it in a positive light, yet we have so little non-traditional testing in other areas that would be more important, since as I mentioned heat & sound are likely non-issues for those willing to fork out the dough at launch anyways. I'd rather see the time/effort spent doing these 'he's not fat he's my brother' re-tests to soften the initial reviews and spend the time on something new and insightful into the architecture, or feature-set that's all.

Which sorta speak to the last part of your comment, to me the value of this data is less than something not being done by others, it's definitely nice to have, but I would value more performance test in CS5 (both OpenGL acceleration and CUDA [which isn't available for the GTX480 yet but is for the 285 so OGL 480 & HD5870/5970 vs CUDA 285]), or else HD playback comparisons etc, this is my issue with the trend.

Other than that it just bugs me whenever a company says that one of their specialized setups needs to be used, but the same consideration isn't given to the other company. While AMD doesn't require this it just would've been nice to get their input on the issue while complying with nV's desire to test under their special situation.

Nothing against you mang, I know you, Don, Paul and Thomas (sounds like members of a UK rock group) aren't vendor specific, it's all good. [:jaydeejohn:5]
 

Cheshyr

Distinguished
Apr 7, 2006
11
0
18,510
Thanks for the article, Chris. It's good to see the extra due diligence on the latest Geforce.

My only comment is on the disregarding of the Metro 2033 results. I agree that they're a statistical outlier, and shouldn't be part of the general comparison... but it might be worth noting that 3xSLI 480 is the only configuration that is capable of 30fps with these settings. That might be a given, but it has a little value of it's own.
 

Cheshyr

Distinguished
Apr 7, 2006
11
0
18,510
To those concerned about 'overloading your power supply': What really matters is the amperage available on your 12V rail.The published wattage of a power supply isn't truly indicitive of that critical number, but can give you a rough approximation. Good power supplies might have more power allocated to the 12V, while bad ones might have less. If you want to do true power supply matching, we would need exact 12V rail amperage values for the video card (and the rest of the system), and then you'd need to examine the appropriate 12V specs for each power supply. Impractical for most people, so manufacturers recommend wattages that should cover system needs regardless of power supply quality.

Chris used a quality power supply that must have had a pretty beefy 12V rail amperage spec. The call-out regarding a quality supply is appreciated.
 

backin5

Distinguished
Oct 23, 2009
146
0
18,680
[citation][nom]dragonsqrrl[/nom]Yes... you "somehow missed them when reading". Despite popular belief on this thread the author does mention the power consumption of the 3-way SLI setup, try re-reading... carefully.[/citation]

If you had read my question carefully, you would have noticed I asked about the temperature under load, not power consumption. And btw, I checked for the third time. I didn't miss it - it's just not there, only idle temperatures are graphed.
 

jbakerlent

Distinguished
[citation][nom]Cheshyr[/nom]To those concerned about 'overloading your power supply': What really matters is the amperage available on your 12V rail.The published wattage of a power supply isn't truly indicitive of that critical number, but can give you a rough approximation. Good power supplies might have more power allocated to the 12V, while bad ones might have less. If you want to do true power supply matching, we would need exact 12V rail amperage values for the video card (and the rest of the system), and then you'd need to examine the appropriate 12V specs for each power supply. Impractical for most people, so manufacturers recommend wattages that should cover system needs regardless of power supply quality.Chris used a quality power supply that must have had a pretty beefy 12V rail amperage spec. The call-out regarding a quality supply is appreciated.[/citation]

Yep, looks like it's got 696W across three +12V rails.
 

Iluv2raceit

Distinguished
May 5, 2010
23
0
18,510
[citation][nom]nforce4max[/nom]I am unimpressed and like the 9600gt sli user I am going to stick with my two 9800gt 1gb setup.[/citation]

You can't be serious!? A single GTX285 would own your 9600GT SLI setup, let alone a single GTX480 would own the GTX285! Oh, and don't forget that your 9600GTs can't play games in DX11 and you can't play games at their highest settings. Time to upgrade for you!! Nuff said!
 

Iluv2raceit

Distinguished
May 5, 2010
23
0
18,510
For those worried about which power supply to use with the GTX480s, here are 4 good options to choose from:

BFG EX-1200
PC Power and Cooling Turbocool 1200
Antec TPQ-1200
Antec TPQ-1200 OC

I am using the BFG EX-1200 to run three GTX480s in 3-way SLI, an Intel 980X CPU overclocked at 4GHz,and 6GB of OCZ Blade 2000MHz DDR3 memory @ 7-8-7-20 1T and all runs great! I will be honest with you though -the three GTX480s do get loud when gaming, so make sure you have excellent ventilation or use water cooling for you cards! I ended up switching from my Antec 900 mid-tower case to a more spacious and better ventilated Silverstone Raven RV-01 90-degree full-tower case and I LOVE IT! Good luck with your next build and make sure to get at least one GTX480... You won't regret it ;-)
 
G

Guest

Guest
How about 5970 + 5870? Or even 5970 + 5850, which would pretty much match two 480's in price. Two 5970's would be less than three 480's in price, and be something a living human could tolerate in terms of noise and power consumption.

 

reix2x

Distinguished
May 18, 2010
2
0
18,510
Why u test only nvidia optimised games? they most test more games.. i think the ati's 5000 series have better scaling... nvidia makes a great job with this 480, but the high price its not great XD i think "nvidiaGTX480*1=ATI5870*1.333"
 

dif4us

Distinguished
Apr 11, 2010
9
0
18,510
Try crunching 20 gigs of 2048 x 2048 Google images to fly over with X-plane(days with no cuda vs hours with cuda)Or convert a two hour movie for your Iphone (8 hours no cuda vs 20 miutes with cuda) No A.A. for a.t.i. get real. Crap drivers.(getting better though) Poor linux support. Physics, yes some people do want it. At these price points I want bloody eye candy. Tessalation whether you think it's hype or not is the only way at this point in time to bring incredible real world detail without massive overhead and nvidia does it better period. For fps both camps are more then adequate. For everything else to me the extra money is worth it. Flame away young flamers flame away.
 

WarDad

Distinguished
May 22, 2010
20
0
18,510
$100 price difference, a few dollars more electic bill.
Hah! Have you priced the 1000w and higher PSUs? What about the reliability? Most of these big boys have spectacular blow ups. Even the good brands only get 4 out of 5 eggs.
A GTX480 3-way SLI system better have a dedicated wall socket. 15 amp circuit breakers are common, giving you a 1800w ceiling, less during brownouts. In some houses one breaker feeds 3 sockets, so you may need to cover the other 2, or pay for an electrician.
 

Casper42

Distinguished
Apr 24, 2007
61
2
18,640
I would have really liked to have seen the lower cost part included in this assessment. I am just talking about the 5850 ($75-100 cheaper) and GTX 470 ($150 cheaper).
Both support SLI/CF and often you get a better $/FPS Ratio with the cheaper cards.

Just like I don't need a 980X and am waiting for his little brother to emerge, I don't particularly need a 480/5870 and can make due with something a hair slower if it has a better Cost/Benefit ratio going for it.
 

BIGthingy47

Distinguished
Sep 26, 2009
7
0
18,510
To some of the people asking about the 5890 on this page, I believe it has finally been released. If you watch Trubritar's videos on youtube he has just recently done a Crossfire X Benchmark with the 5890's for Resident Evil 5.
 

jeff77789

Distinguished
Jun 24, 2009
198
0
18,690
the reason that the radeons are losing like crap its cuz the Intel boards are ADAPTED to use crossfire...........if you used the 790FX chipset and an AMD processor, you would get better results
 
Status
Not open for further replies.