Quote Originally Posted by Svnth View Post
Given all the questions about noise levels, and the uncertainty of [H] videos, it would be interesting if you would do a test comparing all the GPUs you mentioned, and actually test actual Decibels from, say 2 locations, across a couple games. Or maybe just do it for a few of the cards (the 4xx and the 5xxx)?

Would you be willing to do that? I'm sure people would like to see the data for actual games, not a synthetic test like Furmark (especially if AMD throttles this application, which alters results).
I already have a Fluke dB meter on the way as well as a heatsink I am confident I will be able to run passively on my i5 750 system May as well put this to bed once and for all and back up my claims with real numbers.


Quote Originally Posted by zerazax View Post
Thanks for the numbers!

No clue on the HWC thing... its 5-6 points higher than the other ones it seems

Are there enough 2560 x 1600 data points to make a compilation of that as well?
Quote Originally Posted by Funky View Post
Yeah.
Their numbers are way off than everyone else but there might be an explanation for it.

HWC is the only review site from that charts which is running i7 Quad at 4Ghz with turbo on.

Rest all are 3.8ghz or less. ( most are way less than 3.8G )

Firingsquad is the only one which is running 980X but only at 3.3Ghz.

Someone needs to do analysis of how Fermi scales with increase in CPU freq.
So I dont know, that might be the reason for those numbers.
Quote Originally Posted by annihilat0r View Post
The 15% number is pretty much established now I think... Anandtech mentioned 15% in their review too. This difference rises in minimum FPS, though, might affect some people's decisions. Obv this means a lower maximum FPS difference, but who cares about max FPS anyway.

btw majord (does this stand for "major chord"? ) about the HWC review. Your numbers show 22% over 5870 for 1920 with 4AA, but HWC's own compilation shows 18%:


There is a reason for this.

If you look at the numbers on that GREAT chart that was posted, there were a number of games that weren't included. Unfortunately, including two Far Cry 2 results and not including the AvP, L4D 2 and some other results meant the percentage is shifted WAY too far in NVIDIA's favor.

Also, we are using a highly overclocked system to eliminate as much CPU bottleneck (a major problem at 1680) as possible.