TomsHardware needs to fix their GPU table.

[MushBrain]

Now this is of course pertaining mostly to Kepler and Fermi.

I have benchmarked in two games the 580 vs the 670.

EVGA GeForce GTX 580 FTW HC2 on air cooling (custom with a factory OC of 850mhz with 4200mhz memory clock.

vs

EVGA GeForce GTX 670 SC reference with a boost clock of 1176mhz and 7000mhz memory clock.


I tested with settings that will not butcher the video memory of the 580 since it has only 1.5GB Vram.

Thief at 1680x1050 with max settings but shadows on medium, SSAA off.

580 result.

33.7 minimum
73.2 max
54 average

670 result.

39.4 minimum
98.5 maximum
70.0 average

Shadow of Mordor at 1680x1050 with medium textures, order transparency off, everything else maxed.

580 result.

50.12 minimum
115.01 max
75.69 average

670 result.

49.37 minimum
140.52 maximum
96.58 average



Tested with driver revision 378.49 under Windows 10 x64 Pro.

intel XEON E5640 4.2ghz
8GB 1600mhz CL10 memory
ASUS P6X58D-E motherboard
Silverstone ST1000-P 80 plus silver PSU


PLEASE UPDATE your charts, the 670 is on par if not better than the GTX 960.

https://www.youtube.com/watch?v=Eg4MaRVc_9M

 

[damric]

No one cares about crappy Kepler, Fermi, or such low resolutions, sorry. Forum marketplaces and Ebay are flooded with this stuff, hard to even give away anything that isn't at least Maxwell or GCN.

 

[MushBrain]

No one cares about crappy Kepler, Fermi, or such low resolutions, sorry. Forum marketplaces and Ebay are flooded with this stuff, hard to even give away anything that isn't at least Maxwell or GCN.

\What an amazingly helpful post, do you reply like this often?

 

[COLGeek]

Lets keep it civil guys. Thanks.