Did you click 'apply' after setting the custom fan curve? I'd manually set the fan speed to 100% with the slider far left, hit apply, then run the benchmark again.
Your PCIe bus is set to 1.1. This may be restricting the 980's performance slightly in some games/benchmarks. With a 2500k CPU you should be good for PCIe 2.0. The setting will be in your BIOS somewhere.
Well - it may ramp up - there's a button somewhere in GPU-Z that brings up a 3D render isn't there - click that and then see what it says? I've not got GPU-Z here and can't remember what to click... Well - looks like it's getting too hot. What do the benchmark results look like for the Metro/Bioshock/whatever it was, with the fan at full tilt?
The 980 is configured to reach 80°C and then stay there using a combination of fan speed and/or clockspeed reduction. If your 980 is starved of airflow, it seems likely that it's throttling. For an apples to apples comparison I'll run the Metro benchmark later using the same settings, with my clockspeeds at reference, and post the result.
This. Although the default fan profile, by my understanding, favours a drop in clock speed over added noise. I wonder if, like our old pal Shirty says, your card is starved of cool air?
Otherwise, the TIM might be: a) badly applied; or b) not great, and therefore a cooler re-seat might do the trick?
Sorry to be so ignorant. I actually found it and it was doing it's thing, whatever it was so I banged Esc. What was I looking for?
Ran Metro: Last Light again and still rubbish. Options: Resolution: 1920 x 1080; DirectX: DirectX 11; Quality: High; Texture filtering: AF 16X; Advanced PhysX: Disabled; Tesselation: High; Motion Blur: Normal; SSAA: OFF; Total Frames: 6092, Total Time: 171.1806 sec •Average Framerate: 35.62 •Max. Framerate: 84.36 (Frame: 27) •Min. Framerate: 8.61 (Frame: 4450)
I've just run that benchmark on my setup which is obviously not the same as yours but according to this my two cards should basically perform the same as your one card on Metro:LL. METRO LL BENCHMARK RESULTS 06/05/2015 17:03:21 Preset 0 Options: Resolution: 2560 x 1440; DirectX: DirectX 11; Quality: High; Texture filtering: AF 4X; Advanced PhysX: Disabled; Tesselation: Off; Motion Blur: Low; SSAA: ON; Run 0 (D6 ) Total Frames: 11158, Total Time: 171.2481 sec Average Framerate: 65.19 Max. Framerate: 150.06 (Frame: 9787) Min. Framerate: 32.58 (Frame: 7079) I reckon it's a combination of the PCIe Bus thing mentioned above and temperatures keeping your performance down. If you open the EVGA Precision graph does the GPU clock stay constant during the benchmark or does it drop down lower?
When running the render test, have a look to see if the PCIe bus interface as reported by GPUZ changes from "PCIE 2.0 x16 @ x16 1.1" to "PCIE 2.0 x16 @ x16 2.0".