1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

hidden performance in gtx770 capped by driver?

Discussion in 'Overclocking & Cooling' started by DarkStalker, Mar 28, 2014.

  1. DarkStalker

    DarkStalker

    Joined:
    Nov 30, 2007
    Messages:
    28 (0.01/day)
    Thanks Received:
    0
    hi all!

    for months i've been usin msi geforce gtx 770 gaming oc 4gb version... awesome card factory oced to 1085 for gpu then successfully oced to 1185 on default voltage(mem on 8000mhz>don't plan mem anymore to oc it)...but...

    last night i was a little bored and decided to oc further and ofcourse bench it little with 3dmark11.
    on 1185 gpu(and default voltage) i achieve approx. 11670 p score...
    normally then on 1210 i hit (voltage +5) p score 11850...
    then in it strucked me when in afterburner i increased voltage to +9, power limit left on default temp setting 79c and increased gpu freq.to 1230, temps max 70c and when it finished pscore was 13470!!??
    I WAS STUNNED, sadly i noticed that driver automatically restored default settings after finishing benchmark but i never expected p score increase from 1620 points!? rather ~250points!

    could it be that nvidia intentionaly capps card through hidden driver setting or bios maybe?
  2. ISI300

    ISI300

    Joined:
    Jan 31, 2012
    Messages:
    83 (0.09/day)
    Thanks Received:
    4
    Location:
    Turkey, Istanbul
    It would help if posted screenies, But you shouldn't see that much variance from run to run.
  3. Kaynar

    Joined:
    Jan 18, 2012
    Messages:
    615 (0.67/day)
    Thanks Received:
    161
    First of all, when I run 3dmark11 my first run is always about 5% lower than the next ones. Then, it seems you got a real overclocker there! nice! But your numbers seem logical since my older 7970 did P11500 at 1100mhz with a i7 930 cpu...

    Consider that my stock titan and stock 4930k are doing P13500 so u got urself and very nice OC there. (Though with just a little oc I go at 15000).
  4. DarkStalker

    DarkStalker

    Joined:
    Nov 30, 2007
    Messages:
    28 (0.01/day)
    Thanks Received:
    0
    well Kaynar i guess so but to see such increase in points with raising only another 20mhz on gpu
    seems like something not transparent...
    ill try today a rerun to see if i can recreate the score...and post some screenies
  5. Vario

    Vario

    Joined:
    Oct 21, 2005
    Messages:
    2,821 (0.88/day)
    Thanks Received:
    887
    Wow, I run mine at the default 1254/1800 and it gets about 10,850 with stock i7 (3.7), stock 2x8 ram(1333), stock 770 2GB(1254/1800).
  6. TheBrainyOne

    TheBrainyOne

    Joined:
    Mar 30, 2014
    Messages:
    105 (0.91/day)
    Thanks Received:
    91
    Location:
    India
    Use GPU-Z and check for Clock Speeds while you are running the benchmark. There is a (very small) possibility that the extra voltage allowed you to sustain higher clock speeds.
    Elaborate please.

    EDIT: You can also use Afterburner to monitor clock speeds.
    blanarahul says thanks.
  7. DarkStalker

    DarkStalker

    Joined:
    Nov 30, 2007
    Messages:
    28 (0.01/day)
    Thanks Received:
    0
    well guys i forgot to come back here(been busy) but NO, no matter what setting i used i couldnt get that score again...
    lol i do monitor clocks in afterbuner and gpu-z too @TheBrainyOne
  8. Vario

    Vario

    Joined:
    Oct 21, 2005
    Messages:
    2,821 (0.88/day)
    Thanks Received:
    887
    Maybe it was a glitch in the 3dmark software?
  9. kn00tcn

    kn00tcn

    Joined:
    Feb 9, 2009
    Messages:
    654 (0.33/day)
    Thanks Received:
    116
    Location:
    Toronto
    hold on... why should anyone care about an arbitrary score? the real world result is min fps, avg fps, max fps, even better would be a line graph that charts fps over time

    why is the score useless? let's say it does some calculation based on min fps, & you just happened to be 30fps, not hitting a stutter that normally dips to 15fps, this completely changes the score even if avg & max fps are identical

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)

Share This Page