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

Drastic decline in performance.

Discussion in 'Graphics Cards' started by Aksh_47, Oct 26, 2012.

  1. brandonwh64

    brandonwh64 Addicted to Bacon and StarCrunches!!!

    Joined:
    Sep 6, 2009
    Messages:
    18,529 (10.23/day)
    Thanks Received:
    6,042
    Location:
    Chatsworth, GA
    Open CPUZ then run fur mark and see if the CPU rises when the GPU does.
    Aquinus says thanks.
    Crunching for Team TPU
  2. Aquinus

    Aquinus Resident Wat-man

    Joined:
    Jan 28, 2012
    Messages:
    6,172 (6.58/day)
    Thanks Received:
    2,023
    Location:
    Concord, NH
    1: I would also go into your power settings and set it to "High performance" and try again. It is set to power saver, that could impact performance to favor power consumption, not speed.
  3. Aksh_47 New Member

    Joined:
    Apr 24, 2012
    Messages:
    117 (0.14/day)
    Thanks Received:
    6
    Location:
    Goa, India
    Umm so here it is while running furmark.

    [​IMG]
  4. Aksh_47 New Member

    Joined:
    Apr 24, 2012
    Messages:
    117 (0.14/day)
    Thanks Received:
    6
    Location:
    Goa, India
    I did change the power settings in the Nvidia control pannel to 'prefer maximum performance'. Do i need to change any other settings?
  5. de.das.dude

    de.das.dude Pro Indian Modder

    Joined:
    Jun 13, 2010
    Messages:
    7,561 (4.94/day)
    Thanks Received:
    1,966
    yes go to your system control panel -> battery -> and select high performance.


    also can you run furmark for 15 minutes and show the graph?

    the first furmark you posted shows that your card was running at 74C! and the graph was no way near leveling off. might be a heating issue here.

    Also run GPU-Z in the background and go to sensor tab, and have the logging box checked. then report back with both the furmark screen shot, and copy paste from the log file of the GPU-Z


    not sure if its a heating issue, but it can be. 74C isnt that much, but then again its not normal to hit 74C so fast. my GTS450 OC2 hits 80C at max. something isnt right, the thermal paste may not have been applied properly.
    this is what my gts450 OC gives.

    [​IMG]
    Last edited: Oct 29, 2012
  6. Aksh_47 New Member

    Joined:
    Apr 24, 2012
    Messages:
    117 (0.14/day)
    Thanks Received:
    6
    Location:
    Goa, India
    Yes the temps do rise fast but then its constant.. around 75C.
  7. cdawall where the hell are my stars

    Joined:
    Jul 23, 2006
    Messages:
    20,668 (7.00/day)
    Thanks Received:
    2,980
    Location:
    some AF base
    I doubt its a powersupply issue. Sounds like the driver is playing stupid if you can try a fresh install if not try a clean driver install. Remove in safe mode and delete everything nvidia related. Install in normal windows and see how that performs.
  8. Aksh_47 New Member

    Joined:
    Apr 24, 2012
    Messages:
    117 (0.14/day)
    Thanks Received:
    6
    Location:
    Goa, India
    Also, is the score in 3D Mark 11 normal for my CPU? If i can remember, last time i had taken my rig apart, although there was thermal paste on the processor, it was dried out.. could CPU overheating be the problem?
  9. STCNE

    STCNE New Member

    Joined:
    Sep 1, 2010
    Messages:
    242 (0.17/day)
    Thanks Received:
    42
    Location:
    Connecticut USA
    Does Furmark stress the CPU as well as the GPU? If it's a power issue it probably wouldn't act up with only the GPU under stress. You should try running the intelburntest and Furmark together.

    At the very least the burntest will let you know if there is anything wrong with your CPU or RAM so you can narrow down the cause of the problem.

    You also might want to download your HDD manufacturer's testing software and run a S.M.A.R.T. scan to make sure those are working ok.

    If it's a heat issue with the GPU, it's most likely with the VRMs instead of the core. You could try locking the GPU fan to 100% speed with afterburner and see if that helps while you're gaming. I ran into a similar problem overclocking a GTX570 I had a year ago. The core temps were fine but it was unstable on lower fan speeds.
  10. Aksh_47 New Member

    Joined:
    Apr 24, 2012
    Messages:
    117 (0.14/day)
    Thanks Received:
    6
    Location:
    Goa, India
    The GPU does not overheat.. it stays under 70C always.. max was 75C.
  11. Pehla

    Pehla

    Joined:
    Mar 29, 2012
    Messages:
    315 (0.36/day)
    Thanks Received:
    70
    Location:
    brčko dc/bosnia and herzegovina
    not much..,but its something..,u can use afterburner to downclock gpu a litle bit so it would need less power..
    but only a litle! then reun benches..
  12. cdawall where the hell are my stars

    Joined:
    Jul 23, 2006
    Messages:
    20,668 (7.00/day)
    Thanks Received:
    2,980
    Location:
    some AF base
    The GPU is not showing signs of a weakened PSU.
    [​IMG]

    If you look closely at this image you will notice core clock doesn't change. If the PSU was the culprit first of all it doesn't randomly loose performance in games going from a normal 560 to a 560 AMP.

    [​IMG]

    Your 3DMark however is abysmal. It is calling out the CPU not the GPU. Try turning of C1, EIST, C3/C6 etc in the BIOS so that it forces it to run 3ghz the whole time.

    Just FYI since everyone seems to think to run a single GPU you need a 650W unit

    [​IMG]

    That's a rough guesstimate of his usage I feel like his name brand Corsair 430W can output 300W peak.
    Last edited: Oct 29, 2012
  13. Aksh_47 New Member

    Joined:
    Apr 24, 2012
    Messages:
    117 (0.14/day)
    Thanks Received:
    6
    Location:
    Goa, India
    Okay so narrowing down, GPU = No prob, PSU = No prob, CPU = Prob?

    Also, what benches, apps, tweaking, can i use to solve what ever problem the CPU is having?
  14. TC-man

    TC-man

    Joined:
    May 25, 2011
    Messages:
    348 (0.29/day)
    Thanks Received:
    117
    Location:
    Netherlands
    Hi,

    you may need to verify the pins in the 1156 socket/LGA if there's one or several being bent. As for testing the cpu itself, check on the temperature and use CPU stress/burn tools like OCCT or prime95 while using temperature monitoring softwares to check on the temperature (OCCT has them built-in), e.g. Coretemp or Realtemp etc..

    Also you mentioned something about dried out thermal paste/compound? Perhaps you need to re-apply some fresh thermal paste and see if it solves the problem you are having new, assuming the CPU is throttling because of overheating, hence the lower performance overall.
  15. basco

    Joined:
    Jun 29, 2009
    Messages:
    92 (0.05/day)
    Thanks Received:
    21
    Location:
    austria
    crapdaddy said your gpu is throttling and i think he is right.

    look at your furmark pic it says throttling minus 320mhz???

    maybe a pic from afterburner showing your gpu-mhz during gameplay would be nice.

    you could try to lower gpu volt in afterburner and bench again.1,087 is ok but on the higher side for gtx560.
  16. cdawall where the hell are my stars

    Joined:
    Jul 23, 2006
    Messages:
    20,668 (7.00/day)
    Thanks Received:
    2,980
    Location:
    some AF base
    Gpuz shows zero throttling in game.
  17. basco

    Joined:
    Jun 29, 2009
    Messages:
    92 (0.05/day)
    Thanks Received:
    21
    Location:
    austria
    ok sorry overlooked.
    but whats up in furmark then-just wrong reading?
    in zotac forum there are lot of different probs with amp card.

    but maybe this sample is only throttling at specific degree like over 70°
    in games you never exceeded this temp but in furmark you did.
    maybe pic from gpuz during furmark?

    maybe pic with this:
    http://www.ozone3d.net/gpushark/
    Last edited: Oct 29, 2012
  18. Crap Daddy

    Crap Daddy

    Joined:
    Oct 29, 2010
    Messages:
    2,739 (1.97/day)
    Thanks Received:
    1,044
    That was GPU-z render test not a game. Given the low CPU score in 3Dmark I'd say it's the CPU. Run a prime test on the CPU while keeping the CPU-z open and see if it does not throttle down.
  19. STCNE

    STCNE New Member

    Joined:
    Sep 1, 2010
    Messages:
    242 (0.17/day)
    Thanks Received:
    42
    Location:
    Connecticut USA
    The temp sensor only shows the temp of the core. The card will still have problems if the VRMs or VRAM overheat and there isn't any way to check those temperatures that I know of. I don't think that possibility should be overlooked. It happened to me with my old GTX570, the core temp never passed 68C but as long as I was overclocking and had the fan set on a lower speed it was unstable. I can't think of any other problems that could have been solved by turning the fan speed up. The best way I found to test graphics card stability was Unigine Heaven. If anything with my card was unstable I'd get weird globs of colored light or flickering, have the GPU drivers reset, or it would crash the computer all together.

    Honestly I don't think it's a power issue. Corsair make good PSUs and they tend to be conservative in their ratings if anything.

    If the problem is with the CPU or RAM, running a few of the popular CPU benchmarks like the intelburntest of Prime95 will show it. You can use coretemp to watch the temperature of the CPU while testing to see if there is a problem with the CPU cooling.
  20. eidairaman1

    eidairaman1

    Joined:
    Jul 2, 2007
    Messages:
    12,214 (4.68/day)
    Thanks Received:
    1,405
    what room temperatures are you encountering? if the Air Temperature is 80+ Fahrenheit Stock CPU fans tend to rev higher? So it basically means keep temps around 75 in a Room where the computer is.

    Id Say test the video card in another machine.
  21. Aksh_47 New Member

    Joined:
    Apr 24, 2012
    Messages:
    117 (0.14/day)
    Thanks Received:
    6
    Location:
    Goa, India
    The PC is in an air conditioned room with a temp of about 20C = ~70 Fahrenheit.. I will run the Stress tests for CPU, but could anyone tell me what exactly is throttling and is it good or bad?
  22. de.das.dude

    de.das.dude Pro Indian Modder

    Joined:
    Jun 13, 2010
    Messages:
    7,561 (4.94/day)
    Thanks Received:
    1,966
    well looks like the CPU is fucked. throttling is when the cpu gets too hot it lowers its clock and voltage to reduce temps.
  23. Aksh_47 New Member

    Joined:
    Apr 24, 2012
    Messages:
    117 (0.14/day)
    Thanks Received:
    6
    Location:
    Goa, India
    woah.. its the CPU overheating i guess.. temps rising like crazy, within about 10 secs of the torture test using prime95.

    [​IMG]
  24. Aksh_47 New Member

    Joined:
    Apr 24, 2012
    Messages:
    117 (0.14/day)
    Thanks Received:
    6
    Location:
    Goa, India
    Yes, i think CPU overheating is the prob, temps hitting 100C while borderlands is on.. just on the starting screen that too.. whereas the GPU temps still stay to be on the lower side, and the clocks stay at 950 Mhz.
    [​IMG]
  25. cdawall where the hell are my stars

    Joined:
    Jul 23, 2006
    Messages:
    20,668 (7.00/day)
    Thanks Received:
    2,980
    Location:
    some AF base
    Yup cpu is overheating.

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

Share This Page