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

Need Help for overclocking my i3570K - First Time

Discussion in 'Overclocking & Cooling' started by flaubert, Apr 25, 2013.

  1. puma99dk|

    puma99dk|

    Joined:
    Aug 29, 2005
    Messages:
    3,432 (1.03/day)
    Thanks Received:
    775
    maybe u just got a power hungry cpu, bcs since Sandy Bitch it was more a cpu limited then board limit, or that is what i have hear and seen when it comes to oc and that's the same for Ivy....
     
    AlienIsGOD says thanks.
  2. AlienIsGOD

    AlienIsGOD

    Joined:
    Aug 9, 2008
    Messages:
    3,463 (1.53/day)
    Thanks Received:
    1,507
    for example my 3570K is at 40x multi using 1.25 vcore, i intend to push it further in the coming days to see where the sweet spot is voltage wise for 4.3 -4.5ghz
     
    Crunching for Team TPU
  3. flaubert New Member

    Joined:
    Apr 25, 2013
    Messages:
    29 (0.05/day)
    Thanks Received:
    2
    1.25 for 40x is too high......not normal.

    Look this OC by a me friend with my own MoBo and CPU:

     
  4. AlienIsGOD

    AlienIsGOD

    Joined:
    Aug 9, 2008
    Messages:
    3,463 (1.53/day)
    Thanks Received:
    1,507
    1.25 is just fine actually, Cadaveca already mentioned that IB chips are finicky with voltage sometimes earlier in the thread. 1.25 is fine as im still seeing what my cpu will do with that voltage.
     
    Crunching for Team TPU
  5. droopyRO

    droopyRO

    Joined:
    Feb 14, 2012
    Messages:
    154 (0.16/day)
    Thanks Received:
    42
    Location:
    Romania
    Strange temp variation with a i5 3570k with or without overclock, on my rig.
    From 57* to 68* i got those temps playing Mass Effect 2 in 1440p with 29* C in my room, and the CPU@4118Mhz. Now the temps look ok to me but that 9* difference between cores is too big.
    I tried both a MX-3 and Prolimatech TIMs rotating my cooler 90* same temp difference between cores.
    What could be the cause ? Thanks.
    PS: sorry for highjacking the thread.
     
  6. EarthDog

    EarthDog

    Joined:
    Dec 31, 2009
    Messages:
    3,316 (1.89/day)
    Thanks Received:
    683
    That difference between the cores though annoying I'm sure, is fine.
     
    droopyRO says thanks.
  7. AlienIsGOD

    AlienIsGOD

    Joined:
    Aug 9, 2008
    Messages:
    3,463 (1.53/day)
    Thanks Received:
    1,507
    my old q 9450 had 1 core that always read 5-6c hotter than the rest; ie. 45, 40, 40, 40. It was annoying as hell
     
    droopyRO says thanks.
    Crunching for Team TPU
  8. flaubert New Member

    Joined:
    Apr 25, 2013
    Messages:
    29 (0.05/day)
    Thanks Received:
    2
    But if you tri to run CPU at 43-44: VCORE needed?
     
  9. AlienIsGOD

    AlienIsGOD

    Joined:
    Aug 9, 2008
    Messages:
    3,463 (1.53/day)
    Thanks Received:
    1,507
    haven't tried yet :p i dont have alot of free time to be tinkering so im at 4ghz for now

    edit: 42x multi w/ 1.250 vcore, running prime 95 now :D
     
    Last edited: Apr 30, 2013
    Crunching for Team TPU
  10. flaubert New Member

    Joined:
    Apr 25, 2013
    Messages:
    29 (0.05/day)
    Thanks Received:
    2
    Ok AlienIsGOD...;)

    I decided to replace the CPU. We hope to find a more fortunate! Now I decide if to buy another i3570k or switch to a i3770K....:D
     
    AlienIsGOD says thanks.
  11. Hood

    Hood

    Joined:
    Sep 25, 2012
    Messages:
    820 (1.08/day)
    Thanks Received:
    247
    That voltage seems high for such a small overclock. My 3570K was running 1.250 vcore stable at 4.5GHz, but a little hot, so it's now at positive offset .025 and running 1.226 max vcore@4.5GHz under load (stability testing with IBT & Prime95). I was only able to get it stable at those volts by tweaking the Digi+ CPU power settings (Asus Z77 board) to high, ultra, and extreme settings. I'm not familiar with EVGA's BIOS, so I can't be specific, but look for the CPU LLC and power settings, this will allow the CPU to handle higher clocks with less voltage.
     
    AlienIsGOD says thanks.
  12. AlienIsGOD

    AlienIsGOD

    Joined:
    Aug 9, 2008
    Messages:
    3,463 (1.53/day)
    Thanks Received:
    1,507
    its at 4.2 atm with 1.25, prime95 going on its 2nd hour
     
    Crunching for Team TPU
  13. flaubert New Member

    Joined:
    Apr 25, 2013
    Messages:
    29 (0.05/day)
    Thanks Received:
    2
    I buy the new CPU......Intel i7 3770K

    Well hopefully this time...:D
     
  14. droopyRO

    droopyRO

    Joined:
    Feb 14, 2012
    Messages:
    154 (0.16/day)
    Thanks Received:
    42
    Location:
    Romania
    1.25 in BIOS or CpuZ ?
     
  15. AlienIsGOD

    AlienIsGOD

    Joined:
    Aug 9, 2008
    Messages:
    3,463 (1.53/day)
    Thanks Received:
    1,507
    1250 in bios, 1.24 and change in CPU z

    ran 5 hours of prime95 w/o errors and then crunched WCG for 8 hours straight w/o any project errors .
     
    Crunching for Team TPU
  16. flaubert New Member

    Joined:
    Apr 25, 2013
    Messages:
    29 (0.05/day)
    Thanks Received:
    2
    For stability, personally I run prime for about 20mins, which indicates the worst instabilities. If i pass with no errors or WHEA errors I consider it good to go.;)
     
  17. EarthDog

    EarthDog

    Joined:
    Dec 31, 2009
    Messages:
    3,316 (1.89/day)
    Thanks Received:
    683
    20 mins? Not many would consider that remotely stable really... MANY have failures after 20 minutes. I would think a general consensus is 4-8 hours of P95 testing really... But surely 20 mins is not remotely a sign of stability.
     
    AlienIsGOD says thanks.
  18. AlienIsGOD

    AlienIsGOD

    Joined:
    Aug 9, 2008
    Messages:
    3,463 (1.53/day)
    Thanks Received:
    1,507
    mhm googling seems to indicate the same, 4-8 hours.
     
    Crunching for Team TPU
  19. Hood

    Hood

    Joined:
    Sep 25, 2012
    Messages:
    820 (1.08/day)
    Thanks Received:
    247
    I agree, if you're pushing mid to high overclocks and dropping your vcore too low for better temps (as you should be doing), it will show errors/crash within the first 10 minutes, then you know to slightly increase vcore or LLC, and test again. When I think it's all good I'll run it maybe an hour to be sure. I don't run anything near as intense as prime95 in my normal daily usage, and never run games or video conversion more than an hour anyway, so why test for hours, running up the electric bill and using up component life? I know when my system is stable, by observing vcore, VID, temps, behavior of IBT, etc. - and when it's just right you can sort of feel it in the way it responds, boots, opens programs, etc. I never have blue screens/crashes unless I'm pushing a CPU or RAM overclock to the limit, which by the way is a great way to get a feel for your CPU, RAM & motherboard and their limits/strengths.
     
  20. EarthDog

    EarthDog

    Joined:
    Dec 31, 2009
    Messages:
    3,316 (1.89/day)
    Thanks Received:
    683
    Running up the electric bill? You would want to do the math to see how much that actually costs (pennies).

    Using up component life? Ooof, no.. doesnt work that way..


    Stress test proper or take the chance of a BSOD or freezing. Nobody (in their right mind) would call 20 minutes of stress testing 'stable'.
     
    AlienIsGOD says thanks.
  21. Hood

    Hood

    Joined:
    Sep 25, 2012
    Messages:
    820 (1.08/day)
    Thanks Received:
    247
    It's like the difference between leaving a night light on and leaving all the lights on all night (~40 watts vs 350-400 watts), if that's how you run your home, or your computer, your light bill will be higher, obviously. And as for component life, what? you think it lasts forever because it's solid state? Any electrical/electronics engineer will tell you different, and even in my limited experience it's obvious - a wall of my office is plastered with dead motherboards from old systems that resisted all attempts at resuscitation. Add heat and voltage and component life decreases exponentially. and I clearly stated that my final run is at least an hour, not 20 minutes - of course I don't keep running prime after errors are detected, that would be stupid.
     
  22. flaubert New Member

    Joined:
    Apr 25, 2013
    Messages:
    29 (0.05/day)
    Thanks Received:
    2
    Hello to all...and cadaveca....:)

    Finally, i replaced my bad i3570k with a new Intel 7 i3770k....:D

    Made some OC testing and here are the results at this moment......:

    Cpu vcore set to 1.25V and frequency 45: turbo core enable

    [​IMG]


    Cpu vcore set to 1.27V and frequency 46: turbo core disabled.

    [​IMG]
     
    Last edited: May 12, 2013
  23. puma99dk|

    puma99dk|

    Joined:
    Aug 29, 2005
    Messages:
    3,432 (1.03/day)
    Thanks Received:
    775
    r u running stock cooler with those temps? ^^;

    bcs if that's ur Noctua NH-D14 i wonder if it's cooling right or ur airflow ain't working...
     
  24. cadaveca

    cadaveca My name is Dave

    Joined:
    Apr 10, 2006
    Messages:
    14,044 (4.51/day)
    Thanks Received:
    7,202
    Location:
    Edmonton, Alberta

    Much better chip this time, eh? Way more like what you were expecting.
     
    flaubert says thanks.
  25. flaubert New Member

    Joined:
    Apr 25, 2013
    Messages:
    29 (0.05/day)
    Thanks Received:
    2
    Yeah Cadaveca!......:toast:

    In any case, I am unhappy with my Noctua. For this reason I have already purchased and ordered the swiftech h220 liquid cooling.....:D
     

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

Share This Page