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

3.40 temps lower than 3.00

Discussion in 'RealTemp' started by navilor, Apr 4, 2010.

  1. navilor New Member

    Joined:
    Apr 4, 2010
    Messages:
    3 (0.00/day)
    Thanks Received:
    0
    I just know that I am missing something blatantly obvious here. Clues for the clueless here are always appreciated. Also, if this is the wrong area to post this issue, please direct me to the correct location. I apologize if this is so.

    My current hardware is a Core i7 920 on an EVGA X58 A1 motherboard running bios SZ2E. I am running Windows Vista 64 bit Ultimate.

    When running RealTemp 3.00 it would report what I felt was the accurate temperature.

    When running RealTemp 3.40 it reports a 15c drop in temperature.

    When running RealTempGT.exe included with 3.40 it shows three cores but displays the same temperature as RealTemp 3.00. Yes I understand I am not running the six core Gulftown processor which realtempgt.exe was written for.

    I assume that I am doing something completely wrong and that is resulting in a different temperature than I expected.

    Please advise on if I am doing something incorrect or should revert to an earlier version of RealTemp.
     
  2. Mussels

    Mussels Moderprator Staff Member

    Joined:
    Oct 6, 2004
    Messages:
    42,565 (11.40/day)
    Thanks Received:
    9,836
    TJmax settings have changed between versions, likely. Tune it in yourself, to the correct value.
     
  3. burebista

    burebista

    Joined:
    Sep 8, 2005
    Messages:
    615 (0.18/day)
    Thanks Received:
    185
    Location:
    Romania
    IIRC 3.0 doesn't know about Nehalem. Use 3.40 or latest beta.
    But you don't have to adjust nothing. TJMax (in fact is TJ Target %^&*#@! Intel for their terminology) is written directly in Nehalem MSR so no more guess.
     
  4. Mussels

    Mussels Moderprator Staff Member

    Joined:
    Oct 6, 2004
    Messages:
    42,565 (11.40/day)
    Thanks Received:
    9,836
    that was probably the problem then, the earlier one needed manual TJ entered, new one doesnt - hence the discrepancy
     
  5. navilor New Member

    Joined:
    Apr 4, 2010
    Messages:
    3 (0.00/day)
    Thanks Received:
    0
    I had prevously overridden TJMax. Updated information.

    OK. So just to be sure on this because I really don't want my CPU to go out in a blaze of glory...

    My BIOS reports the idle temperature to be around 40C. RealTemp 3.40 reports it to be about 45C. I cannot explain the discrepancy.

    Perhaps this is the result of different CPU thermisters being used by RealTemp and my BIOS.

    A different monitoring tool, SpeedFan 4.40, has a bug against it where it reports temperatures as being 15C below normal. They have to be manually adjusted. Mine are adjusted accordingly per their bug report.

    BIOS reports 40C
    RealTemp reports 45C
    RealTempGT (just for informational purposes) reports 45C
    SpeedFan reports 45C

    Assuming RealTemp is accurate then my BIOS is reporting the temperature as 5C lower than it should and RealTempGT and Speedfan are reporting temperatures accurately.

    Their bug report can be found here:
    http://www.bugtrack.almico.com/view.php?id=1548

    Is this just me being paranoid? I am loathe to lose a CPU due to heat because I do not know the actual temperature. Yes a 5C difference concerns me.
     
  6. Mussels

    Mussels Moderprator Staff Member

    Joined:
    Oct 6, 2004
    Messages:
    42,565 (11.40/day)
    Thanks Received:
    9,836
    whoever said your BIOS has to be right?
     
  7. tigger

    tigger I'm the only one

    Joined:
    Mar 20, 2006
    Messages:
    10,183 (3.18/day)
    Thanks Received:
    1,399
    I never trust bios temps they always seem to report very low.
     
  8. burebista

    burebista

    Joined:
    Sep 8, 2005
    Messages:
    615 (0.18/day)
    Thanks Received:
    185
    Location:
    Romania
    BIOS temperature is from a sensor calibrated by MB manufactures. Usually you should ignore it.
    Core temperatures are from sensors integrated into CPU. Those should be considered.

    But as a rule of thumb Intel core sensors are a mess. Conroe sensors was OK, Wolfdale/Yorkfield were a mess, Nehalem sensors are decent, Clarkdale/Lynnfiled are again a mess. Why? Because Intel made them to be accurate as close as they are to TJMax/TJ Target value. So, as far as you are from TJMax/TJ Target they are more innacurate.
    That's why I keep telling to rely only on one value: distance to TJMax. This is the value read straight from on die sensors. Keep that value >20 in daily use and forget about temps.

    And as a general rant peoples care too much about Intel temps. Don't worry, Intel made them to throttle if anything goes wrong, they're self-protected. I have my friend with his strange hobby to test heatsinks and after months and months of testing came to this conclusion: after testing at 100C in daily use his Nehalem is as good as new. :D
     
    navilor says thanks.
  9. navilor New Member

    Joined:
    Apr 4, 2010
    Messages:
    3 (0.00/day)
    Thanks Received:
    0
    Holy cow. Everyone I had previously listened to in forums and on many websites said that the maximum safe temperature for the Core i7 was 85C.

    It is going to take me a while to adjust to the "new to me" normal. That you very much for your input.
     
  10. burebista

    burebista

    Joined:
    Sep 8, 2005
    Messages:
    615 (0.18/day)
    Thanks Received:
    185
    Location:
    Romania
    Agree, for daily use max. 85C seems a very good advice. I told you that you should keep distance to TJMax >20 and that's translated to 80C on cores. :)
    All I want to say with my friend tests is that Intel CPU's are well build regarding protection for high temps. Not high temps are killing them on long term but high VCore. ;)
     

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

Share This Page