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

Stress Test Error, Need Help

Discussion in 'Overclocking & Cooling' started by DeAtHWiSh, Dec 26, 2007.

  1. DeAtHWiSh

    DeAtHWiSh New Member

    Joined:
    Dec 24, 2007
    Messages:
    197 (0.08/day)
    Thanks Received:
    48
    Location:
    Miami, FL
    Hey when I run the stress test now at anything higher than 3.7Ghz it says,
    FATAL ERROR: Rounding was 0.49609375, expected less than 0.4
    Hardware failure detected, consult stress.txt file.

    Any idea what this means?
  2. PyroX1040 New Member

    Joined:
    Dec 17, 2007
    Messages:
    283 (0.12/day)
    Thanks Received:
    13
    Location:
    Chicago, IL
    What stress tester are you using, aswell as what version of that tester :X
  3. btarunr

    btarunr Editor & Senior Moderator Staff Member

    Joined:
    Oct 9, 2007
    Messages:
    28,253 (11.36/day)
    Thanks Received:
    13,585
    Location:
    Hyderabad, India
    Downclock asap!
  4. PyroX1040 New Member

    Joined:
    Dec 17, 2007
    Messages:
    283 (0.12/day)
    Thanks Received:
    13
    Location:
    Chicago, IL
    Downclock?? with his cooling system, it's sounds like a software issuse not a hardware......
  5. Kursah

    Kursah

    Joined:
    Oct 15, 2006
    Messages:
    7,647 (2.69/day)
    Thanks Received:
    1,592
    Location:
    Missoula, MT, USA
    That's a pretty high OC, you either need more Vcore or you need to downclock. Usually those are the primary reasons for a failure in a Prime test.
  6. Kursah

    Kursah

    Joined:
    Oct 15, 2006
    Messages:
    7,647 (2.69/day)
    Thanks Received:
    1,592
    Location:
    Missoula, MT, USA
    It's a CPU mathematical test, if the CPU is undervolted or overclocked beyond stability, no matter the temps it will send an error like that. I've seen it many times with many different rigs, AMD and Intel and the two solutions I listed above will fix it damn near every time, well in my case it was every time.

    :toast:
  7. btarunr

    btarunr Editor & Senior Moderator Staff Member

    Joined:
    Oct 9, 2007
    Messages:
    28,253 (11.36/day)
    Thanks Received:
    13,585
    Location:
    Hyderabad, India
    Such errors are a result of "good OC, Bad cooling"
  8. PyroX1040 New Member

    Joined:
    Dec 17, 2007
    Messages:
    283 (0.12/day)
    Thanks Received:
    13
    Location:
    Chicago, IL
    With that CPU and that cooling system he should be able to push 3.8Ghz - 4.0Ghz using that core, the highest record on a 6400+ was 4.42Ghz using Liquid Nitrogen...
  9. PyroX1040 New Member

    Joined:
    Dec 17, 2007
    Messages:
    283 (0.12/day)
    Thanks Received:
    13
    Location:
    Chicago, IL
    I think it's a simple power issue or software mis-read... he should be able to atleast test the darn thing!
  10. Kursah

    Kursah

    Joined:
    Oct 15, 2006
    Messages:
    7,647 (2.69/day)
    Thanks Received:
    1,592
    Location:
    Missoula, MT, USA
    He'll need more than 1.5vcore to attain those...it's not always about cooling folks. Every time I've seen that same error, it was due to lack of voltage due to a high OC. Increase Vcore or drop clocks, it will pass either way.

    I'm just lending my experience with this issue since I've seen it many times and know why it comes up. If you guys have a solution, then by all means suggest one. Saying "it should work" doesn't fix the problem.

    :toast:
  11. PyroX1040 New Member

    Joined:
    Dec 17, 2007
    Messages:
    283 (0.12/day)
    Thanks Received:
    13
    Location:
    Chicago, IL
    I've been able to push my CPU, not stable due to lack-o-watts >< but either way I got it to 3.7Ghz using 1.475Vcore >.> so please to be explaining..
  12. sneekypeet

    sneekypeet Unpaid Babysitter Staff Member

    Joined:
    Apr 12, 2006
    Messages:
    21,482 (7.08/day)
    Thanks Received:
    5,995
    Have to agree with this 100% . Its a cpu voltage issue. Iv a bit of voltage doesnt correct this issue, you may hae reached the linit of the CPU, or the stability limit of all your components combined!
  13. Kursah

    Kursah

    Joined:
    Oct 15, 2006
    Messages:
    7,647 (2.69/day)
    Thanks Received:
    1,592
    Location:
    Missoula, MT, USA
    lol..
    Sure you can get there with less vcore. A lot of new OC'ers do and wonder why they have more random crashes or freeze/crash during CPU intensive operations. Performing a stress test such as Orthos/Prime puts a load on your CPU that you won't generally meet, but if you pass it after such a time frame (I go for 6hrs+), then you have a good idea you're stable. If you can't make it 20 minutes...well you're choking your CPU for the power it needs to do the proper processing you want it to do.

    I would rather have a stable OC'd system than run my CPU up with lower vcore for cooler temps. So I'll find my happy medium, somewhat low volts and a nice OC to boot. Take a look at my system specs for a good idea. 90% OC on a C2D with a minimal increase in Vcore. It took a while to find it, and it was worth the extra time.

    Also, feel free to check my guide in my sig. It's a lot of info donated from many on TPU, a lot of good info.

    Either way, that chip needs more volts...or he can run it not stable, but if he loads it the way Orthos/Prime does ever...don't expect it to run like a dream. But everyone has their preferences.

    Good enough for ya Pyro? :toast:
  14. DeAtHWiSh

    DeAtHWiSh New Member

    Joined:
    Dec 24, 2007
    Messages:
    197 (0.08/day)
    Thanks Received:
    48
    Location:
    Miami, FL
    Im using Prime 2004 Ver.0.30, And yeah cooling isnt an issue, and I have raised the V-core to 1.6V, maybe its memory holding me back?
  15. Kursah

    Kursah

    Joined:
    Oct 15, 2006
    Messages:
    7,647 (2.69/day)
    Thanks Received:
    1,592
    Location:
    Missoula, MT, USA
    Could be a number of things...if raising vcore didn't help..like sneeky said, you may have passed the stable limit of that chip. No 2 chips will OC the same...one guy might get 4.0, one might only get 3.6...it's a roll of the dice.

    I would start lowering speeds until you can pass Orthos/Prime for at least 20 minutes. Then run it over night...if you can pass a full night run, at least you'll know you are stable.

    :toast:
  16. PyroX1040 New Member

    Joined:
    Dec 17, 2007
    Messages:
    283 (0.12/day)
    Thanks Received:
    13
    Location:
    Chicago, IL
    I was able to hold that setting for atleast 2 hours before I crashed >.> while running full load on air...
  17. Kursah

    Kursah

    Joined:
    Oct 15, 2006
    Messages:
    7,647 (2.69/day)
    Thanks Received:
    1,592
    Location:
    Missoula, MT, USA
    lol...well if that's how you want to OC...

    I can run this rig OC'd as long as I want...loaded as heavily as I want and it won't flinch.

    I'd take my route over 2 hours any day. Also no 2 chips will OC the same as I stated before...results may vary.

    :toast:

    Edit: And either way, we're not here to talk about our OC's, were here to fix the OP's problem. So I'm still waiting for a solution from you Pyro... :D
    And please don't tell him to run it unstable!
  18. DeAtHWiSh

    DeAtHWiSh New Member

    Joined:
    Dec 24, 2007
    Messages:
    197 (0.08/day)
    Thanks Received:
    48
    Location:
    Miami, FL
    I've brought it up to 4.1GHz before just not stable, I think it's memory.
  19. Kursah

    Kursah

    Joined:
    Oct 15, 2006
    Messages:
    7,647 (2.69/day)
    Thanks Received:
    1,592
    Location:
    Missoula, MT, USA
    Could very well be, and since your mem controller is in your CPU...there could be an issue there. You gonna run a couple instances of Memtest to verify?

    I go either way, for Windows or DOS, let us know what you find! :toast:
  20. philbrown23

    philbrown23 New Member

    Joined:
    Dec 22, 2007
    Messages:
    1,231 (0.51/day)
    Thanks Received:
    147
    Location:
    vermont
    well a successful oc is one that does not crash ever that is unles you enjoy ruining you cpu ;)
  21. DeAtHWiSh

    DeAtHWiSh New Member

    Joined:
    Dec 24, 2007
    Messages:
    197 (0.08/day)
    Thanks Received:
    48
    Location:
    Miami, FL
    Ruining I think is pretty hard with all the protection and super cooling I have, I twould take a long time. Memory seems stable, just it say ddr2-880mhz for some reason???
  22. Hawk1 New Member

    Joined:
    May 8, 2007
    Messages:
    1,249 (0.47/day)
    Thanks Received:
    163
    Location:
    The Big Smoke, Canada
    Agreed. I've had Orthos crash after 18.5 and 19.5 hours of blend test, so I ususally let it go the full 24hrs before I'm satisfied (and even then, I question wether it's F@H 24/7 stable.)
  23. DeAtHWiSh

    DeAtHWiSh New Member

    Joined:
    Dec 24, 2007
    Messages:
    197 (0.08/day)
    Thanks Received:
    48
    Location:
    Miami, FL
    Nevermind, I'm gonna bump up the memory to 470MHz see what happens. Disregard what I wrote before, it's ddr2-880 because the memory is at 440Mhz, lol, shhh, don't tell nobody.
  24. DeAtHWiSh

    DeAtHWiSh New Member

    Joined:
    Dec 24, 2007
    Messages:
    197 (0.08/day)
    Thanks Received:
    48
    Location:
    Miami, FL
    This crap isn't working out for me tonight, I think I'm gonna sleep on it and continue tomorrow. I'll get back to you guys soon. And if anyone has anything they can suggest please do, I will greatly appreciate it. These are my settings before I failed the stress test after 2wks of stability:

    Max Stable MHz for AMD 6400+ X2
    AMD Athlon 64 X2 6400+ (3200MHz)
    Socket AM2 (940)
    90nm
    Currently @ 1.520V VCORE
    Core Speed @ 3760.2 MHz
    Multiplier @ x16
    Bus Speed @ 235.0 MHz
    HT Link @ 1175 Mhz per core
    Cooling: Vapochill Lightspeed
    Evaporator @ -50C
    CPU @ -21C on Full Load during Stress Test

    Case
    Lian Li PC-60 Plus Modified (Silver)
    120mm intake, front (1)
    80mm exhaust, top (1)
    80mm exhaust, rear (1)

    Power Supply
    Antec True Power Quattro
    850W w/ four +12v @ 18 Amps

    Cache
    L1 D-Cache
    Size 64 Kb x 2
    L1 I-Cache
    Size 64 Kb x 2
    L2 Cache
    Size 1024 Kb x2

    Mainboard
    ASUSTeK Computer Inc.
    Crosshair Bios Revision 0904
    Chipset: Nvidia SPP190
    Southbridge: nForce 590 SLI

    Memory G.Skill
    DDR2 PC2-6400 (400 MHz) 5-5-5-15
    Voltage @ 1.85v
    Total: 8192 Mb
    Timings:
    DRAM Frequency @ 470.0 MHz
    FSB/DRAM: CPU/8
    CAS# Latency @ 5.0 Clocks
    RAS# to CAS# Delay @ 5 clocks
    RAS# Precharge @ 5 Clocks
    Cycle Time @ 15 clocks
    Bank Cycle Time @ 24 Clocks
    Command Rate @ 2T

    Graphic Cards
    XFX Geforce 8800 GTS x2 in SLI
    Clock Rate @ 600 MHz
    Memory Clock @ 1900 Mhz
    Shader Clock @ 1500 Mhz
    Memory: 640 Mb
    Bus Type: PCI-E
    Memory Type: GDDR3
    Memory Bus: 320 bit
    Cooling: Zalman VF1000 Ultra Quite VGA Cooler
    VGA Ram Cooling: Zalman RHs88 Heatsink
    Full Load Temp @ 72C
  25. Kursah

    Kursah

    Joined:
    Oct 15, 2006
    Messages:
    7,647 (2.69/day)
    Thanks Received:
    1,592
    Location:
    Missoula, MT, USA
    Well, read the voltages from your PSU under idle and load...how's the vcore droop? 12v drop under load?

    Run memtest through at least one 100% pass (I usually allow it to run many passes overnight for assurance). Drop your CPU/Bus speed to find some sort of stability to work with. Then you can start working your way back up...but I don't think pushing it beyond a stable level will net you much.

    You said it was 2Wks stable at those settings, how long did it pass Orthos/Prime? It may have seemed stable, yet still have been unstable depending on your findings.

    I'll check back when I get off of work, I got about 5 hours to go. :toast:

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

Share This Page