Intel Burn Test Utility

Discussion in 'Overclocking & Cooling' started by 1erCru, Sep 2, 2008.

  1. Riccochet

    Riccochet Off Topic Award

    Messages:
    22,048
    Joined:
    Apr 11, 2007
    When you can run P95 and IBT at the same time for 30 days straight without an error. :)

    I'll stick with my "if it doesn't crash in any app I run it's stable enough for me" theory.
     
  2. samsonJS

    samsonJS Limp Gawd

    Messages:
    308
    Joined:
    Jul 17, 2008
    What if your apps don't crash but there's silent data corruption? ugh.

    I haven't seen a point in doing more than 150-200 passes. After 100 you're looking really good, especially considering the real load you'll actually put on the machine. If ultimate stability is not important than 20-50 passes is plenty.
     
  3. supastar1568

    supastar1568 Limp Gawd

    Messages:
    382
    Joined:
    Jul 19, 2006
    Yeah, i think 100 passes is good for me personally, not that i would ever load it that long in real use. 100 passes is about 6-7 hours, which is enough for those once or twice a year (thanksgiving and christmas) gaming marathons.
     
  4. ShuttleLuv

    ShuttleLuv [H]ardness Supreme

    Messages:
    7,052
    Joined:
    Apr 12, 2003
    Be careful when running this on more low end, generic boards. Ran this on my Biostar P35 mobo and it was never the same after. It was sluggish, shut off by itself once, and then the other day just up died. Worked FLAWLESS before this. Watch yourselves.
     
  5. YeuEmMaiMai

    YeuEmMaiMai [H]ardForum Junkie

    Messages:
    15,029
    Joined:
    Jun 11, 2004
    ^ that was a coinsidence.....

    I ran that test of my machine listed below and it was perfect and passed the test 5 out of 5 times
     
  6. ShuttleLuv

    ShuttleLuv [H]ardness Supreme

    Messages:
    7,052
    Joined:
    Apr 12, 2003
    ^No...it passed on mine too, but a week later instabilities happened and a week and a half later from then I lost the board completely. It was the only thing that would be considered low end, being biostar, even though the board WAS great.
     
  7. Zero82z

    Zero82z [H]ard as it Gets

    Messages:
    27,897
    Joined:
    Jan 20, 2004
    Even so, that was most likely due to an inherently defective board. A properly functioning motherboard should have absolutely no problem passing IBT without incurring any damage.
     
  8. ShuttleLuv

    ShuttleLuv [H]ardness Supreme

    Messages:
    7,052
    Joined:
    Apr 12, 2003
    But how do ya know you got a normal working board until you run it? Eh eh ? ! ;) It just might end up being one of those iffy boards ya got.
     
  9. Ocellaris

    Ocellaris Ginger @le, an alcoholic's best friend.

    Messages:
    18,830
    Joined:
    Jan 1, 2008
    Might as well weed out the bad board then before they go down and take all your data with them :p
     
  10. markm75

    markm75 Gawd

    Messages:
    683
    Joined:
    Sep 12, 2008
    I tried a custom mode, set the test to 6000mb (out of the 8GB, this was what was free).. did 20 iterations.. it succeeded, no errors on the log..

    Then this morning i did "maximum stress", 10 iterations, 1 of the 10 failed...

    Would this indicate instability, with just one failure?

    I've had a heck of a time getting my new rig stable with 1600mhz native ocz ddr3 ram on a p5e3 deluxe board, which is truely only 1333 mhz native, but has overclocking options for 1600 and 1800..

    I've been wondering if i should have bought 1333 native ram and tried overclocking that to 1600..

    I haven't even been able to boot at 1600mhz, the tests i mentioned above are at 1333mhz in the bios.. cpu clock at 3.0ghz on the q9650 (7.5 multiplier, in this case i did up the fsb to 400, to give 1600 on fsb, ram setting at 1.8v)..

    Even at 333 fsb, i've had bluescreens and other issues.. this was with a ton of different skews and settings and higher voltage too.. both on NB and dram..

    So i scaled back, but left fsb higher this time, so far, except for the 1 iteration error with IntelBurn Test linpack.. its been stable, no bluescreens, but it seems doubtful ill ever achieve 1600mhz on the ram side..


    Any thoughts on this?

    Should i have bought 1333 and upped it to 1600mhz instead? Is one failure a true indicator of instability? (i also passed a mix test for 2 hours on OCCT)..

    -----------------
    Check out my Bluray/HDef Editing posting: here.
    TV: Samsung HLT-6187S
    LCDs: WH 24" L2410NM x 3
    PC (HTPC/gaming-FlightSims/server): Asus P5E3-D;Q9650 Quad 3.0ghz, 8GB ocz3g16004gk 8-8-8-26 1.8v. 1600mhz native speed; 12-drive icy dock MB453IPF-B (3ware pcie4 x4 raid5=d), Thermaltake Armor VA8000BWS, Vista x64
    PSU: 850 watt coolmaster
    Video: dual independent ati 4870s
    Audio: Onkyo TX-SR606 and Onkyo HTS780 speakers, Velodyne CT-10 sub; Xbox 360;
    Media Player: Mag NB500MG9 1.1 BD
    MediaRecorder: LG GGW-H20L
    MCSE 2003-Aug 2k6
     
  11. Ocellaris

    Ocellaris Ginger @le, an alcoholic's best friend.

    Messages:
    18,830
    Joined:
    Jan 1, 2008
    Just one failure means you are pushing something in your system past the stability point. How relevant that is differs from person to person, it is up to you if you want to operate your systems at settings that you know have failed.

    Try running 100x maximum test before going to sleep, see what the results are when you wake up.
     
  12. samsonJS

    samsonJS Limp Gawd

    Messages:
    308
    Joined:
    Jul 17, 2008
    Which one of the 10 failed? I'm not sure if it's significant, but sometimes I only get an error on the very first pass and I'm not sure why that is. I suspect it indicates a memory instability but I have no idea if this is true.
     
  13. Mongo

    Mongo only pawn... in game of life.

    Messages:
    273
    Joined:
    Oct 15, 2001
    Hmm thanks for this I will give it a go.
     
  14. Riccochet

    Riccochet Off Topic Award

    Messages:
    22,048
    Joined:
    Apr 11, 2007
    I've been having strange results with IBT lately. Sometimes the whole app will crash on the first pass. Then I can re-run it for 20+ passes without a crash or error. P95 can run indefinatly without error.

    My system never crashes, BSOD's or has any problems with any other app. So, I guess I'll chalk it up to a mystery issue.
     
  15. rockfella

    rockfella n00b

    Messages:
    57
    Joined:
    Mar 15, 2008
    There is a major difference in temps between Orothos and IBT. With Orothos my temps were 63 deg and IBT 74(LOAD). Though this is a very powerful tool, i was wondering whether it is really required as one hour Orothos stable is pretty decent! What else can stress our system so bad that we need to make sure our PCs pass this IBT?
     
  16. Ocellaris

    Ocellaris Ginger @le, an alcoholic's best friend.

    Messages:
    18,830
    Joined:
    Jan 1, 2008
    Ack, just one hour of Orthos testing is nothing :eek:
     
  17. rockfella

    rockfella n00b

    Messages:
    57
    Joined:
    Mar 15, 2008
    ^ And 3 tests of IBT (max stress) ?
    Temps went upto 74 :eek:
     
  18. Ocellaris

    Ocellaris Ginger @le, an alcoholic's best friend.

    Messages:
    18,830
    Joined:
    Jan 1, 2008
    3 tests is not much, thats only a few minutes at max load. If you want some more confidence 20+ is a good place to start.

    Another thing to keep in mind is that 74C isn't exactly melting the chips or anything. You would be fairly shocked to see what some OEM systems run at for temps after they get a bunch of dust built up in there. I have a Macbook Pro (the original Intel model) for work and that thing is too hot to rest on my lap without a pillow. I've checked it under load and that has been up to 95C and running fine.

    The problem is you do not know. People always want better optimized games and apps, and developers work to get everything they can from the hardware. If you take a game like Crysis, that put a significant higher load on people's systems than most people have previously experienced. People with "100% stable" GPU overclocks were getting artifacts and crashes and had to clock back down.

    So do you stress test for the lowest common denominator or for the worst case scenario? When it comes to stability, I think should test for the worst case scenario to have the best chances of success.
     
  19. Eruditass

    Eruditass Gawd

    Messages:
    686
    Joined:
    Jul 7, 2002
    i've heard this a lot before and am wary because of it.

    however, i just ordered a bunch of components for a new computer so if they don't pass a whole lot of this IBT, they are getting RMA'd!

    *in reality i don't think i'd be able to part with it.
     
  20. Khaydarin

    Khaydarin 2[H]4U

    Messages:
    2,534
    Joined:
    Dec 13, 2005
    This looks like it's an awesome tool; I'll definitely be using it for the new PC I'm building next weekend. I'm glad we can all get past the endless P95 tests at last...
     
  21. samsonJS

    samsonJS Limp Gawd

    Messages:
    308
    Joined:
    Jul 17, 2008
    I'd go for a combination of tests Khaydarin. Prime95 still does things that IBT doesn't do and people have had to bump vcore to get both stable. I've found another corner case for my machine is folding (cpu + gpu) and watching hi-res video (720p or better) in VLC, so after other tests I leave a movie playing all night as another stability test. Some like to run 3dmark or similar with prime95 or orthos simultaneously. One test alone is inconclusive.
     
  22. ep3w

    ep3w Gawd

    Messages:
    626
    Joined:
    Apr 4, 2008
    I am using vista 64 bit with 4gb of ram and I noticed when i run the burn test, my temps reach about 48C, where as with orthos I will hit 55C which seems to be opposite of most peoples results. Is something not quite running right? Task manager says both cores are loaded 100% and one time I tried opening firefox while it was running and it said I was running out of memory so I know its using it all.
     
  23. Khaydarin

    Khaydarin 2[H]4U

    Messages:
    2,534
    Joined:
    Dec 13, 2005
    I see what you're saying. Do you recommend I start with one test or another?
     
  24. Zero82z

    Zero82z [H]ard as it Gets

    Messages:
    27,897
    Joined:
    Jan 20, 2004
    The order won't really make a difference.
     
  25. samsonJS

    samsonJS Limp Gawd

    Messages:
    308
    Joined:
    Jul 17, 2008
    Except for the time you spend testing! I recommend 5-10 passes of IntelBurnTest for quick tests and then 20+ passes later. After that I would leave prime95 overnight when you have time and the same for the others.
     
  26. Dan_D

    Dan_D [H]ard as it Gets

    Messages:
    54,634
    Joined:
    Feb 9, 2002
    I use this test on my Skulltrail setup. Worked very well. It let me know what was stable and what wasn't? What more can you ask for?
     
  27. Khaydarin

    Khaydarin 2[H]4U

    Messages:
    2,534
    Joined:
    Dec 13, 2005
    I'll use this method then.
     
  28. Dan_D

    Dan_D [H]ard as it Gets

    Messages:
    54,634
    Joined:
    Feb 9, 2002
    That's the method I used for my own setup. That's how I ensured my 4.0GHz overclock was stable. When I switched to a 4x2GB setup for 8GB of RAM my system was no longer stable at 4.0GHz and I had to back down to 3.6GHz for both processors.
     
  29. Katalysis

    Katalysis [H]ard|Gawd

    Messages:
    1,813
    Joined:
    Aug 8, 2008
    At what point do I shut this thing down out of fear? It ran fine but I freaked out when I saw what I assume to be my Northbridge hit 60C, the highest core temp on my Q6600 was a 66, about 11 degrees higher than Prime. This was on auto voltage, but I still got scared.
     
  30. Zero82z

    Zero82z [H]ard as it Gets

    Messages:
    27,897
    Joined:
    Jan 20, 2004
    Don't shut it down out of fear unless something crashes. Just leave it for whatever length of time you need to depending on the utility you're using, and don't intervene unless something goes wrong. The temps you were getting are perfectly safe.
     
  31. corey407woc

    corey407woc [H]Lite

    Messages:
    103
    Joined:
    Aug 24, 2008
  32. Zero82z

    Zero82z [H]ard as it Gets

    Messages:
    27,897
    Joined:
    Jan 20, 2004
    That's the idea.
     
  33. Vinnard

    Vinnard Gawd

    Messages:
    681
    Joined:
    Jul 7, 2008
    Wow, that does get it hot! 61c on my e7200 @ 3.5, prime gets it up to 56
     
  34. tnt3k

    tnt3k [H]ard|Gawd

    Messages:
    1,909
    Joined:
    Dec 3, 2006
    hmmm my q6600 G0 went up to 65 max then never went any higher. idles at 40 tho. however with the IBT after about 5 minutes it crashes. I've had to constantly bump the voltage higher
     
  35. Zayla74

    Zayla74 [H]Lite

    Messages:
    64
    Joined:
    Apr 26, 2008
    I really like IBT. It may not be the perfect test, but is able to find most any problem in a matter of minutes.
     
  36. Katalysis

    Katalysis [H]ard|Gawd

    Messages:
    1,813
    Joined:
    Aug 8, 2008
    I've replaced Prime with this test, because it's so quick and easy. Should I still run a 24hr Prime test if I can pass 5 IBT's in a row?
     
  37. Sovereign

    Sovereign 2[H]4U

    Messages:
    3,080
    Joined:
    Mar 21, 2005
    My 3.2GHz Q6600 survived 8 hours of Prime but bombed out in seconds (BSoD) with this. I'm not an "absolute" stability purist, and again this is probably the most unrealistic load condition imaginable so I'm not concerned. It is nice though that you can get a result in minutes rather than hours.
     
  38. NoxTek

    NoxTek The Geek Redneck

    Messages:
    9,267
    Joined:
    May 27, 2002
    Sweet jesus...


    [​IMG]


    (check the 'Maximum' temps recorded in RealTemp... :D :D :D)

    Uhm I guess I'll back down on my overclock until I can afford a cryogenic chamber for my rig... :eek: At least it made it through the test! I kept watching RealTemp the whole time and thinking "OK if it gets past 90C I'm going to stop it... "

    Guess the ol' Tuniq Tower just can't cut the mustard....
     
  39. Apdravya82

    Apdravya82 Limp Gawd

    Messages:
    142
    Joined:
    Oct 29, 2008
    LoL,

    my system shuts down in the 90's....

    I'm not sure where my problem is with heat ATM

    Scythe Ninja qx6700 even at low voltage i idle in high 30's to mid 40's

    Intel burn test raises temp 40c

    I've changed fans, reseated and reapplied arctic silver.

    Only conclusion is the CPU was a free oem sample to my work, perhaps something with the chip.

    Or

    Stupid as mounting clips instead of screws on the heatsink.
     
  40. Zero82z

    Zero82z [H]ard as it Gets

    Messages:
    27,897
    Joined:
    Jan 20, 2004
    My CPU has hit 93C while running IBT :p.