Printable View
It uses Intel's Linpack math libraries. While Prime95 uses 100% of your CPU, there are still lot's of wait statements involved. The linpack libraries are designed to get the data you need as fast as possible. You could try running the tester at lower priority to see if that fixes it.
Edit: Also make sure you didn't use too much ram. Figure out how much you need based on free ram.
As you suggested i've tried to run the test with less memory but the results still remain the same: test run fine, but pc gone crazy!
So i've swapped to the default settings, to see if these issues still persist.
Well, even with stock settings my system goes crazy during test: usb wireless dongle lost, everest nb & sb going crazy and then stopped reading and lcd poster starts flashing.
I'm running win Xp pro 32 bit.
Anyone else experienced strange issue like these?
BTW, i've didn't run with lower priority. i think if it runs at lower priority it doesn't use 100% cpu so the test will be useless, am i wrong?
It will still use 100% of the cpu, but if there is a function call to anything else in the system then windows will run it first.
No, I guess I don't know how to read the result?Quote:
Originally Posted by Loser777
I've never had any signals of any sort of instability, 0 WU errors running boinc wcg, gaming, etc. Uptimes for days on end, I guess it's good enough for me :D
Ya I got the same thing, I even tried running completely stock bios defaults and the numbers were different. This program def gets the CPU nice and toasty thoughQuote:
Originally Posted by loonym
Yeah, it's a better stability tool for servers where you need 100% data integrity.
So it should look like this? Same settings on my pc, I just selected less ram... meh close enough :D
http://lakesidepc.com/burntest_.jpg
Just tried this, system passed (woot!), but holy crap, it got HOT!
Realtemp reports the maximum temperature at 74*C. That's over 10 degrees hotter than Prime95 gets at these voltages... Yeesh. But if it does beat out prime 95 for showing stability issues, and only takes 8 or so minutes to complete, wow, I'm all for it. Really neat little program!
-Zigosity
Mine was hitting 105c and throttling back, I need a new cooler this Coolermaster sphere is worse then a stock one! :mad: Contact is good on CPU and everything + good case airflow
Residual needs to be a value of less than e^-09 ie e^-10, e^-11, etc, and Residual Normalised value should be roughly between e^-02 to e^-04. Ie 0.01 to 0.0001. Outside these ranges indicates memory or cpu errors.
I rather use Windows Powershell v2 and PSPlus Graphical IDE front end. Guess most people here wouldn't have the faintest idea what it is! But its nice. Its a scripting shell based somewhat on *nix shell environments. V1 was basically a blue command prompt, V2 has become a graphical IDE with great features.
http://s3.supload.com/thumbs/default..._ps_-_lp64.jpg
Your NB/RAM needs more volts or your powresupply can not keep up with the demand.Quote:
Originally Posted by panheimer
So my system look pretty good.. No higher then 50c
http://i223.photobucket.com/albums/d...Shawns/IBT.jpg
Yeah that looks good. Very stable.
this is pretty sweet, my temps were a little warmer than p95 1.36v 4200mhz i hit 63c after 5 instances and they all good.
It should be the max amount of RAM you can use... if you set it lower, it won't give accurate results... it's like running Small FFTs on Prime vs Blend.Quote:
Originally Posted by loonym
Not testing everything...
I also don't like running just five tests, I usually run ten... sometimes my system will just BSOD after 6 tests if it's not stable. (even though the first 6 tests were all error-free)
Ahhh, sorry... I still think I'll keep the setup though ok? ;)
I know your perspective... I'm not comfortable knowing that my system will fail a test no matter how extreme it is.
As suggested i've tried again:
- lowered priority, upped nb voltages 1 notch an run 10 tests: no residual norm variation but system went crazy;
- lowered priority, upped nb voltages 2 notches and 10 run tests: system went crazy and error wents out;
- lowered priority, swapped everything at stock settings an run 10 tests: no residual norm variation but system went crazy;
I think i'll have to wait for newer release of this sweet test tool, as i'm not confident seeing nb and sb temps going throttling around from -64°C and 0°C, lcd poster flashing, and usb wireless dongle going crazy.
Quote:
Originally Posted by panheimer
It's not the tool, I got that same problem with my Maximus Extreme when I tried to stress test it with other tools.
Anybody else than panheimer ran this with a rampage?
yes, on Q9450 @3.6ghz @1.328V (LCC enabled)Quote:
Originally Posted by RealTelstar
and... im afraid of runing this... core temps on 2 cores goes to ~90 in only 2min, other two are on more acceptable 80, but since i dont know what temperatures are real im afraid of runing it... and system is ~15h prime stable otherwise with temperatures on hot cores not going past 74....
i will try default clocks, just to see if i get any problems that he is reporting...
edit: works fine for me with default clocks, no wierd LCD blinking or something, works as it should!
With tools did you use? And how did you solved this issues?Quote:
Originally Posted by [cTx]Warboy
I run occt, prime95, 3dmark 06, super pi and none of theese gives me problem.
Yes. Here!Quote:
Originally Posted by RealTelstar
But i have to lower my Settings to got this heavy Stresstest successfully running. Normaly, for my daily needs and for 24/7 Distributed Computing, my Rig (Q6600 G0, Vista 64) runs on 8x478, for the Linpack Test i have to lower to 9x416. Main reason is, my Low-Price Memory. 8 GB Mushkin SP2 DDR2-800. In my daily Settings it runs fine at DDR2-956 5-5-5-18 PL8, but with the Linpack Test, i can only run one Test, then BSOD.
http://www.abload.de/thumb/linpakiq1.jpg
Helli
damn, the cpu actually burns with these computations