Need Help Identifiying Orthos Error?

[MNFirstBlood]

Hey Guys Im new with using Orthos Stress Test and was hoping someone could help me indentify the following error.

1:FATAL ERROR: Final result was A58FC8B4, expected: 3FC3CCD7.
1:Hardware failure detected, consult stress.txt file.
1:Torture Test ran 35 minutes 8 seconds - 1 errors, 0 warnings.
1:Execution halted.

I could not find stress.txt file. I think it may be my psu because my temps were fine at max load 52c and my ram is not oc'd yet. I'm hoping its my psu can someone confirm? thanks in advance.

[Stormcrow]

PSU? No, your PSU only supplies power to the system. Thats a memory error, it glitched and returned the wrong value then was expected. So your system isn't overclocked at all? Go into the BIOS and check your voltage settings for the memory and manually set them and the timings if you have to. If everything checks out, you may have a bad stick of RAM especially if it generates an error in under an hour.

Oh yeah, there isn't a stress.txt file for Orthos. The guy who converted it from Prime95 forgot to include it with the program, but all it says is that your system is unstable to make sure basically everything is set up right.

[MNFirstBlood]

My thoughts were that the psu was not delivering enough power to the system. I know the psu is not the error but maybe the reason. Yes my timings and voltage are correct as to manufactures specs. I had my cpu overclocked during the test to 3.4 at 1551fsb
but did not have my ram overclocked. Last night ran memtest for 12hours with no errors. With my cpu at 3.2 1441fsb. In my bios I have the voltage set to 1.425 on the cpu but it only reads 1.38 and the voltage reading is in green type my thoughts that I was possibly not delivering enough power to substain those clocks. The temp only reached 51c and max load It should be stable. I also noticed that my ram stays at 800mhz untill I clocked my cpu up to 3.4 my ram then went down to 761mhz on its own so I figured my system is thirsty for power. Anything else I should do to see if the ram is faulty besides running memtest86? I really dont think its the ram.

[Elysium]

I think I see the problem... most motherboards have this thing called "VDrop" or VDroop". This is primarily a power saving utility, but can get really annoying when overclocking. You said you had the voltage set at 1.425v, but it only reads 1.38v? That would be the VDroop, not your PSU.

The good new is I think that there is a mod for the Evga 680i board that can fix the VDroop... Here it is: Vcore Voltage Not Accurate on 680i

And a pic:


(sorry is so big)

[MNFirstBlood]

Thanks yeah I did the vdrop mod today I found that on the evga site with exact same pic. Orthos has run stable now but my ram still dropped to 775 when i clocked the cpu past 3.4ghz. I have it set to unlinked any ideas why the ram would clock down on its own?

[Elysium]

Originally Posted by MNFirstBlood

Thanks yeah I did the vdrop mod today I found that on the evga site with exact same pic. Orthos has run stable now but my ram still dropped to 775 when i clocked the cpu past 3.4ghz. I have it set to unlinked any ideas why the ram would clock down on its own?

Yeah, does that on my board too... I think it's because the ratio between the RAM and the FSB (when you set the CPU at 3.4GHz) wasn't as "clean cut" as it used to be... like there was a "remainder" of some sort, just like in math (since FSB and RAM can't do decimals). Hope that makes sense...

[MNFirstBlood]

Yes, that is pretty good for v-drop with a simple pencil You can do a bit better by shading the resistor a little more throughly.

Your memory is downclocking because when you set it to unlinked, the board is using dividers to get as close to your requested memory speed. Sometimes you are right on the cusp of a different divider at a given FSB speed, so by raising the FSB you force the board to use a different ratio and it's still trying to keep the memory at your requested speed.

Yes I found this on evga as well pretty much explains it.