Errors in Super Pi, and Prime95! What exactly do they mean?

more vCore for Prime95 small fft fails

Agree with the post underneath me...

Dont Use Prime95 myself i use orthos for testing, however they both seem to come up with the same errors the whole rounding thing. basically wat it means is your system is not stable. you;ll need to post ur system specs & bios settings here for people to b able to help u better.

Normally upping the vcore wud help also turning things off in the BIOS such as intel speedstep. like i say, u need to post ur specs for the experts on here to help u better.

Like Silverion jus said, more vcore  

Orthos is Prime!   It's just wrapped in a prettier package.

But yea, like the others and the great Christopher Walken say,

"Needs more vcore"

To be more specific... Bump vcore by .05 or .1 and retest. Use Prime or Orthos though, Super Pi is over too quickly. If it fails, bump it again and retest. Just remember to watch temperatures closely as you increase vcore. They can get out of control fast  

I'll mess with it again come the morning! Core # 2 likes to fail the quickest. Oh... I have a b3 proc with two cores that are hotter than the other two. 0 and 1 maintain like 4 c hotter than 2 and 3. My G0 appears to only have a single flame thrower core, This time its 2 thats out of control! Imagine that. Its the one that fails the test the fastest. Is that normal for a core or two to be that amount hotter than the others? Thats odd, especially being core 2. Zero I can see being out, or 3! Got me, so I am asking you all! Its neat to see what goes on with others hard ware! (No pun, yuck!)

--Lupi!

I have noticed that pre fetch likes to help me get errors in Prime95 when doing the blend test. If I just jump right into testing a new configuration, and pre fetch is loading junk into RAM, my blend test will fail a core. Anyone know of a way to disable pre fetch? It drives me nuts! And hogs my RAM, though its a zillion times faster at releasing RAM when needed, as opposed to loading whatever its pre fetching into RAM, that is like 5-10 megs a second in pulses.

Does that happen to anyone else? Or say, you run the blend test on all four cores and then run super pi in the background doing the 32 million while the blend test is going on. Poof, one core fails.

--Lupi!

Hmm, got her cranking away at the normal every day rate of 3.6. x9 multi with 1600 fsb. My Ram for some reason cant run at its advertised rate, new it was trash when I got the huge voltage version in the mail. I dont know why the processors speed would effect the rams performance, but everyone else seems to go with 1:1 with the FSB, so I have it in sync, and am testing the timings from 1066 5-5-5-15, to 800 4-4-4-12. Its been blending nicely with prime 95 for over an hour and a half...

And yes, I needed VCore! Its set to a modded 1.4, it hits 1.39 under full load. But I also needed to lower the Rams speed. It slowly needed to go below its 1066 as I upped the FSB (or processor speed.) At 3.2, I had to make it something like 966, then 3.4 had to be around 933. I just synced it up at 3.6 and will see if the 4-4-4-12 is 8 hour stable, then tighten it a lil more. Voltage for the ram is at 1.9 since its at 800, not 2.35 @ 1066. (Had at 1.85 for normal 800/5-5-5-15, went to 1.9 for 4-4-4-12.)

Thanks!!

--Lupi!!

What's your 3dmark06 score?  

Thats a very good question! never ran it before.  ~ I will have to get the progie. Stopped the blend at 8 hours with no errors, and I had to install a bunch of junk at the same time, and use AIM and skype, blahblha.... I am currently on my 780i board... havent started with playing yet, but I did have trouble xfering my already installed winblows on my raid zero array. If only when it installed the new motherboard resources it would restart itself! Once it does all the searching and installing new drivers, it trashes my old usb junk for new ones and that causes my mouse and keyboard to stop working, and it takes like 3 mins longer after it stops working for it to say "windows needs to restart in order for the changes to take effect." You've all seen it, but I cant click on the restart now!! And if I force re start it like I have to, it doesnt save the new drivers, lol. And then the only way for me to even get into windows is the use last known good settings, and it starts all over, since it loads again with the old boards resources. Then the mouse and keyboard USB drivers are changed and I lose the mouse and keys, then a couple mins later it happily asks if I would like to restart my computer now so it can apply the changes!

Arrrrrgggghhhh!!

--Lupi!

It appears that VCore solves all sorts of problems, even a BSOD that says Invalid memory access, tried to write to read only memory. Hows that one for fun! And to get my b3 stable at 3.6 I needed a drooped vcore of 1.56. That was like 1.71250 in the bios, 1.64 in windows, and 1.56 under full small ftt load.
Yup, got the picts to prove it! Although on air, the cores are over 90c. I just wanted to see if I could find its breaking point with out... well, breaking it! Its very happy at 3220 and doesnt show much of a heat increase over stock, and the voltage is 1.45 in the bios, 1.40 dropped and 1.34 drooped. Is that right? Drop, droop, whatever! I have a huge one on this 780i, lots of weird lil bugs, but lots of features!!!
Thats a huge VDroop and drop! And the higher the voltage goes, the larger it gets! At the high settings, its total droopage was from 1.64 to 1.56! Thats huge, at 3770 my G0 (both q6600s) needs 1.44, drooped to 1.42! Thats small!
Anyways, thought ye all would get a kick out of it. Dont know where the free pic hosting is, or I would post em, if you wanna see, just ask!

--Lupi!

Sorry, thats on a lapped arctic square w/lapped processor with three fans on high blowing on it, lol, effectively lowered the temps by 3c, lol!!

Anyone else have huge test temps just to see what it can take?? Someone has to have some BALLberrings of steel, right??

--Lupi!

Not many people are willing to do that. We're poor, see?

Hehe, I Love it!