Sign in with
Sign up | Sign in
Your question
Solved

Occt vs. prime95 vs intel burn test vs y cruncher

Last response: in Overclocking
Share
December 23, 2012 6:38:56 PM

Which test is better to see stability and why?

Im having trouble keeping my i7 3770k 4.6ghz stable at 1.275v - 1.29v in all programs....
well my 4.6ghz oc at 1.28v would be stable with 2 hours on occt and pass intel burn test but when I run prime95 v27.9 after 30min the program would lock out and stop.

2 out of 3 stress test means its stable?

a c 78 à CPUs
a b K Overclocking
December 24, 2012 1:23:36 AM

If any stress test (or other program) shows instability, it's unstable.
m
0
l

Best solution

a b à CPUs
a b K Overclocking
December 24, 2012 5:18:39 AM

My method.....


Set my overclock and get it to boot into windows.

First test with IBT for a few rounds. Ten on normal settings and then put the stress level up and right click the start button on IBT and select eXtreme stress level. This is where I find out the highest temperatures my processor gets. If I manage to pass the IBT tests; then I'm happy.

Once I get it to pass a few different arrays of IBT tests. I'll move onto a few hours of Prime95. If it lasts for two hours on Prime95 I'll schedule in a 25 hour run of Prime95. If it passes 25 hours of prime95; it's stable, almost guaranteed to be stable.

IBT is good at finding major instabilities while prime95 is good at finding the "little" instabilities that IBT can't bring to the surface. Sometimes it just takes 24 hours of extreme math problems to find out if your processor is stable to the T or not.
Share
Related resources
December 24, 2012 7:11:53 AM

Best answer selected by geekdgreen.
m
0
l
December 24, 2012 7:48:19 AM

Yeah ive been following the same method...ive done 2 hours of OCCT with linpack AVX capable checked, then 20 passes of Extreme IBT. I guess the most recent build of Prime95 is the issue (27.9). Right now im testing my different overclocks on 3 different builds the 26.6, 27.7, and 27.9. I dont know if anyone else is having same issue with the 27.9 version of prime.

Also whats the stock voltage of the cpu pll and VTT cause i think my MSI z77 mpower is automatically overvolting the cpu pll and undervolting the VTT settings. Ive read the the cpu pll should be at 1.5 and the VTT at 1.11 but my board has it at 1.8 for the cpu pll and 1.05 for the VTT.

Can anybody explain how ram effects a cpu overclock. I know it affects the base clock overclocking so it wouldnt be an issue of ivy bridge but for amd fx. But ive realized when I RMA'D my patriot ram and started using these cheap super talent 8gb ram, trying to find either overclock on my desired voltages are harder to get.
m
0
l
a b à CPUs
a b K Overclocking
December 24, 2012 10:18:35 AM

With keeping the BCLK at the stock 100, you're not overclocking the ram unless you do it in the ram's settings. Adjusting the BCLK is just like adjusting the FSB on an oldschool rig. However, most of today's chipsets don't take kindly to such boosts like they did back in the day.

Also, I'm believe PLL should be around 1.8 and the VTT at 1.05 Mine are like this and they are on stock settings

CPU PLL 1.832
VVT 1.051v

All that said, I've been stable since day 1 of my overclock and haven't been happier. Honestly for types of memory? I don't see it being much of an issue unless you've got the timings off or the ram is just, bad.
m
0
l
December 24, 2012 1:20:28 PM

Yea I know I shouldnt touch the BCLK on ivy bridge cpu's...i only messed with the fsb on my previous 8120fx chip. Well the current ram im using seems to be a joke but does get the job done...Patriot G2series are getting a lot of heat for dead stick of ram on newegg and I just wish i could get my Replacements in soon cause with the switch of ram my cinebench score dropped at-least .20 points and that's ridiculous.

So far ive realized the newest build of prime95 are causing issues. Ive been running the one that was released sometime around may (v27.7) for 4 hours and going strong.
Just wish i couldve got a 4.6ghz at 1.27v instead of 1.29v.
m
0
l
!