Overclocking E6400 - so far 3400mhz at 1.35Vcore- need advice to push to the limit

[RussianSensation]

My system comprises of:
- E6400 2133mhz
- 2gb Corsair XMS2x1024 5400C4 1.8V
- Gigabyte GA-965P-DS3 F7 bios
- ZM-NB47J northbridge
- Big Typhoon
- OCZ PowerStream 420 (30A on 12V rail)

MCH, FSB and CPU are all at stock voltages so far. RAM is +0.2V (or 2.0V effective) at 5-5-5-15

Prime 95 Blended test - 1 instance - 13 hours

I can't seem to figure out how to run 2 instances of prime 95? When I try to open 2 prime 95 executable files, the one I have open just flashes and nothing happens. I tried to set processor affinity to 0 or 1 but then it runs on one cpu or the other.

The maximum ram I have tested was 922mhz at 2.0V so I know my ram will easily let me go to 450FSB+.

I am hesitant to increase voltage prior to finding a stable overclock on stock voltages - and I want to run dual prime 95 instances.

Any help is appreciated.

 

[m21s]

Running Prime95 for Dual Cores

 

[RussianSensation]

Thanks M21, I'll try this out overnight and see if the system is stable with dual instances and return back with my results for some more overclocking goodness

 

[n7]

3200 MHz on defaults may not be very default.

Keep in mind many mobos auto increase vcore & the other voltages if you OC & leave the voltage settings auto.

I'd suggest manually settings things.

Also, you can quite safely up to at least 1.45V before i'd get too worried, as you'll likely not get more than 1.43V out of that with vdroop.

 

[n7]

Oh, & i don't see much point in effing around with P95, when there's a much easier alternative: SP2004 Orthos: http://sp2004.fre3.com/beta/beta2.htm

 

[RussianSensation]

Originally posted by: n7
Oh, & i don't see much point in effing around with P95, when there's a much easier alternative: SP2004 Orthos: http://sp2004.fre3.com/beta/beta2.htm

Thanks, i put on dual instances of prime95 since 7am today. I'll check the status today after work.

How long should I run Orthos for? Is it more stressful?

Yes, I did set the voltages to Auto. However, CPU-Z shows from 1.264 to 1.280 and Easy Tune 5 shows 1.300V, throughout stress testing at load (and same in idle). Should I not trust any of these numbers?

Before I used to have Motherboard Monitor 5.3.7.0 to check for Vcore and voltage rail stability on the PSU for my P4 2.6 system but that doesn't seem to be compatible with the DS3 motherboard as far as I can tell. I'll try setting it manually to 1.300V and see if the system is still stable.

I will also try 1.43-45 as you have recommended once I reach instability on stock voltages. Is it safe to increase MCH by +0.1 and FSB by +0.1v?

 

[harpoon84]

Originally posted by: RussianSensation

Originally posted by: n7
Oh, & i don't see much point in effing around with P95, when there's a much easier alternative: SP2004 Orthos: http://sp2004.fre3.com/beta/beta2.htm

Thanks, i put on dual instances of prime95 since 7am today. I'll check the status today after work.

How long should I run Orthos for? Is it more stressful?

Yes, I did set the voltages to Auto. However, CPU-Z shows from 1.264 to 1.280 and Easy Tune 5 shows 1.300V, throughout stress testing at load (and same in idle). Should I not trust any of these numbers?

Before I used to have Motherboard Monitor 5.3.7.0 to check for Vcore and voltage rail stability on the PSU for my P4 2.6 system but that doesn't seem to be compatible with the DS3 motherboard as far as I can tell. I'll try setting it manually to 1.300V and see if the system is still stable.

I will also try 1.43-45 as you have recommended once I reach instability on stock voltages. Is it safe to increase MCH by +0.1 and FSB by +0.1v?

Orthos is just an automated program to run two instances of prime - it's just more convenient I guess.

And yes, MCH +0.1V is fine, not sure what you mean by raising the FSB voltage, there is no such thing?!?

 

[RussianSensation]

Originally posted by: harpoon84

And yes, MCH +0.1V is fine, not sure what you mean by raising the FSB voltage, there is no such thing?!?

I think my BIOS actually has that. I upped MCH and FSB by 0.1V. Manually set CPU to 1.3500V

E6400 @ 3400mhz The temperature is at load running 2 BOINC SETI calculations.

Will try to go higher tomorrow. I'll check for you the exact name of the FSB voltage setting to make sure I didn't make a mistake.

 

[n7]

That's extremely nice considering how low your vcore is.

 

[MichaelD]

Originally posted by: RussianSensation

Originally posted by: harpoon84

And yes, MCH +0.1V is fine, not sure what you mean by raising the FSB voltage, there is no such thing?!?

I think my BIOS actually has that. I upped MCH and FSB by 0.1V. Manually set CPU to 1.3500V

E6400 @ 3400mhz The temperature is at load running 2 BOINC SETI calculations.

Will try to go higher tomorrow. I'll check for you the exact name of the FSB voltage setting to make sure I didn't make a mistake.

Wow! NICE OC! :thumbsup: Especially considering your low temps. If you can do dual P95 for 12 hours w/o errors, you're golden. I'm hoping for similar results...hopefully today, finally.

 

[smopoim86]

I got mine to 3.8 stable on water, was like 1.65 volt in bios though, 1.56 actual voltage. I don't think you could use that much voltage on air though

 

[RussianSensation]

Originally posted by: smopoim86
I got mine to 3.8 stable on water, was like 1.65 volt in bios though, 1.56 actual voltage. I don't think you could use that much voltage on air though

wow thats amazing. I don't think I'll attempt 1.65 on air cooling though like you said.

 

[smopoim86]

Do you have to have 1.35 to get to 3400? i can get to 3400 at 1.3

I still think i got lucky on my e6400.

 

[MichaelD]

I'm on air (AC Freezer 7) and I'm sitting at a boringly-normal 3.2GHz @ 1.320 volts. I just hard set the CPUv to the default VID which is 1.320v, set it to 8x400 and here I am. Easy. And lazy.

I haven't tried a lower voltage and I really don't care about a higher clockspeed...for now.

Aside from the standard "try all the voltages and see", what lower voltage do you think I would be Prime95 stable at?

I'm 17 hours stable at the current 1.320v; I'd call that stable.

 

[RussianSensation]

Originally posted by: MichaelD

I'm 17 hours stable at the current 1.320v; I'd call that stable.

That is most likely stable. However, I find it odd that sometimes prime 95 passes yet something like PCmark04 crashes say a Virus scan test. So I am not so confident as of late that prime95 is 100% bulletproof for checking stability.

I believe Intel spec still allows for up to 1.36. Therefore I do not see a significant reason for you to lower voltages even more, maybe to drop cpu temperature by a degree or 2.

 

[RussianSensation]

Originally posted by: smopoim86
Do you have to have 1.35 to get to 3400? i can get to 3400 at 1.3

I still think i got lucky on my e6400.

At 1.300 it would give me red error messages in BOINC and would abort calculations at times. I just raised it to 1.3500 right away without testing any of the voltages in between so I am not sure if it can work at say 1.325. I am going to try to get to 3.6ghz.

 

[MichaelD]

Originally posted by: RussianSensation

Originally posted by: MichaelD

I'm 17 hours stable at the current 1.320v; I'd call that stable.

That is most likely stable. However, I find it odd that sometimes prime 95 passes yet something like PCmark04 crashes say a Virus scan test. So I am not so confident as of late that prime95 is 100% bulletproof for checking stability.

I believe Intel spec still allows for up to 1.36. Therefore I do not see a significant reason for you to lower voltages even more, maybe to drop cpu temperature by a degree or 2.

Thanks, RussianSensation; I appreciate the reply. I agree with you; hours more of fiddling for a drop of maybe 2C isn't worth it to me.

I've run all the other benchmarks except PCMark 2005. Going to do that in a little while. It takes like 30 minutes...and I want to game.