my highest OC is getting lower and lower

[kool168]

my CPU could do 2.9 Ghz @ 1.55V and it was prime stable for more than 13 hours, after few weeks, my PC freezed at 2.9 Ghz.

then i reduce the clock to 2.8 Ghz @ 1.425V, prime stable for more than 13 hours. after couple days, my PC freeze again and of coz prime 95 failed within 30 mins. then i up the voltage to 1.475V, it was fine and passed prime95, after few days, it is not stable again, but i up the voltage to 1.5V, now it works again.

why my highest OC is getting lower, and i need to add more voltage all the time?
i do folding all the time. i am on water, load temp never excess 50 degree, and i never take it to 1.6V, so what's causing that? is that folding thing slowly killing my CPU? or other hardware problems causing that?

there is something wrong with my video card, and i am waiting for the rma replacement, but i doubt the video card could affect my cpu...
checked the PSU, it has no problem.

[Absolute_0]

How did you check the PSU? Inconsistent power is what i started thinking, because maximum OC shouldn't be decreasing so rapidly unless you run dangerous volts, which you aren't.

Also, it is summer, have you been checking up on your ambient temperatures?

[kool168]

How did you check the PSU? Inconsistent power is what i started thinking, because maximum OC shouldn't be decreasing so rapidly unless you run dangerous volts, which you aren't.

Also, it is summer, have you been checking up on your ambient temperatures?

most of the time the PC stopped responsing while i was sleeping, so i have no idea what was the ambient temp. while i am beside the PC, with both folding programs running, now the CPU load temp is 44 degree during load.

werid thing is why 1.425V could do 2.8 Ghz few days ago, but now i need 1.5V?

i am worry about if the situation continues and i keep adding more voltage to stay at this clock and my CPU will finally die some day.

[SoF]

maybe it's not your CPU, maybe it's a DFI board?
Strange that I only read such stories with ppl on DFI boards, I'm that lucky with my gigabyte :D

[Kain XS]

Many CPUs and memory suffer Electromigration.

Thats why it starts to getting more difficult to reach a better score (Uusually happens more when memory is dying).

They eventually loose their habilities and in some time theyre not stable even in stock.
Jope its not that what your suffering but maybe a faulty PSU or maybe a bad motherboard (which is not using the rails withing time to time).

For more info. ( I made a guide of Electromigration... but its in Spanish) read here: http://en.wikipedia.org/wiki/Electromigration

[J-Mag]

it's also getting hotter during the summer, so unless you have climate controlled perfectly... (for all i know you are in australia though)

Edit: Looks like you updated your profile. Canadian, eh?

[kool168]

so i asked the tech support on DFI and here is their response:

No motherboard company warrants any overclock or the pushing of your parts beyond their specification. In fact in tech support we are really told to leave the overclocking to the customer and his responsibility and what he learns or does not learn from others who wish to assume the responsibility of pushing components beyond specification.

Personally I would guess your cpu memory controller is losing functionality from having run 2900Mhz and not the 1800Mhz it was designed to do.

what do you guys think about his response? is that true my cpu is dieing becoz i push it too far? but i have seen people with my stepping hitting 2.8 Ghz average... why my cpu would die becoz i clock it to 2.9Ghz with only 1.55V?


it's also getting hotter during the summer, so unless you have climate controlled perfectly... (for all i know you are in australia though)

i know heat wasn't an issue at all becoz i am on water. the PC wasn't freeze resulted from high temp, it needed up the voltage constantly to keep the same clock.

[einCe]

mem controller dying seams feasable. happened onmy 3000 venice

[J-Mag]

i know heat wasn't an issue at all becoz i am on water. the PC wasn't freeze resulted from high temp, it needed up the voltage constantly to keep the same clock.

Yeah well your ambient is higher and water can never get your chip cooler than ambient... Temps will go up in the summer no matter if you use water or air.

[Kain XS]

so i asked the tech support on DFI and here is their response:

No motherboard company warrants any overclock or the pushing of your parts beyond their specification. In fact in tech support we are really told to leave the overclocking to the customer and his responsibility and what he learns or does not learn from others who wish to assume the responsibility of pushing components beyond specification.

Personally I would guess your cpu memory controller is losing functionality from having run 2900Mhz and not the 1800Mhz it was designed to do.

what do you guys think about his response? is that true my cpu is dieing becoz i push it too far? but i have seen people with my stepping hitting 2.8 Ghz average... why my cpu would die becoz i clock it to 2.9Ghz with only 1.55V?



i know heat wasn't an issue at all becoz i am on water. the PC wasn't freeze resulted from high temp, it needed up the voltage constantly to keep the same clock.

There is a little posibility.
I have never seen a processor working like that (dying like that) but it could be the memory controller due to Electromigration.
Read the post i wrote above. Its the same DFI technician told you but better explained...

[sealion]

maybe it's not your CPU, maybe it's a DFI board?
Strange that I only read such stories with ppl on DFI boards, I'm that lucky with my gigabyte :D


You find more problems because more people have DFI opposed to the few who own gigabyte. Say there are 3000 of us DFI users and say 50 have problems... it's only a 1.6% defective rate. Even that is stretching it.

Update: fixed the math


Hey I use to get 3ghz with 1.52vcore then made a dumb mistake in reseating the heatsink. Oddly enough the temps were almost the same but I could never get 3ghz stable ever again. Adding more voltage just adds to the heat but the stability never increased. Recently put back on my xp90 and now 2.9ghz 1.45 is possible.

[naelq]

i would say it's due to amb./CPU temps ;)
IMHO, a one should look for Summer OC & Winter ones! :D

[cky2k6]

sorry to be a stickler, but 50/3000 is 1.6%, not .016% i agree though, obviously more problems are reported with dfi since more of us use them.

[kool168]

i would say it's due to amb./CPU temps ;)
IMHO, a one should look for Summer OC & Winter ones! :D


if it's heat issues, then why upping the voltage and it becomes stable again? now there is no way i can run 2.8 Ghz with 1.425V, but it was possible few days ago?

[naelq]

more heat > unstable OC > more voltage to get it stable ;)

[lross78550]

We Clock um, we cook um then we buy some more and play again. Welcome to the hobby. We are sorry for your loss, but sounds like its time for a trip to newegg.... maybe a conroe is in your future :)
Lee

[kool168]

more heat > unstable OC > more voltage to get it stable ;)

anyone can second that?
if it is true, then i am extremely happy since i can give up my shopping on FX-60/62 for comparable performance.

[fullup3]

I also think you are losing oc due to summer ambient temps. My processor will not run as high as in the winter and spring. I was running mine at 3000mhz with 1.48v 24/7, and now i'm at 2800 w/1.32v due to the heat.
I am also on water.

just a thought.