My first OC: Intel i7 920 D0 @3.8ghz. Not entirely convinced. Opinions please

[Maegirom]

Hello, sorry for my bad english.
I have just done my first OC and I would like you to tell me your impresions.

My rig:

Intel i7 920 D0
mobo: Asus P6x58D premium
RAM: G.Skill Ripjaws 3x4gb (tri channel) PC12800 1600mhz CL 9 9 9 24
cpu fan: Noctua NH-D14
Corsair AX860 psu
Asus GTX680 direct cu2 top & Gigabyte GTX680 windforce oc in SLI

I've done a 3.8 ghz OC with the following information :
BCLK 200 , multi x19 ( 3.8GHz ) . The memories are left to their default speed ( 1600mhz )
V core: 1.225 in BIOS but CPU -Z in windows shows me 1,216V ( I'll talk about this later)
Frecuency Uncore 3208mhz ( twice the speed of RAM)
QPI frecuency : AUTO
VTT ( QPI / DRAM voltage) : 1.22V
PLL : 1.8v (default )
IOH : 1.18v I've raised a little on the subject of SLI
ICH : 1.10V (default )
VRAM : 1.5V ( recommended vRAM for my model )
CPU differencial amplitube : I know it's said it's okay to put 800mV but I left AUTO (off, I think), because I have also activated the LLC (LOAD LINE CALIBRATION)
Energy saving is enabled intel speedstep and C1E . The C -State is off because I realized that I had it turned off before the OC process, and I ran so for years . So better to leave well enough alone ...

TEMPERATURES: (measured with coretemp and AIDA64)
In idle cores give me about 34 º and Tcase ( sheath temp ) about 26-27 º C

After 4 hours of stress with prime95 in blend ( round checking off) and 4 hours SmallFFT mode, the highest peak has been given by the core 1 at 72 , But it has been a blip, cores have been pulled about 68 º C almost the entire process
The TCASE at full load has 52-53 º C

One strange thing: Skyrim pulls me the core temp up to 65/66ºC oO' How is this possible that skyrim pulls the cores almost as prime95???

NOW , CONSIDERATIONS .

First, voltage. I guess it has done Vdrop because I put 1.225v in bios and both cpu-z and AIDA64 gave me 1,216 in windows. Which seems odd since it was activated LLC. Its function is just correct the Vdrop, right? And is this a real gap , or just a reading failure? I mean, was it 1.225V but it was reading 1,216v or it was actually 1,216v ?

In addition , the voltage is the same all the time in both idle and full load . I thought that it would lower voltage in idle due to speedstep and C1E . (So it was before overclock, had less voltage at idle than full)
The clock frequency fluctuates right. It's at 2.4ghz in idle and 3.8GHz in full load , but the voltage does not change. Is this ok???

What do you think about these temps? Can I improve it somehow? actually these are the first values I've tried , I have not tried yet to reduce vcore or anything. Do you think I can set an even low vcore and will this lower the temperature down ? Are these safe temperatures for 24/ 7 for a long life (years) operation? And when summer arrives, will I have to squeeze a little less ?

Furthermore, the noctua FANS , gave me 1430 rpm constant in both idle and full. But I think this is already the 100% so it can't go faster I guess


Well, that's all (I hope) sorry for the long speech. Give me your opinions pleaseeeeee .
Thanks


Thank you.

 

[Brantyn Gerik]

Skyrim is a cpu intensive fans. Your temps are actually really good so i wouldnt worry about them

 

[Maegirom]

Any idead about the vdrop? I have LLC activated so it shouldn't have any vdrop.
Do you think I can lower vcore even more? let's say 1.20V or so?

PD: I have experimented some hangs playing (No BSOD, just hangs or CTD). Maybe a low IOH voltage? Remember I'm running 2 gtx680 in SLI. Now I have it at 1.18v. Maybe should I rise a bit, let's say 1.2v? I have read this is safe, right?

Thanks.

 

[Brantyn Gerik]

.02 change in voltage is just a software error, if you wanted accurate readings you would use a volt monitor and a motherboard with v-checkpoints. 1.18 vs 1.2 will be ever so slightly more stable but really think about how miniscule the increase in power is.

also that processor is generally safe up to about 1.3, whereas anything after that requires serious cooling and can cause damage to the processor.

another thing to check which i didn't see from above is your power supply (i could just be blind) the 680's require 42 amps on the 12v rail to be used at max load effectively, and especially when overclocking. This could very well be the source of the "hangs" you are experiencing.

If you have a very high quality psu (850 watt and good quality) then with your temps i would suggest raising the voltage on your cpu by another .1 or .2 This will only increase temps moderatly (max maybe 3 degrees under load) but will improve stability.
same goes for your graphics cards. increasing the voltages only minisculy may solve the hanging problem at the expense of some heat. And generally with the stock bios, overclocking programs like msi afterburner will only let you go so far on the voltage as to keep you within safe operating limits.

 

[Brantyn Gerik]

Typically when you are going for stability with an overclock, you increase the voltages to increase stability. You get higher temps this way, but the overclock becomes more stable.

 

[Brantyn Gerik]

At a certain point however, you get MASSIVE gains in temp for minimal gains in performance, and you often also run into a max overclock (processor simply wont overclock any higher even if you are under the thermal limit)

 

[Maegirom]

Brantyn Gerik, my psu is a Corsair AX860 (860W). More than enough for my gtx680's SLI.
I think the problem (little problem, just 2 or 3 hangs in several days...) is not in vcore. If so, then I wouldn't have passed 8 hours of prime95 with no errors. The thing I say is maybe that IOH voltage is low? (IOH manages northbridge and PCIe- hence GPU- comunication). Mine is at 1.18V. DOn't know if this is too low for the SLI, although default is 1.10
thanks

 

[Brantyn Gerik]

Typically most people don't even need to change that, so highly doublt IOH voltage is low. honestly everything about your system is completely perfect aside from these hangs. I think raising IOH voltage to 1.2 would be a good start, and maybe raise the voltage ever so slightly on the cards themselves to see if this stops the problems

 

[Maegirom]

Ok thanks. Actually I have not had any problems or hangs for a several game sessions. Maybe it was just a casual or random game teasing...
You know, Skyrim has crappy behaviour...

Thanks to everybody
Regards

 

[Brantyn Gerik]

Skyrim has horrid behaviour, especially after modding XD

 

[Maegirom]

Hi, I'm refloating this thread since today, my pc has not started up...

I have been running perfectly during 2 months with this OC, but this morning I had the "overclock failed" message when turning my pc on.

I have increased a bit (the minimum allowable increase) vcore and vtt and it seems it works right now, but my cores have gained 2ºC.

The thing that bothers me is Why the the hell my pc doesn't star up with an OC that has been stable during 2 months???

May it be that the components have deteriorated during these 2 months to the point of needing more voltage?

Thanks