• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

E8500 OC results

S

Samury

Limp Gawd
Joined
Feb 7, 2006
Messages
199
I've just started successfully OCing, so I thought I'd share my results and seek out some tips.

I'm using the system in my sig (CPU stepping A, Revision E0) along with a Scythe Zipang cooler.

Currently I have it set to:

400 MHz FSB
1.325 V vcore (showing 1.296V in CPU-Z and 1.30 V in Hardware Monitor)
mult 9.5
or 3.8 GHz

Core temps:
41 C/ 38 C min
56 C/55 C max under Prime 95

So far everything seems stable, though I haven't run Prime 95 overnight, but it's been running for over 25 minutes w/o a problem (1024k test passed completely).

When I tried upping it to 420 MHz at up to 1.35 V (moving incrementally on the voltage) Prime 95 crapped out during the first or second 1024k tests. 410 seemed stable, so I moved to 420 and have had no luck yet. Perhaps it's just more volts?

Ideally I'd like to get it over 420 MHz FSB, ideally 450 MHz, so are there any suggestions? Is there any sort of sense to FSB/vcore ratios that people have found?

I had tried using EasyTune6 to bump it to 450 MHz and fiddled with the vcore around 1.4 volts and several settings below, but that didn't work either, so maybe it's just a useless endeavor to get that high a clock speed (4275), but I've read around and 4.0+ should be fairly easy. Maybe I should start at 420 MHz and 1.4 V and work my way down until I find something stable?

I'd prefer to stick with only the FSB and vcore, while modifying the RAM to stay as near to 1066 MHz as possible, preferably below; maybe even further below, closer to 800, would be more stable?
 
Update: Core 1 failed after 36 minutes with a rounding error in P95. M14155775, 768k length.

Core 2 is still going.

Guess I need more voltage.
 
you may need to adjust the MCH v, try 1.22v to start and then start messing with mch reference. For duallies people like .74 or .70-.68 when going higher.

Also your cpu/board may like cpu reference of .785

Make sure that your PL(static tRead) is not holding you back, so set to 10 for now
 
SpeedEuphoria said:
you may need to adjust the MCH v, try 1.22v to start and then start messing with mch reference. For duallies people like .74 or .70-.68 when going higher.

Also your cpu/board may like cpu reference of .785

Make sure that your PL(static tRead) is not holding you back, so set to 10 for now
Click to expand...

Wow, big thanks.

I did everything you said, mch ref at .70, vcore at 1.3625, and FSB at 430 and was completely stable in Prime 95 for 14.5 hours. Max temp topped out at 60 on both cores.

I am one happy overclocker.
 
You can prob lower the Vcore some, I assume your talking about 1.3625 in bios. Try like 1.325v see if it works. Or try 450FSB, but you may need more MCH voltage to do it, like 1.30v MCH

For 1.3625v in bios and LLC enabled you should be able to get 4.2-4.3Ghz.
 
After checking in my bios, I noticed that the MCH ref was up, as well as the CPU ref. This was after I upped the FSB to 440. Would that have done it, or did the system auto-adjust it at some point?
 
Those are upped when Cpu term and MHC Voltage is upped if i recall, Also, Devide em and make sure their higher than 63 & 69 percent, i dont recall which is for what, but im sure someone will correct me and say it.
 
No its not an issue its just something you need to understand, goes like this:

The MCH and CPU references automatically change when you change VTT(cpu termination). Its trying to keep the ratio that you know should work. Now ET6 will allways report the references as if VTT is on 1.2v, so most with this board talk about the references and refer to what ET6 shows or using 1.2v vtt.

So what you do is leave vtt at 1.2, then adjust the references to whatever, like what i suggested(.785v cpu and .70mch), then go back and raise VTT to what you want. Notice it will change the reference voltages automatically.

Also on this note, all the dram termination/references will change automatically to 1/2 of what you set for the dram voltage. So if you change dram voltage all of those references will change automatically. You wont notice if you have it set to "auto" but if you manually set the refereence voltages you will see them change in bios.

Now you have a better understanding of how the bios alters things and will understand what your seeing.

Good Luck
 
Ah, thanks for the info. I had set it all in the bios, but that makes more sense. Upping the speed didn't work very well, so now I'm trying 430 with a lower vcore.
 
I got the voltage down to 1.35. Below that and it BSOD'd in P95.

Any tips on how to get it below that? Or is that just how this particular chip will work?
 
did you try different cpu reference? and or VTT? You can try running small FFT in P95 to try and isolate the issue to cpu or mem/MCH. Small will put more stress on the cpu, so if it fails that its likely one of the issues I say just above or thats the limit and you will need more Vcore to go further

How long till it BSOD'd?

If it fail's blend quickly then Try and run memtest and loop test #7, see if you get errors. If so then mch ref needs to be adjusted and/or mch voltage needs to be raised.
 
No, I haven't adjusted the CPU ref or VTT since I didn't know which way would be better, if it matters.

The BSOD happened at various times, but probably after at least a half hour. The last one was while I was asleep, so no idea.

Should MCH be raised or lowered? I assume raised, but want to be sure before I make any changes.

PS - thanks again for all the help. You've been much more concise than all the guides I've read.
 
New Update:

The most recent pass was with the large FFT test in P95, which was fine.

However, the Blend test failed after a few hours.
 
You must log in or register to reply here.
Back
Top