Need help and advice on my OC settings! (Again...)

[MUff1N]

I pretty much got the basics down but need more added info from someone here.
Here's the thing.
Whenever I go passed 3.4Ghz, I get it stable (at least it "seems" to be) at 3.46, 3.5, etc, at the lowest voltage settings I can and still pass an OCCT test, but after a day or two it becomes unstable because the chipset drivers become corrupted along with misilanious OS files.
Anyways, after I reinstall the chipset drivers & fix the system files (checkdisk) all is well again with the OC settings I had input.

What would cause this the most? NB or SB bridge voltage?
This will narrow it down some for me as to what needs a bit more juice so I don't have data corruption.

NOTE: Keep in mind the multi is locked on this CPU @ x10 so the FSB is a multi of that aka 3.4GHz is 340 FSB.
Thanks!

[ccperf721p]

The southbridge controls the hard drives.

Are you running the occt 1 hour test? For 24/7 use I like to see 8 hours OCCT / Prime95 stable. I prefer Prime because the test it fails on will help me determine whether it is memory or cpu that drops first

[MUff1N]

Originally Posted by ccperf721p

The southbridge controls the hard drives.

Are you running the occt 1 hour test? For 24/7 use I like to see 8 hours OCCT / Prime95 stable. I prefer Prime because the test it fails on will help me determine whether it is memory or cpu that drops first

So you're saying it would be the NB then?
But I"m running such low FSB compared to most out there but I guess it all depends on how everything together in the system as a whole comes together as to what is enough voltage to make the NB stable.
That's also the case on my memory as it stays below 900Mhz and they can clock to 992Mhz.

I usually run a custom 30 min test but I guess I need to run at least an hour or 2 to make sure my settings are stable for long term usage.

I"m no longer in a hurry so to speak to get good OC results.

Btw, at 1.450v vcore my temps are decent now...but it seems this still isn't enough for a 3.60-3.66GHz OC...rrr

[ccperf721p]

I have never seen the bios of the BX, but it should be southbridge voltage and not northbridge. All of the 975X boards I had would do 400 without having to raise voltage or maybe required a little bump.

I would also take a look at the memory, what voltage are you running to it at 850 4 4 4? 2.0 - 2.125 is pretty common on some of the older stuff to hold 4 4 4 timings after 850..

[MUff1N]

Originally Posted by ccperf721p

I have never seen the bios of the BX, but it should be southbridge voltage and not northbridge. All of the 975X boards I had would do 400 without having to raise voltage or maybe required a little bump.

I would also take a look at the memory, what voltage are you running to it at 850 4 4 4? 2.0 - 2.125 is pretty common on some of the older stuff to hold 4 4 4 timings after 850..

Awh...gottcha! Southbridge voltage increase it is...

These are the lowest stable voltage settings (I thought) for 3.46Ghz...

3.46Ghz Settings
Host Clock Frequency set to 346
CPU vcore set to 1.3875v with Enhanced Powerslope Enabled
Front Side Bus Voltage Override: 1.300v (1.250 is default)
MCH/ICH Voltage Override: 1.525v (Default, increase to what?)
Memory set to 4-4-4-12 CL @ 2.12v / Memory frequency: 667 / 1066 FSB Strap (FSB/DRAM 4:5) @ 875Mhz


What voltage would you recommend ccp on the Southbridge to keep my data from being corrupted??? (MCH/ICH)
Thanks forhand...
As of this morning (Mon 25th) I'm back to 3.46Ghz & increased the SB to 1.550v.
I ran Prime for an hour with no errors but who's to say I won't have them on a longer test.
I'm going to do that this afternoon when I take a nap.

Once I get this figured out, I'm going for the 3.66Ghz OC (1Ghz OC) on my CPU!
I reconfigured all my case fans to where I've got good cooling now & the case is Cold and with the Zalman on the CPU it can handle up to 1.4625v on the vcore and still be within
thermal specs...

I took out some of the fine metal mesh fan grills and replaced them with fan radial fan grills (pic below) to allow for more air throughput on the 2 80mm exhaust fans in the back of the case.
Then I added a high volume 120mm Scythe intake fan in the front that puts out 110 cfm & use the Zalman fanmate 2 control to control the speed.

I found out that after a period of time, the Zalman fan has gotten pretty quiet, so I wanted to see how loud it would be running at full speed connected directly to the CPU fan header.
Well guess what, it's not loud at all running at 2650 rpm! So this really cools the CPU good now!
You'd think all of this would be a bunch of fan noise, but without the mesh obstructions in the way, the fans got very quiet and let a ton more air in and out of the case!
The loudest fan in the lot is the Scythe 120mm fan, but that's at full speed!

I've already clocked my CPU to 3.67GHz to see but it wasn't very stable, but the temps were very good!
CPU-Z 3.67GHz OC Validation

[MUff1N]

I'm convinced now that I've got a not so good mobo, because it won't stay stable for me for any amount of time.
(1-2 weeks at a time with anything over 3.30GHz, currently back to 3.46GHz)

The only way I can make it stable again is to re-install my chipset drivers & bios each time which is a royal PITA.
I've done everything I can think of to get it stable and it just won't stick...& since pushing it some (Up to 3.67GHz) I've noticed when the bios will post it now takes 5 seconds for it to continue to Windows boot where as before it was 1 second so I think the bios chip is screwed up some & thats the reason it won't stay stable OCed for more than a couple of weeks at a time. (It's a permanently soldered in bios chip too)
Too bad because with this Zalman it will keep the CPU cool even with a full 1GHz OC on it. (Already tested it @ 3.67GHz to see but it wouldn't stay stable for nothin)

Of course I can't complain about the mobo & have it sent back via warranty because my warranty doesn't cover overclocking and that's the only time it gives me problems...
As a matter of a fact OCing voids my warranty so I can't say anything about it at all...

So I just have to wait it out until it either goes bad all together, or my warranty runs out & I'll just replace it with a better one like an Asus Rampage Extreme or P5NE.
By then the prices should be very low on these boards as the new mobos that support the new i7 Core processors will be out.
Then I can OC it a full GHz like I want it to be...

[richvanrooyen]

Hey Muff1n...

As you are aware, I have never been able to get this board stable at speeds above 3.4 ghz... I'm a little puzzled why you expect to get higher here...

I do have the Zalman also, but it just isn't stable. Are you able to hold at 3.4 ghz (that is stable)? (not 3.46 ghz)?

I think the point of overclocking (from what I have heard here) is to find the fastest, most stable level your mb and cpu will go... There isn't any promise that you will (a) be able to overclock the machine and (b) have a dramatic overclock. Try to keep that in perspective while you are doing this. I found myself getting frustrated and trying for higher and higher speeds... I would have probably blew my machine up if I hadn't thought about this or reset my perspectives.

I would sit at the stable level and count your blessings that you managed to overclock this fast! 3.3 ghz or 3.4 ghz is nothign to sneeze at!

[PorPorMe]

He's just a little anoyed that I hit 3.0Ghz so easy.

[MUff1N]

Originally Posted by richvanrooyen

Hey Muff1n...

As you are aware, I have never been able to get this board stable at speeds above 3.4 ghz... I'm a little puzzled why you expect to get higher here...

I do have the Zalman also, but it just isn't stable. Are you able to hold at 3.4 ghz (that is stable)? (not 3.46 ghz)?

I think the point of overclocking (from what I have heard here) is to find the fastest, most stable level your mb and cpu will go... There isn't any promise that you will (a) be able to overclock the machine and (b) have a dramatic overclock. Try to keep that in perspective while you are doing this. I found myself getting frustrated and trying for higher and higher speeds... I would have probably blew my machine up if I hadn't thought about this or reset my perspectives.

I would sit at the stable level and count your blessings that you managed to overclock this fast! 3.3 ghz or 3.4 ghz is nothign to sneeze at!

That's all well and good Rich but I want what I want!
This CPU can easily do a 1GHz OC (3.67GHz) so I want it!!!
If that means replacing the mobo to do it then that's what I'll end up doing after this one either goes out or the warranty is done.

No, 3.4GHz doesn't stay stable for more than a couple of weeks.
The only OC that is 100% stable for indefinite periods of time is 3.30GHz and that just won't do...hehe!

Right now I got it back up to 3.46GHz and for the time being it's stable.
As soon as it becomes unstable again I'll just reinstall the chipset & bios again to make it stable.

And No Porpor, I wasn't upset about your OC! As a matter of a fact I'm happy for ya!

[PorPorMe]

I know that, Muff!n. It's hard to do sarcastic humour in print!

[Daniel ~]

Really? ":O}

[richvanrooyen]

Hey Muff1n.... All the more power to you... I just need my machine to be slightly productive also (which kind of sucks).... Hope you get there and get stable!

[MUff1N]

Originally Posted by richvanrooyen

Hey Muff1n.... All the more power to you... I just need my machine to be slightly productive also (which kind of sucks).... Hope you get there and get stable!

And there-in lies the difference...This machine is used mostly for Gaming, so I want the most performance I can get out of it!

[MUff1N]

I "think" I may have found the magic solution to this board staying stable!
Since I have disabled C1E (Enhanced Halt State) it's been rock solid up to 3.5GHz! (3.5GHz I have stable)
Seems the vcore voltages were erratic with C1E enabled under load.
Now when you look at the charts during testing it's almost a perfectly straight line across where as before it was jumping all over the place & I think this was the problem.

I'm now (right now) currently testing @ 3.60GHz and getting close to having it stable...this board seems to like higher SB voltages.
15 mins so far on Orthos & no errors *fingers & toes crossed* & with any luck (& a bunch of patience) I'll have a new stable speed for my Zalman to actually work on to keep cool...

EDIT: As I was finishing up this post it Crashed.
No errors reported in Orthos, just a black screen & reboot!

Anyone have any ideas why it did this or what would cause this?
Seems strange that for 15 mins there's no errors then Wham-0!!!
Dead...

Any & ALL Suggestions & or Ideas are Welcome!
Voltage increase somewhere that's needed to try & stabilize this?

Settings used during the crash...

3.60Ghz Overclock Settings:
EIST: Enable
C1E: Disable
Host Clock Frequency set to 360
CPU vcore set to 1.4500v with Enhanced Powerslope Enabled
Front Side Bus Voltage Override: 1.250v
MCH/ICH Voltage Override: 1.675v
Memory set to 4-4-4-12 CL @ 2.16v / Memory frequency: 667 / Reference Frequency
266 (FSB/DRAM 4:5) @ 900Mhz

[MUff1N]

So nobody has any ideas or suggestions on how to fix this huh?
Look at above post~

[Daniel ~]

When it crashes...does it just quietly reboot without any messages or warnings? If so I've corrected this on my machines in the past by upping the V-core... when temps don't permit this... well I'm at the limit of what I know how to make it do.":O}

[MUff1N]

Thanks Daniel for at least trying to help me with this, but I'm sure it's just a Not so Good mobo that's the problem.

I always stress test my setting to make sure it's stable so voltage isn't the problem.

The problem is random. One time it will just silently reboot without saying anything at all, other times its a BSOD & endless reboot cycle.

So to fix this I'm going to have to replace the mobo.
As you have seen, a good mobo makes a world of difference & OCing relatively easy instead of a PITA Headache.