Anyone added CPU-NB voltage or bumped up the NB speed during your overclocking trials? Changing them both seems to play no part in gaining stability at all

Was trying to get Large FFT prime stable on my PII 550 BE at 3.8GHz, at about 1.465v vCore, everything else stock. Couldn't get it stable for some reason till I ran my RAMs at 2.0v.

Happily running at 3.8GHz now, but without any cores unlocked. What terrible luck.

Just did, much thanks if you can help me out

I've been poking around the settings already for two days straight, feel like giving up hope to get extra cores already.

The first thing I did was to update to the modded BIOS and set ACC to Auto. Core Levelling is at Auto as well. Managed to boot into Windows with four cores, but Worker #2 would immediately fail after running Prime95. The other three would continue running. Upped the vCore, no difference. Shouldn't it be Worker #3 or #4 that should fail?

Hmm, someone told me that in Prime95, Worker #1 will always be allocated to Core 1, so on and so forth, so which I assumed if Worker #2 fails, it is because of the second core. But then again, why is the second one failing when it should be working correctly?

Also, can AOD individually disable or enable cores?

Ah, never thought about actually disabling the core from Windows. The BIOS doesn't allow disabling of individual cores, however I could test it by setting the processor affinities in the Task Manager. I'll try your method as well and see if it works.

That's what I'm guessing as well. Initially I assumed that the working cores will always go in the correct order, Core #0 and Core #1 are the ones that will always work, and Core #3 and Core #4 would be the faulty ones (in my case). From this example it's quite probable that it's not the case. It also doesn't make sense that AMD would have to dump the chip if the first two cores didn't work when binning for an X2

Will let you know the results later in the evening. Thanks again

Finally some progress. I tried disabling the problematic core by changing the processor affinity and Prime95 now gives no problems. Changing the number of processors in MSCONFIG doesn't do any change, though as the problematic core is the second one (changing the number of processors to 3 still includes the affected processor).

Are there any manual ways to disable the problematic core via Windows, or to automatically set processor affinities?

This post has been edited by colinwong89: Aug 3 2009, 02:18 PM

My proc gives out after 90 passes of LinX. Heat is under 60C, and upping the vcore doesn't solve the problem. What else can be done to make it stable?

My humble OC...



In the middle of another LinX run, increased the NB to 2400MHz for testing.

No probs kuntawakaw, any advice is appreciated

So far that's my setting to get Prime stable, haven't tested going lower yet. I also heard that Phenom IIs don't really like high temperatures, so I plan on cutting the voltage down right after I get my overclocks certain. Or maybe you're just more lucky than I am

So far, done 56 LinX passes at the toughest settings possible, and I'm aiming for at least 100 passes. Sounds excessive no doubt, but I'm the kind of person which stability has no compromise

I want at least an X3 as well, I'm sure that my PC can use the third core but the BIOS shuts off the good one instead of the defective. Searched the whole day for info but nothing at all

edit: Why doesn't CPU-Z detect your processor model correctly?

This post has been edited by colinwong89: Aug 5 2009, 03:03 AM

Overclocking my NB speed (2200 to 2600) improved my SuperPI by half a second. Try reading Dolk's Research, the helpful tip posted by antonio

LOL, it's up to you to find out which of his advice can be taken seriously

That SG fella with the 0925 is really something. All their 0925s have all the luck, and mine can't even get the extra cores stable let alone pass 3.8GHz with ease

Does anyone know what is Core DID setting in the BIOS? The settings is either 'Divided by 1' or 'Divided by 2'.

Those with Biostar boards, does anyone use the G.P.U Phase Control when overclocking? It's that option where is limits the phases to save power, and I just realised I've left it on throughout all my testing.

Will it affect overclocking or whatsoever?

I used to have lockups on Large FFT tests, but seems now like I'm not having it anymore. And all the while I was trying something else to fix the problem...

No harm done, I myself forgot and missed out on that little option Will run more tests later on during the day.

Ganged mode, in short uses a 128-bit path controlled by a single controller which is good for single applications, where as unganged mode uses two 64-bit paths controlled by two controllers, which works better for multitasking and such.

As time passes more and more applications are going to be multi-threaded, so going unganged should be better in the long run

I'd give it a few weeks before something smokes

Try setting your command rate to 2T. For me I could not see any difference in putting 1T or 2T, but 2T helped in increasing clock speeds.

How about the Core Contact Freezer? It's at about RM155 by our forumers, and it's as good (if not better) than the CM V8.