Using Advanced Clock Calibration (ACC) to increase Phenom ii overclock

It appears that ACC might be helpful to overclocking the 955

http://www.bit-tech.net/hardware/cpus/2009/05/11/overclocking-amd-s-phenom-ii-x3-720-be/1

http://www.overclock.net/t/547258/any-idea-what-acc-advanced-clock-calibration-do-to-phenom-ii-x4/0_100

I'll add more info as I scour the internet, but it seems more than one individual found ACC +2%+ helped them out. Who knows how reliable that information is though, is that for prime95 12+ hours stable, or just booting, but that's what this discussion is for.

On a side note, I was curious if just increasing CPU-NB speed would help cpu stability, as well as increasing CPU-NB voltage. I seem to see a few people saying that, but I haven't seen anyone actually report that it helped, they just suggest it without having done it themselves.
Edited by Belial - 11/20/12 at 4:48pm

I am monitoring VRM temps on the hottest 2 phases, and I make sure they don't go above 85*C. I am currently downvolting, since it seems like heat is causing instability (prime95 fails around 5-8 hours if core is at 55*C or more the entire time... i think, not sure, maybe chip just isn't stable at 3.6ghz+).

I've talked to Biostar about VRM temps as well. My VRM temps are fine.

What? How is this possible? Do motherboards really affect overclocks? I thought it was just a matter of "hey your motherboard will either blow up at that voltage due to the VRM temps/quality or it won't" and FSB limit.

What revision were you on? Saying 955 doesn't really say anything to me since c2 and c3 are basically 2 totally different processors.

And I've found a million 'side' factors to be causes for overclock limitations as much as anything else. Like I couldn't boot even with 1.55VID and 1.55vcore on 3.6ghz+, but I can boot easily on 3.6ghz+ on 1.4vcore with 1.25VID and +2 cpu over voltage, for some reason I have to have over voltage to overclock and to reach 3.6ghz+, and I also couldn't POST if I manually set my HT link speed, no matter if it was 1.6ghz, 1.8ghz, or 2ghz (this is with no FSB tinkering btw), but if I set HT link speed to auto, i could post with any speed.

I'm currently testing ACC on auto, 3.7ghz@1.488vcore, and I'm stable so far on like 3 hours. This same system was failing 3.6ghz at 1.520vcore with ACC disabled. But we'll see if it matters after 8 hours, since my overclocks have mostly been failing around 4-7 hours, so maybe i'll fail in 3 hours. I also think temps might be way I was failing too, i think with long periods of time on 55*C+ causes crashes after 5+ hours, since while I was crashing on lower vcore, it seemed like I was crashing around the same time no matter if my vcore was 1.568 or 1.488.
Edited by Belial - 11/9/12 at 6:26pm

^ I don't understand. And won't that only matter at higher voltages? I'm going for a 24/7 overclock here, not a benchmark, and since I'm on air, I won't be pushing more than 1.568v at the most, although it seems like I'll probably settle for 1.488-1.504vcore.

Yes, I'm totally aware of this. I don't understand how phases relate to actual max overclockability for the same chip.

My understanding of motherboards is:

1. It has the right holes and sizes (atx, am3 socket, pciex16 x 2, 10/100/1000 lan port, etc)
2. VRM is high enough quality so you don't blow out your VRMs when overclocking (or worse, at stock) see: msi motherboards
3. Maximum FSB (which isn't a big deal, you can increase chipset voltage and ~250 on a low end board with a locked multiplier should allow enough configurations to reach the perfect system overclock)
4. Actual chipsets (not a big deal, even cheap budget boards tend to support overclocking and have the features you need, which is generally overclockability and discrete GPU support)

I'm aware of what phases are, what VRMs are, how they work, what the components are. I was confused by the claim that my max overclock limit was somehow limited by my motherboard, as if getting a higher end motherboard would increase my clock on the exact same CPU. Never mind that such a claim makes no sense either anyways, what exactly about a different motherboard would allow lower or higher overclocks?

Now I get blowing out your VRMs is a very real concern, but at the moment I do not intend to go over 1.55vcore because of the VRM temps. I have found that at 1.568 vcore, my VRM gets to 90*C+, which with researh and talking to Biostar, is okay for a few times but will lead to PCB discoloration over long term usage, but my CPU temps are a limiting factor before VRM temps, and I'm going to be settling around 1.48-1.5 vcore anyways, where VRM temps hover around 60-70*C instead, due to 'the heat wall'.

Im aware of the temp and voltage limits of a pii c2.

Thanks for your input but i had no confusion on this and there's nothing in my op that would suggest that ;/

I think you got it backwards. Higher Vcore = higher VRM temps.

Im running a c2 955. c2 and c3 chips are two totally different chips, it's comparing apples and oranges. I'm aware heat is a limiting factor with phenom ii x4 955 c2 chips. I'm asking if ACC can be used to help system stability.

I'm running 1.488vcore right now. I'm struggling to make 3.6ghz work on that. With ACC, I think I might be able to make 3.7+ ghz work on it. I also think voltage isn't my limiting factor, and maybe lowering voltage might help my overclock stability (or moreso than raising it).

So far I've run prime95 for 4 hours on prime95 at 3.7ghz at 1.488vcore. It looks promising. Will have to leave it on overnight to know for sure though.



http://www.overclock.net/t/959145/jerryrigging-mofset-heatsinks/0_100#post_12741129

This is my system.