NEW MMO-Champion CPU Overclocking Leaderboards

[Asera]

You can probably take me off the list, I have a 3770 now, non K variant, thus cannot overclock it. :P

[DeltrusDisc]

Nope, once you're on, you're on! Just because you replaced the CPU doesn't mean we can't leave it up there as a reference!

[Gravewyrm]

@DeltrusDisc The TjMax for the 6800K is 74c according to a few places ive found around the web. Core Temp (which shows absolutely NO temp readings for the chip, yet tells me the TjMax) says its @ 80c. the 80c would correlate to the insta-shutdown I experienced on my second go at a stability test. Spiked to 87c /insta off.

The first spike it did (84c in AIDA64 Extreme) the system stayed up, but I killed the test within 5 seconds of the spike. In both cases, the voltage went WAY over what I had set to max. Both instances, max voltage was set to 1.512v. 1.574 is considered the ultimate max to go, as it has been proven stable via P95 @ 5.0Ghz by a review I found. I'll see if I can dig it up again. The spike voltage hit 1.620. Way out of my comfort range.

I've tried everything to get the voltage locked down, and not fluctuate, but its impossible. It can flux a tiny amount (0.006v) to insane amounts (0.7xx+).
With it currently underclocked, its maxes at 1.272v, idles at 0.868v (+/- 0.012v).

[Asera]

Nope, once you're on, you're on! Just because you replaced the CPU doesn't mean we can't leave it up there as a reference!

My 2500K is eternal!

[DeltrusDisc]

My 2500K is eternal!

So long as mmo-champion remains, so too does your 2500K. :P

[Gravewyrm]

Figured out some tweaks, and got her back up to 4.7. Screenies here --> http://imgur.com/a/e34xF
Funny note here, once I opened Photoshop and saved the 2 screenshots, I closed PS, and CPU-Z said I was dropped to 20x100Mhz, and voltage was stuck at 1.416v. It never went over 1.404v within that 30 mins of testing.

[DeltrusDisc]

Figured out some tweaks, and got her back up to 4.7. Screenies here --> http://imgur.com/a/e34xF
Funny note here, once I opened Photoshop and saved the 2 screenshots, I closed PS, and CPU-Z said I was dropped to 20x100Mhz, and voltage was stuck at 1.416v. It never went over 1.404v within that 30 mins of testing.

How very weird. So if I recall correctly, Prime95 and AIDA64 are both problematic for you?

[Gravewyrm]

How very weird. So if I recall correctly, Prime95 and AIDA64 are both problematic for you?

P95 was only a prob at 5.2Ghz. But that was most likely due to needing more voltage, and I refused to go higher.
A64 seems to only give me issues past say, 2.5hrs on average. Ill be setting up that 4.7 test for P95 tonight. If it makes it to 8hrs, I'll then try A64. Planning on starting the test around 3AM (CST) or so.

- - - Updated - - -

Managed to get stuff done quicker than I expected, and started up the 4.7Ghz test with p95. Lost a worker in 20 mins, so bumped the voltage a bit, and was solid for 4 hours. Rebooted after stopping the test and giving everything a 10min cool down. Booted up, loaded up A64... 10mins in dropped the test on me... backed off to 4.6, no voltage change...35 mins /crater. Backed off to 4.4, still same voltage.. 1hr .. /dump. Frustrated as all hell, I went down to x38. Set all voltages to Auto. This 300Mhz below the stock speed. Exactly 2 hours into A46's test, it borked. There is something going on with A64. What is is, I have no clue, but when this things BELOW stock speed by a considerable amount, and it craters 2 hours in.. It has to be the test itself.

x38 is one of this chips "low power" states. 3800 MHz, 0.9V is its mark. Its voltage hit 1.464, set to Auto, @ x38. So, either the board is being retarded (Most recent bios installed, so doubt its that) or A64 is demanding a shit ton of un-needed voltage. the TDP of this chip is rated at 100w. That's at default clock base of 4.1Ghz. During this last failed test, it sucked down a stupidly high amount of 164w...@ 3.8Ghz!

I'm currently back to "Optimized Defaults" in the BIOS until I can find out what in particular is causing this wildly erratic behavior.

[DeltrusDisc]

Yeah, this sounds like a deeper problem. Back before you posted here, you said you'd managed a 'stable OC'... remind me what did you use to confirm its stability and what was the length of time tested?

[Gravewyrm]

Yeah, this sounds like a deeper problem. Back before you posted here, you said you'd managed a 'stable OC'... remind me what did you use to confirm its stability and what was the length of time tested?

My "Stable" tests were done with both A64 and P95, but were only 2 hour tests. Both of my sig block badges were done with the 2hr test with both softwares. I have less issues with P95. In P95, I did large in place FFT's, as to stress mainly the CPU.

Also, the version of A64 I used was the Trial version of A64 for the tests that "passed" both sig block speeds. I now have the full version of A64 Extreme.
I'll tinker in P95 later, and see what I can come up with. Made it to 4 hours @4.7. Only stopped it to get into the A64 test. Ill keep ya updated.

[DeltrusDisc]

Awesome, glad to hear it. I would definitely post on other forums like overclockers and see if anyone else can help you further. There might be a club of knowledgeable people there with similar (or the same) CPU(s) as you, so that might help!

[Gravewyrm]

I'm done. This chip literally can NOT do a test longer than 3.5 hours, no matter the clock, no matter the voltage, no matter the test. Tested from 3.8Ghz with default voltage, all the way to 4.7 @ 1.512v. It refuses to get past 3.5 hrs in either P95 or A64.

I'm officially throwing in the towel with this CPU+MoBo combo. Its flat out impossible. Voltages fluctuate too much (Yes, even with LLC at multiple diff levels). APM on or off, C6 on or off, Multiplier literally locked in to x[insert number here], it does not matter. NB voltage tweaks didn't matter. The power draw is flat out insane at even 4.5Ghz, and this things turbo is 4.4. Multiplier dipping to x38 when everything is set up to any number higher adds to the flux. Even tried the latest beta BIOS (F5a), same behavior.

[chronia]

Any weird error messages that you are getting? I've stressed a few Apu's (Axx-5x00 series, as each machine i build gets stressed, even at stock to make sure i deliver a good build), and at stock clocks/volts i never really had a issue. Did not have the luck yet to be able to build a OC'd machine yet based on a Apu so cant comment on that.

[teamedward]

Somewhat off-topic, but the AIDA link on the initial post still links to the Prime95 ftp and not AIDA. Not sure if this is intentional or not.

[DeltrusDisc]

Somewhat off-topic, but the AIDA link on the initial post still links to the Prime95 ftp and not AIDA. Not sure if this is intentional or not.

Absolutely not intentional! Thanks for pointing it out though, I thought I'd changed everything over!

[Draculla]

I overclocked my i7 4770K to 4.2GHz with Coolink Corator DS

[Gravewyrm]

I'll just drop this right hur....







Name: Gravewyrm
Processor: AMD A-10 6800K
Clock: 4.70Ghz (47x100Mhz)
Voltage:1.464v Peak, 1.368v Nominal
Cooler: Corsair Hydro Series H60
Screenshot LINK: http://i.imgur.com/UdzlAQm.png
Valid CPU-Z link: http://valid.canardpc.com/sahbup

[chronia]

I'll just drop this right hur....







Name: Gravewyrm
Processor: AMD A-10 6800K
Clock: 4.70Ghz (47x100Mhz)
Voltage:1.464v Peak, 1.368v Nominal
Cooler: Corsair Hydro Series H60
Screenshot LINK: http://i.imgur.com/UdzlAQm.png
Valid CPU-Z link: http://valid.canardpc.com/sahbup

Nice, how did you fix it? Or you just got a new Apu?

[Gravewyrm]

I tore the system apart. Re-assembled everything. During that process, I discovered that the cooling block had become loose, and that the TIM was in fact not covering 100% of the die surface. re-applied TIM(never, EVER use the "Grain of Rice" method, as this is what caused this issue in the first place), put her all back together, and ran an 8 hr stress test @ 3.8Ghz, to break in the TIM, and re-adjusted the block. Once that was done, set up the voltage @ 1.414v, 47x multiplier. Brought the voltage up until I could stably log into windows and run its WEI. Set up A64 for a CPU burn only. This is where I discovered just exactly how both the BIOS and A64 work for voltages.

Bios shows what you have set as the BASE voltage. Not peak. During A64's test, the voltage shot up exactly 0.050v, so I adjusted my base BIOS voltage accordingly.
After that, re-ran the test, and the voltages are what you see in the screenshot. The voltage shown in the screenie on CPU-Z is in its lowest Pstate, 100% load. It held this setting for 3 hrs+.

Just to clarify .. the 1.414v that is shown is BIOS is actually its "Standard" voltage for a moderate load @ 100% CPU usage. During the A64 test, the max voltage hit 1.464v @ 55c. Later today I'm going to be tinkering with lowering that high end voltage and try to shave off a degree or 3 while remaining 8hr+ stable. Results will be posted in a cpl days.

[DeltrusDisc]

Great to hear! Though a .05v jump does seem a little... high. I'll go ahead and get you added! Congrats!