Originally Posted by cadaveca
You're approaching it wrong, IMHO. 115 BCLK...USB3, audio, and other devices need to be disabled. USB3 starts to have issues @ 105 BCLK already, but if you do not use, you'd not see it. I leave a thumb drive plugged in, and when it drops out...I know BCLK is too far.
To get 2700 MHz ram, you run 125 BCLK divider + some BCLK. What you are doing is pushing the 2400 MHz divider, which is pretty weak to begin with....some CPUs just are not stable without significant voltage increase (I have one that is like this).
Based on what you report...you are having the issues you should expect doing that sort of clocking. Nothing wrong at all...just how it works. try something else.
I kinda know but I'm trying to figure out what I really need to get past this, I thought of trying to lower the cpu multi way down but I haven't gotten to it yet.
You're saying usb3 cuts out?
I can disable audio and usb3 modes maybe, maybe if I look around my house for an adapter I can try it without usb at all.
But I'm honestly doubting it has anything to do with the pch at this point so far, but then again I don't know enough about this board yet.
I know my vga are fully capable of 150mhz pcie, at 2.0 though never had 3.0 before on my amd...
I just don't wanna go with the higher straps.
I know it might be impossible to get higher clocks fully stable at the 2400mhz mem div...
Which kinda sucks too.
2700 boots and benchs fine at low voltages, but can't pull of y-c at any voltage (at least to 1.8v as far as I tried it).
But then again she's slower at that speed unless go for lower timings.
2431.62mhz 9-11-11 is faster then 2736mhz at 10-12-12, except for latency (aida64) and dpc latency which is better at the 2700'ish.
And it works just fine I think at 114mhz bclk.
I actually failed y-c at 4.5hrs, but I'm re-checking now at a lower cpu multi, I went past 4.4ghz on stock voltage... (was checking both bclk and cpu speed at the same time, 4.45ghz)
I've done 10hrs at a higher vccsa and 3hrs at the same vccsa as now.
At least I know how to get away from that weird bandwith bug I get.
What's odd, at this peticuler spot where I'm at, if I change from a cpu multi of 38->39x or vice versa the system doesn't power down to switch the multi and the bug never shows it's self.
Anyways I can at least test the board now, before I didn't wanna do anything until I fixed that bandwith prob.
What I'm trying for now is to verify everything at 8hrs y-c.
Then try inching past 4.4ghz cpu to 4.45ghz and figure out how I can get that stable without touching llc and pwm, thinking maybe if the pll overvoltage setting gets me stable then cool
It's worth trying, Right now I'm running 4321mhz lol, at least that's what aida64 is saying (I know up to 4.4ghz is stable at stock and 4.5ghz is close, 4.45 should be closer).
I can't just jack up the pwm and everything because I don't have the side of my case on, I have it on so I can switch vga bios'es when in dos for the time being.
I've got fans ready to cool my mem and pwm but for now I gotta deal with stock llc and pwm speeds for the cpu pwm.
I don't wanna overheat it and have probs with that, like my previous ch4-f.
One more thing before i end my post here...
I was thinking about checking with my watt meter to see if higher straps use full power 24.7 or not.
It's just they don't idle the cpu voltage and that bothers the heck out of me...
Otherwise I'de be messing with it.