• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Strange OC behavior (2500k)

wabbitseason

[H]ard|Gawd
Joined
Jun 16, 2010
Messages
1,511
Edit: Solved the problem, see post #4 for an explanation.


Hello folks,

Weird problem overclocking a 2500k rig (MSI Z68MA-ED55 mobo), with the following settings:

PLL: Auto
DRAM frequency, voltages, and timings: appropriate
Spread spectrum: Enabled
Vdroop Control: Auto
CPU I/O: Auto
System agent voltage: Auto
DDR_VREF_xxx: All set to"Auto"
PCH: Auto


I can boot into windows and run LinX flawlessly at the following core voltages/multis:

1.25/38
1.3/40
1.375/44

Temps are good. The problem is that when I restart, the system always fails to POST and the mobo reverts back to the last "stable" configuration. For example, if it just failed to post at 3.8--and I'd tried 4.4 previously--after a restart it'll actually revert to the 4.4 settings, not default!

I would think if these were unstable they'd be failing to boot into windows, let alone handling LinX for hours at a time. But somehow the restarting process is revealing some instability. Anyone have any idea what the heck is going on here? Thanks.
 
Last edited:
I was going to answer the reason of your problem then I remember you tried to discredit me.. and suddenly forgot the answer:D:rolleyes:.. goodluck trying to find the issue.. but I will throw you some light and its related to "vdroop" start with that.. =)..
 
Last edited:
Trolling replies aside, I'm sure there's some folks here who've had a similar problem. If vdroop were the problem I'd be getting voltage-induced bluescreens under load, not POST failures.
 
Problem solved, explanation for posterity:


The problem had absolutely nothing to do with vdroop, as Araxie indicated so confidently. Apparently he needs little help from anyone to discredit himself.

MSI z68 and z77 boards have an important feature missing--vcore voltage offset(VVO). On other boards which have VVO, you simply set a fixed vcore for your OC in the BIOS and leave EIST/turbo enabled. The system intelligently modulates the voltage according to the current frequency of the CPU, and everything is gravy.

Without VVO, if you set a fixed vcore in the BIOS it will be fed to the CPU regardless of load or frequency. The only way to get an intelligently-modulated voltage on a board without VVO is to use automatic vcore voltage. Unfortunately, the "auto" setting usually chooses retardedly high voltages (when I tried this, it used 1.4V for 4.4GHz!). For reference, the chip is 24 hours prime stable at 1.3V 4.4GHz, so that's a hugely unnecessary overvolt.

All this leaves MSI z68/z77 owners with 3 mediocre options to overclock:


1. Old-school overclocking

Disable all powersaving features, including EIST and turbo, and set a fixed vcore. Set the multiplier AND 'core ratio limit' to the same value (i.e. 44 for 4.4GHz). The CPU will always run at the max frequency and vcore set in the BIOS. The obvious downside is more idle power consumption and heat than is necessary.

2. Turbo-mode overclocking

Leave powersaving features enabled, including EIST and turbo, but set vcore to 'auto'. Keep the multiplier at stock, but set the 'core ratio limit' for each core to your desired overclock. 'Core ratio limit' defines the max turbo frequency for each core. This is the only way to have both vcore and cpu frequency intelligently modulated according to load on these boards. The major downside is that the 'auto' vcore is usually far higher than your chip needs, wasting power and adding heat. The upside is low idle vcore.

3. Multiplier overclocking
Leave powersaving features enabled, including EIST and turbo. Set the multiplier AND 'core ratio limit' to the desired value, and set vcore to a fixed value appropriate for it. This option allows your cpu frequency to move around according to load, but vcore remains fixed at the value you set. The upside is a far lower load vcore than the value automatically chosen in option (2). The downside is higher idle vcore than option (2).


I was trying a bastardized combination of these three, being unfamiliar with the MSI BIOS and its limitations. Hope this helps somebody!
 
and again the error in post have to be related to vdroop and the way MSI apply vdroop under load.. when Vdroop its applied it actually what do its decrease a certain amount of voltage under load.. so if you select a fixed voltage of 1.4v the total amount of voltage under load will be much less (depending on the vdroop level)... (also way different on how in other boards like asus its applied the load line calibration).. so, how this affect your issue? in the moment the system its turning it will run at full performance so the processor run at max frequency and voltage, if your voltage isn't high enough to compensate voltage drop by the MSI board it will be simply out of voltage enough to feed the CPU triggering the "Failed CPU Overclocking warning" and its typical in most MSI boards... so yes when applying a fixed voltage + Vdroop the voltage in idle its in fact higher than load... (many users here can explain it too :) ) and that can be another instability issue during typical loads and daily usage(but no related to your issue actually).. so yes, machine POST issues and failed to overclock issues have to be related A LOT to vdroop.. as i said in my post i thrown just a hint.. :). btw the thing with msi not supporting Offset voltage its old as hell and well known in the users.. :)
 
Last edited:
Back
Top