BSOD Going From 4.5 to 4.6 on i2500k

meths

meths

meths

Soldato
Joined
12 Mar 2005
Posts
4,406
BSOD Going From 4.5 to 4.6 on i5 2500k

My system is:
i2500k
Asrock z68 Pro3 mobo
8gb Corsair XMS set at 9-9-9-27 1.65v
Sapphhie Toxic 6950 unlocked to 6970.
Soundblaster Fatal1ty
Enermax Liberty 650W Modular PSU
Silver Arrow CPU cooler.


Started overclocking my system last night and all was well up to 4.4ghz at 1.25v.

Set it to 4.6ghz and BSOD. Upped the voltage to 1.38v and it'll boot to the desktop but fails on the fifth pass of Intel Burn Test with a BSOD. Temps are fine at 1.38v. I tried upping the VTT voltage to 1.076 from 1.05 but that didn't make any difference.

If I drop it back to 4.5ghz it's rock solid at 1.25v.


I'm going to start trying to find the fault tonight but was just wondering if you guys had any suggestions of obvious stuff that could be the problem?


Edit: Quite possibly power related. In the Asrock Extreme Tuning Utility, if I click on the IES tab I get an instant BSOD at 4.6ghz. I don't really know what it's doing but the explanation on Tom's Hardware is:
As a counterbalance to hot-and-hungry overclocking, ASRock’s Intelligent Energy Saver allows users to save a few watts by dynamically shedding power phases under low CPU loads.
Click to expand...
 
Last edited:
Are you following a guide at all or going in pretty blind? after x44 there are other setting you typically have to start changing other than just the multiplier and core volts.

Here is an okay guide but it's not your motherboard, i'll see if I can find the one which covers multiple main boards.

EDIT: Lol, forgot the link.
 
Last edited:
Thanks. I'll have a read at that later.

I've just been using the manual for the motherboard and looking up any settings I didn't understand in the net. I've set nearly everything I thought was relevant manually in the bios.

I've got a couple of ideas about what the problem is but it just seemed a bit extreme that 100mhz was the difference between a stable system at low volts and a system that BSOD's at the drop of a hat. I thought the instability would come in more grdually than that e.g. stable at 4.5 and erroring in IBT at 4.6.
 
I think your problem will be Internal PLL Overvoltage needs to be set to enabled in bios.
Once you find the cpu's limit with it disabled it usually gives another 500mhz or so.

If your stable at 4.5 with 1.25v then 4.6 should be ok at about 1.27 - 1.28v , deff no more than 1.3v needed
 
Last edited:
That would be ideal if that's the case.

Kinda had my heart set on 4.8ghz so I had a nice round 1.5ghz OC. :)

Anything above that is a bonus.
 
Like the other guys said, it could be a variety of things, the guide will go into detail but the usual culprits are PLL overide (over 4.6ghz overclocks), Vcore and Load line calibration if your getting crazy vdroop.
It could even be that your cpu hits the wall at 4.6 but that seems less likely as you have jumped the vcore a lot with no sucess.
 
From what you guys are saying PLL override seems to be the best bet.

I thought I had almost ruled out vdroop by setting the voltage to 1.38 which is likely much more juice than is needed for that clockspeed. I also kept an eye on the voltages when running the last test and didn't see any noticable drop under load. I'll adjust the LLC setting to 2 later which on my board should ensure a stable voltage under load. Currently I've got it set at 5 which is basically LLC switched off.
 
I would guess PPL overvoltage. Don't forget you probably won't be able to recover from sleep with PPL OV enabled so sleep should be disabled in windows.
 
Well I've tried every setting under the sun to get this stable at 4.6ghz but no joy. Looks like I'll have to settle for 4.5ghz.

I already had PLL override enabled and LLC made it last a little longer but not much. I was sticking 1.45v through the chip and LLC was making it go up as high as 1.47v and the temps were getting a bit high.

I also tried bumping up CPU PLL from 1.802 to 1.89 but still didn't do anything.

I did find out that Spread Spectrum enabled was making my BCLK frequency fluctuate below 100 for some reason.

I'll screendump my bios pages and post them on here to see if you lot have any other ideas but failing that I'll start seeing how low I can get the voltage at 4.5ghz.

Actually I'll try 4.7ghz first in case it just doesn't like 4.6 for some obscure reason.
 
Seems a bit odd that its crapping out at 4.6 even with 1.47v, surely it can't be hitting the mhz wall at that speed.
You were extremely unlucky with it if thats the case, would be the lowest clock I've seen on one.

I thought all SB cpu's were capable of at least 4.8ghz with some just needing more volts than others.

Are there any newer bios's for your board you could try?

Another thing you could try is disabling PLL overvoltage and then see how high you can get it.
If it is hitting the mhz wall at 4.6 then without PLL Overvolt enabled it should start getting flakey at around 4.1, 4.2 should bsod pretty quickly.... I'm doubting its that bad a cpu though.
 
Last edited:
That cpu's got plenty more left in it than 4.5 then but no idea why its not playing ball, its deff not vcore related thats for sure.
 
Bos said:
That cpu's got plenty more left in it than 4.5 then but no idea why its not playing ball, its deff not vcore related thats for sure.
Click to expand...

Just did some testing on it. Passed 20 runs of IBT on maximum and 8hrs plus of Prime 95 at 4.5ghz so it really is as stable as it's gonna get at that speed.

PLL and LLC are both off.

Now to find out what is causing the crash at 4.6...

Edit: There's also a beta bios that I can try.
 
Just been having a look at bios options for Asrock boards and found a lot of folks suggesting these settings for higher overclocks of 4.6 - 5ghz+.

Turbo Boost Power - Manual
Short Duration Power Limit - 250
Long Duration Power Limit - 250
Core current Limit - 250

You might already have em set at this but worth a try if not. :)
 
Bos said:
Just been having a look at bios options for Asrock boards and found a lot of folks suggesting these settings for higher overclocks of 4.6 - 5ghz+.

Turbo Boost Power - Manual
Short Duration Power Limit - 250
Long Duration Power Limit - 250
Core current Limit - 250

You might already have em set at this but worth a try if not. :)
Click to expand...

Yeah, already did that.

just now I've turned off speedstep and that vanderpool thing and it seems a little bit more stable. Still failed after 6 runs of IBT but it's further than I got before.

Turned PLL back on and just about to stress it again. Fingers crossed.:)
 
I think you might have been right on your OP about it being power related m8.

Your board only has 4+2 phase VRM which is probably struggling to supply enough clean power when you get over 4.5ghz.
 
You must log in or register to reply here.
Back
Top Bottom