I think I just Killed My CPU

Aleman

Aleman

Aleman

Associate
Joined
3 Mar 2010
Posts
16
Intel i5-750 in a Gigabyte GA-P55M-UD2 Motherboard with GSkill Ripjaw Ram and a OXCZ moDStream 600W PSU. Arctic cooling Freezer 7 Pro rev 2 Cooling.

Ran Prime 95 Small fft earlier today for 90 minutes no problems, highest temp recorded was 76C . . . but as I didn't have the fan on teh Freezer 7 that was to be expected, still the system was stable . .. put the Fan on and got core temps around 26 degrees lower using Prime 95.

Decided to do an over clock

Base clock to 190
System memory Multiplier to 8 (Taking memory to around 1550Mhz)
CPU VCore to 1.35
QPI/VTT to 1.37
PCH Core to 1.14V
CPU PLL to 1.9V
DRAM Voltage to 1.64V

System booted normally, and straight into Windows. Started CPU-z and overclock had been applied now running at 3.8GHz . . . Core Temp showed around 47C . . . Started Torture test of Prime 95 . . . Temperatures crept up to around 80-82C . . . a bit warm .. . . left the system running for 30 minutes . . . came back and all was shut down.

System now refuses to power on, get a slight spin of the fans as I switch on but that is it . . . Guess I've killed it? . . . suppose I should clear my CMOS and see what happens then?

Any other ideas?
 
Yes, reset the cmos and try again.

It would be sensible to overclock much more gradually than this, only raising voltages when required. What you have done is set them to roughly the highest safe limits and put bsck up. This is a bad plan, as it makes getting the system stable very difficult and overvolting very likely.

Welcome to the forums.
 
I doubt its dead more likely the Bios playing silly buggers with you i have been there before and thought i had killed something but reset the bios if not down to good old process of elimination.
 
Thanks for the reply John, Clearing the CMOS hasn't made the slightest difference.

Bit of an expensive lesson to learn "Take advice from magazines with a pinch of salt. The is no such thing as a 'safe' overclock". Time to order a new i5 I guess
 
You shouldnt really leave the system while its stress testing, especially when its the first overclocked run and youve just jumped in on some pretty ambitious settings.
 
Aleman said:
There is no such thing as a 'safe' overclock". Time to order a new i5 I guess
Click to expand...

Wrong conclusion. If the hardware is dead, then the overclock you set was not a safe one. Given how aggressive it was this is hardly a reason to give up on the game entirely, there are many entirely safe overclocks. You just didn't pick one. If you took your advice from custompc then you have my sympathy, they've been handing out outrageous advice for ages now.

Switch the psu off, take the cmos battery out, then unplug the computer from the mains. Go and make a sandwich. Come back 15 or 30 minutes later, put the cmos battery back in, plug in the computer, turn the psu switch back on. It'll probably be fine.

I'm fairly sure you haven't killed it.

edit: if you were worried about it overheating, then stress testing without the fan was a bit silly. Nonetheless, the most likely thing to happen if it overheats is for the computer to switch off and reset the cmos to defaults.
 
Last edited:
When I said I had left it . . . I hadn't actually gone away, apart from to go and get a cup of tea toward the end . . . That's when things went awry. If it was overheating I would have thought the CPU would have throttled back? . . . . but thats probably negated by the over volting.

BELGIUM!!!
 
TBH with 80+C on the cores and 1.37 QPI/VTT there is a good chance it has fried something... you might wanna check the CPU socket as well for signs of arcing it has been known to happen with certain foxcon sockets.
 
JonJ678 said:
If you took your advice from custompc then you have my sympathy, they've been handing out outrageous advice for ages now.
Click to expand...

I mentioned no names ;) Still CMOS battery removed . . . Time for a quick drink and try again I guess. The funny side of this is that this is a new rig, that my wife paid a good chunk towards . . .because at Christmas she turned my old one on . . . and took the PSU and Graphics card out . . . not looking forward to telling her today when she wakes up :rolleyes:
 
OK So having removed the BIOS Battery and slept on it, the machine still fails to POST, given these parameters
Base clock to 190
System memory Multiplier to 8 (Taking memory to around 1550Mhz)
CPU VCore to 1.35
QPI/VTT to 1.37
PCH Core to 1.14V
CPU PLL to 1.9V
DRAM Voltage to 1.64V
Click to expand...
What is most likely to have failed, CPU, MoBo or RAM as I need to reorder parts to replace.

It's my own damn fault really, I was naive enough to think that the settings they had published were 'safe' .. . and certainly I guess some hardware will cope with those settings . . . mine didn't . . . and I should have checked somewhere else before applying them . . . . Lesson Learned
 
It's pretty damned hard to kill a modern CPU after a few minutes or even hours or days even with insane voltages.

When you removed the cmos battery did you short the reset cmos jumper?
 
Well you could do a bit of diagnostics. Take all but one RAM stick out, try to boot, if it fails, try a different RAM slot; when you've tried all the slots, try a different stick of RAM. Repeat.
If your graphics card's power-hungry, try taking out the power cables. If it beeps it's probably a good sign (although I don't actually know if that's the mobo beeping, signifying an intact mobo, or the graphics card beeping)
 
I didnt mess with voltages that much tbh.

I just put my multiplier on 20x , clock at 200 (4ghz) , and the cpu volt was on 1.35 so left it. Its running faultlessand very stable.

Its probably the volts that has killed something.

Doesnt every motherboard have a fail safe now where it shutsd down if its going to do damage?

My Asus P7P55D Pro does as i mistakelnly put a massive overclock (something like 5ghz) and it refused to let me save it and restart, instead it just said "unsuccesful overclock"

Matt
 
i can remember leaving the room whilst stress testing my system with the stock cooler, came back a few minutes later and the cpu temp was reading 90c, i was like s### turn of prime quick:eek:


i can only suggest clearing cmos (jumpers) and remove & install memory modules, perhaps in different slots. replug all wires etc.

earlier this week i had to remove the mobo etc to install a new nortbridge heatsink & rebuild. i did this after been awake for over 30hrs - not a good idea:eek:

anyway on rebuild the system appeared not to boot, turned off, on, etc etc. cleared cmos and nothing. could hear the psu etc but no post. i thought id killed something. regretted doing this big time and then realised i hadnt plugged the monitor cable in. what a numpty:D
 
hyperst said:
Intel spec for QPI/VTT is a max of 1.21v
Click to expand...

That can't be right. I've always read that it's 1.35 or is it different for i5?

To the OP, I seriously doubt you've fried anything. Yes them voltages are quite high but I've seen much higher.

RoEy
 
Last edited:
JonJ678 said:
Switch the psu off, take the cmos battery out, then unplug the computer from the mains.

Come back 15 or 30 minutes later, put the cmos battery back in, plug in the computer, turn the psu switch back on.
Click to expand...

Could you clarify whether or not you did the above, exactly as written? You need the psu unplugged and the battery removed at the same time really.

Trying sticks of ram one at a time is a good plan. Make sure you've got a speaker connected up to the board so you can hear diagnostic beeps. Unplugging the graphics card power is likely to make a awful noise and kill the motherboard if you leave it for long enough, I don't think this is a good idea. I believe it's the motherboard screeching but I'm not sure about this either, scares the life out of me every time I forget the peg cables.
 
JonJ678 said:
Could you clarify whether or not you did the above, exactly as written? You need the psu unplugged and the battery removed at the same time really.
Click to expand...
Yeah Did that exactly . . Left it overnight and tried it out this morning . . . Nothing fans gave a little twitch

Thanks for all the help and reassurance . . . I suppose I really need to stop panicking and think clearly and logically about this.

symptoms are the machine fails to post . . . and although the Fan blades twitch when I power on the fans fail to start.

I have shorted the BIOS Reset Jumpers but not with the battery out . . . That is probably something to try next. I have removed the cooler and CPU to check for arcing . . . nothing visible.

Firstly I suppose I should look at the PSU and see if it is supplying current . . . I have another (el cheapo) PSU on my rebuilt old rig (Socket 7 don't laugh) that I can swap out . . . if only for confirmation. . . . then I guess the memory needs to come out.

As Michael Winner would say "Calm Down Dear . . . its only an advert"
 
You have to short the BIOS reset jumper with the battery out. Also try with just 1 stick of RAM and try various slots.

Sometimes puters do funky things for no apparent reason, frustrating but you will eventually learn the ins and outs of your setup.
 
You must log in or register to reply here.
Back
Top Bottom