Diagnosing processor failure

[pvc]

I have been getting an increasing number of BSODs during normal usage so I have been trying to diagnose the cause. i2500K + P8Z68 V-Pro. I am able to force a crash almost immediately using Prime95's Blend test. Temperatures are in a decent range and the crash happens before the CPU has even had a chance to heat up. I ran memtest to make sure it wasn't a RAM issue. Any other suggestions to pinpoint the fault will be greatly appreciated.

Edit: I should also mention the processor has never been overclocked.

 

[Nenu]

Temps are in decent range isnt helpful.
What program are you using to measure temps and what are the temps?

 

[ccityinstaller]

Does the system crash if you run the Small FFT option of Prime95? Blend is usually used to find errors in memory or with the IMC (memory controller)..If you can run the Small FFT option ok, then the cpu isn't the likely culprit..I have seen memory that will pass Memtest for hours and fail right away in Prime as well..

Try running the Small FFT option for at least 30-45 mins and watch your temperatures as well with something like RealTemp and report back and we will see what we can come up with.

 

[cyclone3d]

Are you using the stock heatsink with the stock thermal paste? If so, it could be that the thermal paste has started to seperate and is causing a hotspot which can cause lockups but will not necessarily show up as high CPU temps.

When was the last time you cleaned the dust out of the system?

What power supply are you using and how old is it? It could very well be a bad power supply.

Does the motherboard have any bulging capacitors?

How clean is the voltage coming from the wall?

Have you tried reseating the RAM? What about the CPU?

 

[pvc]

Temps are in decent range isnt helpful.
What program are you using to measure temps and what are the temps?

The cores idle at 31-33 degrees. I have been using CPUID HWMonitor.

Does the system crash if you run the Small FFT option of Prime95? Blend is usually used to find errors in memory or with the IMC (memory controller)..If you can run the Small FFT option ok, then the cpu isn't the likely culprit..I have seen memory that will pass Memtest for hours and fail right away in Prime as well..

Try running the Small FFT option for at least 30-45 mins and watch your temperatures as well with something like RealTemp and report back and we will see what we can come up with.

I see a BSOD as soon as I start the Blend test, so no real time to notice any temperature spike in HWMonitor. I will try the Small FFT test and get back with the results.

Are you using the stock heatsink with the stock thermal paste? If so, it could be that the thermal paste has started to seperate and is causing a hotspot which can cause lockups but will not necessarily show up as high CPU temps.

When was the last time you cleaned the dust out of the system?

What power supply are you using and how old is it? It could very well be a bad power supply.

Does the motherboard have any bulging capacitors?

How clean is the voltage coming from the wall?

Have you tried reseating the RAM? What about the CPU?

I am using the stock HSF.

The system is pretty clean. I clean it with compressed air every couple of weeks.

It is an older power supply. Maybe about 4.5 years old? A corsair TX650W.

I see no bulging capacitors on the motherboard.

I use a UPS with voltage regulation.

I haven't done this yet. I do have 4 sticks of RAM. I will try to run the tests again, cycling through removing one stick of RAM each time. I will also reseat the CPU with some fresh thermal compound.

 

[Jayllo]

I haven't found memtest to be foolproof with ram either.

Are you able to check the cpu & system voltage?

PSU might be the culprit as someone else mentioned.