CPU spikes followed by black screen requiring reboot

[martialshot]

My build should be in my signature, if any more detail is needed I will be adding to the description as needed. I recently added an all in one water cooling system to the CPU and that has not caused or affected the issue. The issue began roughly 2 weeks ago. I added an additional 4 gigs of RAM 2 x 4 of the same make below but a slightly higher clock speed (local store was out of the clock speed I had) At this time the internet would disconnect every 10-20 mins for roughly 3-5 seconds at a period of 5-10 seconds, and then would reconnect with no issues (tracked this through the router/task manager) I switched the ethernet cord from one of the NICs on the back of my computer to another NIC and moved it over a port on the router and this issue disappeared (therefore unsure if related) HOWEVER recently with no establish-able pattern or period the computer will spike CPU then the spikes occur roughly 3-5 seconds apart last for one second for a duration of 5-10 minutes followed by both monitors turning black. I have not waited this out to see if it rectifies itself, as it requires a hard boot to become usable.

For the record there is no overclocking, everything is at stock values (can add screen caps tonight). Also for the record the water cooling for just to do it to do it. There are no heat issues (will post screen caps again tonight) during usage or during idling. I have updated the drivers for the graphics card very recently as I assumed this was a graphics card issue at first. I have started the RMA process with GTX card but I doubt that is the issue.

Any ideas? Any additional details needed?

 

[rds1220]

My build should be in my signature, if any more detail is needed I will be adding to the description as needed. I recently added an all in one water cooling system to the CPU and that has not caused or affected the issue. The issue began roughly 2 weeks ago. I added an additional 4 gigs of RAM 2 x 4 of the same make below but a slightly higher clock speed (local store was out of the clock speed I had) At this time the internet would disconnect every 10-20 mins for roughly 3-5 seconds at a period of 5-10 seconds, and then would reconnect with no issues (tracked this through the router/task manager) I switched the ethernet cord from one of the NICs on the back of my computer to another NIC and moved it over a port on the router and this issue disappeared (therefore unsure if related) HOWEVER recently with no establish-able pattern or period the computer will spike CPU then the spikes occur roughly 3-5 seconds apart last for one second for a duration of 5-10 minutes followed by both monitors turning black. I have not waited this out to see if it rectifies itself, as it requires a hard boot to become usable.

For the record there is no overclocking, everything is at stock values (can add screen caps tonight). Also for the record the water cooling for just to do it to do it. There are no heat issues (will post screen caps again tonight) during usage or during idling. I have updated the drivers for the graphics card very recently as I assumed this was a graphics card issue at first. I have started the RMA process with GTX card but I doubt that is the issue.

Any ideas? Any additional details needed?

From what you said here:

HOWEVER recently with no establish-able pattern or period the computer will spike CPU then the spikes occur roughly 3-5 seconds apart last for one second for a duration of 5-10 minutes followed by both monitors turning black.

It would seem like something is causig the CPU to spike and then the CPU overheats and shuts down.

 

[martialshot]

From what you said here:

HOWEVER recently with no establish-able pattern or period the computer will spike CPU then the spikes occur roughly 3-5 seconds apart last for one second for a duration of 5-10 minutes followed by both monitors turning black.

It would seem like something is causig the CPU to spike and then the CPU overheats and shuts down.

How can I monitor the CPU heat? Iam using CPU-Z and the Intel Extreme tuning utility to monitor and Ive never seen it above 35-45 degrees Celsius. Is this incorrect?

 

[martialshot]

Here's the current values, Ill try and grab the spike;

 

[rds1220]

I don't really know anything about Intel Extreme so I don't how reliable it is. I use CPUID Hardware Monitor.

http://www.cpuid.com/downloads/hwmonitor/1.21-setup.exe

You can also use your BIOS to get another input and then compare all three. If two are the same or fairly close and one is way off the two that are close is pobably the correct temps.

 

[martialshot]

I don't really know anything about Intel Extreme so I don't how reliable it is. I use CPUID Hardware Monitor.

http://www.cpuid.com/downloads/hwmonitor/1.21-setup.exe

You can also use your BIOS to get another input and then compare all three. If two are the same or fairly close and one is way off the two that are close is pobably the correct temps.

Looks as if its correct

During the spikes I do not see a response in temperature nor do I see the process in the resource manager. Ill cap next time or try and force it over VPN.

 

[rds1220]

Before you put in the new RAM did this ever happen? What happens if you go back to your original configuration with just the old RAM sticks?

 

[martialshot]

Before new RAM: Odd networking issue with internet disconnect
Add new RAM: CPU spikes followed by black screen
Remove new RAM: CPU spikes followed by black screen
Just new RAM: ....actually I haven't tested that....

Overall the build is slightly unstable

I can repeat the tests tonight and ensure the results are the same...do you have the link for mem test off hand? I should probably do one of those as well to see if its RAM stick or motherboard RAM slot.

I did do a prime95 blend test for about an hour with no issues...any idea how long this should run?

 

[rds1220]

Try just the new RAM and see what happens. Also try Prime 95 for atleast a couple hours and Memtest 86+.

 

[martialshot]

Memtest is good with one stick of old RAM running for 2 hours the Prime 95 for a 4-6

 

[martialshot]

Started the Prime 95 test at 8PM EST so will end at 8AM EST tomorrow then will add just the new RAM in and play around a bit and see what happens

 

[martialshot]

So the EVGA tech talked to me about the video card while I was starting to RMA it and he mentioned the following "
If the issue persists i would highly recommend on checking if the PSU is providing consistent power with the card installed. Go into your BIOS and look for the +12v reading, this is typically under 'PC Health Status', System Monitor, Hardware Monitor'. A healthy reading (official ATX spec) is between 11.4 - 12.6v, and stable. However, different brands have different tolerances for overvolting and under volting, if it is outside 11.9v and 12.30v, or if you observe consistent fluctuations of more than .05 volts then this may be a sign of concern for a degraded or failing +12v rail. However, BIOS will only show you an idle reading. You can try monitoring the +12v in Windows under load, using a free program like HWMonitor"

Completed this....its fine.

So far testing includes P90 Blend test for 12 hours.....successful

Memtest 2 hours on one stick will do the rest tonight.....successful on first stick

HWMonitor test.....Voltage is fine

Anyone else have any ideas?

 

[martialshot]

 

[martialshot]

I just crashed again and pulled as MUCH data as I could

 

[martialshot]

Try just the new RAM and see what happens. Also try Prime 95 for atleast a couple hours and Memtest 86+.

This is with just the new RAM

 

[thenh813]

This is probabally the wierdest lock up problem I have heard of.
I honestly dont know what to say besides maybe check the voltage with a Multimeter
on one of the Molex or PCIE to confirm it is proper and maybe feel heatsinks and voltage
regulators during the spikes to verify that nothing is overheating.
It could be some sort of driver/hardware conflict occuring intermittently so for the
internet disconnecting untill the connection (and thus IRQ) connection was changed and the
graphics card got the old IRQ and now has the conflict problem. Just my theory.