dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
953
share rss forum feed


Gordo74
Premium
join:2003-10-28
Monroeville, PA

PC Freezing in/ouside of Windows... What would cause this?

Click for full size
Click for full size
Hey all.

Around Christmas time, I built a gaming PC from my brother. Lately, he has complained about it locking up and rebooting. He states that it runs normally for 1-2 hours then freezes. The first error he got was a "Vertex compile error", but then started getting other random errors for programs and the like. He says it happens wether it is idle or load.

My first thought was temperatures, so he checked those. They are in the 40-50C range. All is well there.

The error he started getting was "Vertex Compile Error" which led me to believe GPU. So I had him update the graphics card. No dice. So I had him take out the video card and run it off the onboard Z68 Graphics. It was still happening.

So then I had him run Memtest. The first pass through, it came up with crazy errors, so I suspected RAM, so we tried 1 stick of RAM and changed the sticks as well, all exhibited the same behavior, errors, then freeze. I attached the screenshot of the memtest errors (I think that's what they are under the bottom left section) and then the freeze, where c0 shows all over the screen.

That lead me to rule out hard drive as well, since it was happening outside of Windows/when booting off of USB.

Then I remember that it was a 2500K processor which I had overclocked to 4.0ghz, so I had him set it back to stock and it still freezes.

Here are the specs:

500w OCZ Fatality PSU
MSI Z68 Motherboard
Intel i5 2500k
ATI 6870
64GB M4 SSD
1TB Samsung HDD

Any ideas?


CylonRed
Premium,MVM
join:2000-07-06
Bloom County
Do you have a spare PSU around?


Gordo74
Premium
join:2003-10-28
Monroeville, PA
said by CylonRed:

Do you have a spare PSU around?

That was my next step as well. I do have a 460w Cooler Master.

I'm trying to accumulate as many ideas as possible before I head to my brother's house tomorrow.


Veloslave
Geek For God
Premium
join:2003-07-11
Martinez, CA
reply to Gordo74
I would get rid of any and all OC'ing... back to defaults. Try new RAM

I don't think a PS is going to cause those errors.
--
Mom was right.... I NEED fiber!


CylonRed
Premium,MVM
join:2000-07-06
Bloom County
He already went back to defaults and it still had issues.


Gordo74
Premium
join:2003-10-28
Monroeville, PA
reply to Gordo74
Any other ideas before I head over today?


CylonRed
Premium,MVM
join:2000-07-06
Bloom County
For ram issues I think there are only 3 possibilities - bad ram, bad ram slot/motherboard, or bad PSU.


Gordo74
Premium
join:2003-10-28
Monroeville, PA
said by CylonRed:

For ram issues I think there are only 3 possibilities - bad ram, bad ram slot/motherboard, or bad PSU.

I'm not so sure it is RAM issues though. I've tested each stick individually and it still freezes in multiple different slots.


CylonRed
Premium,MVM
join:2000-07-06
Bloom County
You would need to check the ram in a known good PC, to know definitively if the ram is bad.


koitsu
Premium,MVM
join:2002-07-16
Mountain View, CA
kudos:23
reply to Gordo74
This problem looks to me like one of the following:

1) A UEFI or BIOS problem, pertaining to PCI memory range allocation overlaps,

2) A UEFI or BIOS problem pertaining to broken microcode or broken processor support for your exact model/revision of CPU (stepping matters),

3) A very, very faulty on-die MCH (this would require a CPU replacement),

4) An extremely faulty motherboard,

5) A PSU that has serious/major voltage problems (affecting the CPU and possibly DIMMs, and who knows what else).

Any of these could cause the behaviour you're seeing. I can assure you 100% the issue is not with your hard disk or with your graphics card. I'm about 99% certain the problem is not with your RAM/DIMMs.

You're going to need to start investing money into replacing parts until you can figure out which part is the faulty one.

Because the system is unreliable, doing a UEFI or BIOS update (which may include CPU microcode updates) is out of the question -- it may crash/fail in the middle of the update and result in a completely non-working system.

So the first piece to consider replacing is the motherboard or the PSU -- they're the cheapest ones. You will need to replace one thing at a time to rule out the exact piece (i.e. do not replace the motherboard and the PSU at the same time, because then you'll never know which of the two was the broken part). If you've replaced those and the issue continues, then you know it's the CPU (probably the on-die MCH, which may have been damaged as a result of overclocking).

My gut feeling says that the issue is probably #3, because the MCH is no longer on the motherboard but on-die on the CPU. I have lectured folks here on the forums in the past about overclocking. If you care about your hardware, and care about stability, simply put: do not overclock. Ever.
--
Making life hard for others since 1977.
I speak for myself and not my employer/affiliates of my employer.


Gordo74
Premium
join:2003-10-28
Monroeville, PA

1 edit
reply to Gordo74
This is a 2500K processor. It is literally made to overclock safely. Intel sells these for this sole purpose. It was at stock voltage at a 40x multiplier. Hardly earth shattering considering the i5-2500 (non K) turbos to these frequencies.

After some more testing and another PSU, as well as the RAM in another PC, I have whittled it down to either the motherboard, or, as koitsu mentioned, most likely the memory controller on the CPU die itself.

My brother is bringing his CPU over to me and I will try it in my system to see if it is the CPU itself or the motherboard.

n_w95482
Premium
join:2005-08-03
Ukiah, CA
I was going to say, 4 GHz for a 2500K is cake.

When you tested with one stick of RAM, did you try the same stick in all of the slots?
--
KI6RIT


koitsu
Premium,MVM
join:2002-07-16
Mountain View, CA
kudos:23
reply to Gordo74
There is no such thing as "safely overclocking". I am well aware of who the xxxxK series is marketed/catered to (cough cough), but there is absolutely no guarantee in any way/shape/form that it "guarantees safely overclocking". If you want stability and want to greatly diminish the chance of hardware going bad (specifically CPU, MCH, VRMs, RAM, or motherboard components), do not overclock. Otherwise this is the risk you take.

I look forward to seeing what the root cause turns out to be.

Also remember: using a different CPU (specifically if it's a different model or stepping) may actually "mask" a problem with a PSU (if the problem is with the PSU). So that's something to take into consideration as well.
--
Making life hard for others since 1977.
I speak for myself and not my employer/affiliates of my employer.


Gordo74
Premium
join:2003-10-28
Monroeville, PA
said by koitsu:

Also remember: using a different CPU (specifically if it's a different model or stepping) may actually "mask" a problem with a PSU (if the problem is with the PSU). So that's something to take into consideration as well.

I have tried three separate power supplies now. I believe the problem lies in the motherboard or memory controller on chip.

I'll update as this unfolds this week.


Gordo74
Premium
join:2003-10-28
Monroeville, PA
reply to n_w95482
said by n_w95482:

I was going to say, 4 GHz for a 2500K is cake.

When you tested with one stick of RAM, did you try the same stick in all of the slots?

There needs to be a stick in DIMM1 for the system to boot. I switched out each piece of RAM into that slot, as well as tried a known good stick of RAM in DIMM1 with the same errors.


koitsu
Premium,MVM
join:2002-07-16
Mountain View, CA
kudos:23
reply to Gordo74
said by Gordo74:

I have tried three separate power supplies now. I believe the problem lies in the motherboard or memory controller on chip.

I'll update as this unfolds this week.

Fully agree -- definitely one of those two components.
--
Making life hard for others since 1977.
I speak for myself and not my employer/affiliates of my employer.


Gordo74
Premium
join:2003-10-28
Monroeville, PA

1 recommendation

As a follow up, I tried 3 different procs in this board. All 3 had the same issue. I tested the chips in other motherboards and they worked fine. This leads me to believe it is the motherboard itself. It's the only part I have not tried swapping out.

bbear2
Premium
join:2003-10-06
94045
kudos:5
reply to CylonRed
said by CylonRed:

He already went back to defaults and it still had issues.

Exactly what process did you use to go back to defaults? If it was BIOS menu selection, then that's not good enough.

If you didn't try it, do try the "RESET1" button. I also recommend you use the "Clear CMOS Juniper JBAT1 (WITH POWER OFF).

There are also CPU Phase LEDs on your system. Have a look at those and tell me what they are displaying BEFORE you do the resets above and then after.

If your unsure about any of this, see the manual. »www.msi.com/product/mb/Z68A-GD80···v=Manual


Gordo74
Premium
join:2003-10-28
Monroeville, PA

1 recommendation

said by bbear2:

said by CylonRed:

He already went back to defaults and it still had issues.

Exactly what process did you use to go back to defaults? If it was BIOS menu selection, then that's not good enough.

If you didn't try it, do try the "RESET1" button. I also recommend you use the "Clear CMOS Juniper JBAT1 (WITH POWER OFF).

There are also CPU Phase LEDs on your system. Have a look at those and tell me what they are displaying BEFORE you do the resets above and then after.

If your unsure about any of this, see the manual. »www.msi.com/product/mb/Z68A-GD80···v=Manual

I did both the software as well as the battery pull, both times it stating that the BIOS was successfully reset to defaults.


moarram

@optonline.net
reply to Gordo74
»www.extremetech.com/computing/11···formance

Just saying if OC was so bad, pretty sure that's a great way to lose money...

OP how are you making 12G of memory? It likely is a mobo issue, just want to make sure the ram is set up right =p

dave
Premium,MVM
join:2000-05-04
not in ohio
kudos:8
Reviews:
·Verizon FiOS
reply to koitsu
said by koitsu:

There is no such thing as "safely overclocking".

Sure - it seems self-evident to me that if all 2500K processors could safely and reliably operate at 4GHz all of the time, Intel would be selling them as 4GHz processors rather than 3.3GHz processors.


Gordo74
Premium
join:2003-10-28
Monroeville, PA
reply to moarram
said by moarram :

»www.extremetech.com/computing/11···formance

Just saying if OC was so bad, pretty sure that's a great way to lose money...

OP how are you making 12G of memory? It likely is a mobo issue, just want to make sure the ram is set up right =p

2x 4GB, 2x2GB


Gordo74
Premium
join:2003-10-28
Monroeville, PA
reply to Gordo74
As an update, I changed the motherboard for a different one and the computer is up and running error free.

Thanks all for the sound board!


Gordo74
Premium
join:2003-10-28
Monroeville, PA

1 recommendation

reply to Gordo74
Just to update this thread - It was not entirely the motherboard's fault. Every hour, Windows would still freeze even though memtest was now clean.

I looked around a bit more and it seems the M4 SSD had a firmware issue that after 5000 hours powered on, it would freeze every hour. That was the problem that was encountered here. Updated the firmware and it was fine. No more issues.


C0deZer0
Oc'D To Rhythm And Police
Premium
join:2001-10-03
Tempe, AZ
said by Gordo74:

Just to update this thread - It was not entirely the motherboard's fault. Every hour, Windows would still freeze even though memtest was now clean.

I looked around a bit more and it seems the M4 SSD had a firmware issue that after 5000 hours powered on, it would freeze every hour. That was the problem that was encountered here. Updated the firmware and it was fine. No more issues.

Suddenly I'm afraid of my SSD RAID idea, since even getting SMART information from a RAID is next to impossible.

But i'm glad to see you were able to get it fixed.
--
Because, f*ck Sony