Errors in Test 7 and 9 - CPU or Motherboard?

[k3rn3l]

Hey all,

So first off, Ive tested the ram in another machine and it received no errors, so this is why Im trying to figure out if its CPU or Mobo. I get hundreds of errors in tests 7 and 9 and a few in test 6. The reason I started doing the tests is because my computer has been randomly shutting off when under load, and more so when ram usage was high. Ive eliminated everything but CPU and Mobo, so now am trying to narrow it down to which one is the culprit. Unfortunately I do not have another board to test the cpu, or it would be much easier.

Specs:
CPU: AMD FX-8350 (Not currently OCd)
Mobo: Gigabyte 990FX-UD3
RAM: 16GB Corsair Vengeance 1866 (2x 8GB)

When I run tests like OCCT or Prime95 using very little ram, the tests go for quite a while before the system shuts down. If I run those tests with high ram usage, the system shuts down almost immediately. Now, due to my chip having an integrated memory controller, I have been leaning towards the CPU being the issue, but Ive been told recently that, "Yes, it has an integrated memory controller but the CPU and memory still needs DRAM voltages from the motherboard and the motherboard still provides the electrical connections between the CPU and RAM" , so that kind of threw me and now Im wondering if it may infact be the Mobo.

Any help would be much appreciated, and if any other info is required, please let me know.

TIA,

Brad

[David (PassMark)]

Clearly something is wrong.
You could try,

• Underclocking the RAM slightly.
• Raise the voltage levels on the RAM slightly
• Trying different RAM in the same MB. It could be more of an incompatibility issue rather than an outright fault.

But if the fault doesn't go away, there isn't much other choice that to swap the CPU or MB. I don't think there is any tests you can do in software to say with certainty which part is at fault. Might even be something more obscure, like power supply issues, EMI or overheating.

[k3rn3l]

Clearly something is wrong.
You could try,

• Underclocking the RAM slightly.
• Raise the voltage levels on the RAM slightly
• Trying different RAM in the same MB. It could be more of an incompatibility issue rather than an outright fault.

But if the fault doesn't go away, there isn't much other choice that to swap the CPU or MB. I don't think there is any tests you can do in software to say with certainty which part is at fault. Might even be something more obscure, like power supply issues, EMI or overheating.

Thank you for the reply. This is what Ive been thinking but was hoping someone would say something else, lol. So either CPU or Mobo. Ive not heard of any compatibility issues with this configuration, but will do some further research. I have swapped the PSU, so Im fairly confident its not that. As far as over heating, I had been doing some fairly substantial OCs, upwards of 5.0Ghz at ~1.6v on the Vcore, so employed some extreme cooling. At Idle I was seeing sub zero temps, under load it didnt go above ~50c, so also confident its not an overheating issue. Ive also forced this issue to happen by running OCCT/Prime95 and monitored the temps before shutdown and they were anywhere from ~30-~50c.

Well, I guess this is a good excuse to get both a new CPU and Mobo, lol.

Thanks again.

[David (PassMark)]

I had been doing some fairly substantial OCs, upwards of 5.0Ghz.... At Idle I was seeing sub zero temps....

Sounds like the temperature cycling tests that a lot of electronics vendors do, when they want to break things.

[k3rn3l]

Sound like the temperature cycling tests that a lot of electronics vendors do, when they want to break things.

Whoops, Lol. Its actually interesting that you mention that though. Ive been trying to remember when this all started happening, as its been going on for quite a while, and now that I think about it, I believe this did in fact start around the time I started getting sub zero temps on the CPU. That being said.. I guess the more likely component to have been damaged by that would be the processor. Would you agree?

Thanks again.

[David (PassMark)]

I don't really know.
Large temperature changes causes expansion and contraction of the components effected. Could be a bad solder joint, could be something internal in the CPU. Generally there is no way to fix any of this stuff outside of the factory environment. So you are left with option to swap out components until the problem goes away. Maybe you can pick up a very cheap 2nd hand Phenom CPU to see if it is the CPU at fault.

[orioon]

Some components may behave off when you reach sub zero temperatures.
When putting a GPU on LN2 for example it is often required to heat up the memory in order to still reach high stable clocks.
Could be the same with normal system memory I assume.

Also some chips don't like cold at all. This was referred as "Cold Bug" in the past.
(e.g. no POST when temp is below -30°C or unstable operation)

What is your exact temperature?
Did not see much issues when cooling with dry ice for example.