Test found 2 errors a few seconds after starting

[paulunderwood]

Do you have the RAM in the right slots?

[Beinish]

Originally Posted by paulunderwood

Do you have the RAM in the right slots?

Yes. In slots 2 and 4.

[paulunderwood]

Have you monitored the CPU temperature (with installed software) when running Prime95?

[Beinish]

Originally Posted by paulunderwood

Have you monitored the CPU temperature (with installed software) when running Prime95?

Yes, on small FFTs it reached a max of 95C but averaged at 92-93C (max temp for this CPU is rated for 95C).
On large FFTs it seems to max at 91-92C and averaging at less.
I ran small FFTs for 6-7 hours a couple weeks back and Large FFTs for one hour today, which was the first time.

Never ran Blend. I don't know which settings I should give it.

[paulunderwood]

That seems on the high side for temperature. What cooler and paste did you use when fitting the cooler?

[Beinish]

Originally Posted by paulunderwood

That seems on the high side for temperature. What cooler and paste did you use when fitting the cooler?

Stock Wraith cooler with preapplied paste.
I never reach high temps while gaming, which seems to cause my crashes mostly, I know these are high temps generally but it's a stress test so it makes sense. By the way, on the small FFTs test there were no errors.

[paulunderwood]

Originally Posted by Beinish

Stock Wraith cooler with preapplied paste.
I never reach high temps while gaming, which seems to cause my crashes mostly, I know these are high temps generally but it's a stress test so it makes sense. By the way, on the small FFTs test there were no errors.

[Prime95]

If small FFTs ran OK, temps are not the issue. I bet it's the RAM (or the RAM controller). Up that RAM voltage or lower that RAM speed.

[Beinish]

Originally Posted by Prime95

If small FFTs ran OK, temps are not the issue. I bet it's the RAM (or the RAM controller). Up that RAM voltage or lower that RAM speed.

Upped the voltage to 1.30 on 2133mhz. Looks like worker #1 and #11 are erroring out immediately this time. Previously it was #1 and #2. Then after 15 minutes, #12.

I can keep this running for a bit more, or maybe up the voltage to 1.35 and test again.
If that one will error out, I can return it to Auto on the voltage and just run the RAM at a lower speed.

Thoughts?

[moebius]

I have samsung dies too and they run stable at Cl 16 and 1.36V with 3200Mhz. I think you've got a bad ram-kit (if not, it's your Mobo, i personally use MSI X470 priced around 100 Euro)

[Beinish]

Originally Posted by moebius

I have samsung dies too and they run stable at Cl 16 and 1.36V with 3200Mhz. I think you've got a bad ram-kit (if not, it's your Mobo, i personally use MSI X470 priced around 100 Euro)

I might just go ahead and buy new RAM tomorrow, but I would love to be certain that it isn't just a bad CPU/Mobo. The local store where I bought the CPU/Mobo combo told me to bring them in for testing, but I wasn't able to get errors with any other stress test I used (AIDA64, OCCT). Googling the "rounding error" in Prime95 gave me mixed results thinking it's either the memory or CPU.

I don't know, would love the input of more people before I got ahead and buy new hardware.

Originally Posted by Prime95

If small FFTs ran OK, temps are not the issue. I bet it's the RAM (or the RAM controller). Up that RAM voltage or lower that RAM speed.

Upped it to 1.35, no change.
Then I set it down to 1833mhz(?) and still no change. Then to 1600mhz, no change,then 1600 + 1.35v dram, no change.

I reset to the default settings,
I pulled out one stick, same result.
I put in the other stick, same result.