Ryzen RAM issue, motherboard or processor at fault?

Jimmy Weirdarms · 14 Feb 2021 at 19:04

Hey folks,

I built a Ryzen system a year and a half ago, with an MSI x570-a pro (yeah the cheap x570

, a 3800x and 4x16GB sticks of HyperX Predator RGB 3200MHz DDR4. It's been working fine. Sometime in December I upgraded the processor to a 5950x.

Last week, the system stopped booting and the "RAM" light on the "EZ-Debug" panel lit up. I took everything out and reseated it, then it booted fine, but every couple of hours my (linux) system would hard lock, and one in three reboots would still give the RAM light and no boot.

So I thought it was a dead stick. I tested each of the sticks individually in slot "B2", which is easiest to get at, with memtest86, and they all passed. I put the same stick in "A1" and it boots but memtest86 reports some errors. Back in "B2" it's working fine. It works fine in "A2", with it in "B1" the system didn't boot first time but has now memtested just fine on a retry, I think that failure was an aberration.

So ... it looks like the RAM itself is probably fine, but the "A1" (leftmost) slot is causing issues. Question is, does this point to a motherboard problem or a CPU problem, with Ryzen?

I *could* try the 3700X from my partner's PC but swapping CPUs is a hassle I'd rather avoid, but if it's the only way to diagnose I guess I could

denigor777 · 14 Feb 2021 at 22:25

I think you are taking the right approach, carefully moving elements around to logically isolate the problem.

More things to try could be:
- running the ram at lower speed. To see if the issue is always-there or if it's down to timings and/or voltages.
- trying your partners RAM rather than CPU, if it's easier to swap out.

I's also suggest many internet searches for signs of others seeing your issue.

As you've found a way to reproduce the issue (any stick in slot 1 causes an observable error) I think you are close to resolving this overall. You do have the latest bios release in?

Jimmy Weirdarms · 15 Feb 2021 at 11:01

Yeah I'm on the latest BIOS. That's one of the things I tried to see if it would help

Unfortunately this morning it's decided not to boot with RAM in the other three slots (beepcode, debug light). I'm thinking slot 2 is also unreliable. Dammit, really am going to have to swap CPUs around, and various other things. I'll comb the web today looking for similar stuff I guess. At this point I think it's an RMA, I'm just not sure which part!

mickyflinn · 15 Feb 2021 at 12:49

Jimmy Weirdarms said:
Yeah I'm on the latest BIOS. That's one of the things I tried to see if it would help

Unfortunately this morning it's decided not to boot with RAM in the other three slots (beepcode, debug light). I'm thinking slot 2 is also unreliable. Dammit, really am going to have to swap CPUs around, and various other things. I'll comb the web today looking for similar stuff I guess. At this point I think it's an RMA, I'm just not sure which part!

Swap the cpu around and use you wifes in your system if compatable.

Could your cpu cooler be to tight ?

Jimmy Weirdarms · 15 Feb 2021 at 12:58

mickyflinn said:
Could your cpu cooler be to tight ?

I don't think so, but I guess it's a possibility. I guess if that's the case, whatever it is squashing is now wrecked...

mickyflinn · 15 Feb 2021 at 13:03

Jimmy Weirdarms said:
I don't think so, but I guess it's a possibility. I guess if that's the case, whatever it is squashing is now wrecked...

If overtight it could be flexing the motherboard causing errors doesn't mean it's permenant damaged but worth looking at.

Jimmy Weirdarms · 15 Feb 2021 at 13:13

Fair enough. Worth checking out then just in case. It's an AIO 240mm with a fairly good mount, but as I'm going to be taking it off anyway, a quick loosen won't hurt.

It's been on there over a year though, and the processor has been running fine since December. And yeah, the processors are compatible, we both have the same motherboard.

mickyflinn · 15 Feb 2021 at 13:18

Jimmy Weirdarms said:
Fair enough. Worth checking out then just in case. It's an AIO 240mm with a fairly good mount, but as I'm going to be taking it off anyway, a quick loosen won't hurt.

It's been on there over a year though, and the processor has been running fine since December. And yeah, the processors are compatible, we both have the same motherboard.

Test the memory in your wifes pc also see if it recreates the problem.

Test cpu first if all good then add the memory.