X10SDV-TLN4F, Crucial Rdimm, ECC errors

[chalde]

Hi all

I've recently bought a Supermicro X10SDV-TLN4F (rev-b) and 2x32gb crucial Ddr4 ecc rdimm, but I'm receiving a lot of ecc correctable errors.

The errors are thrown semi-randomly. Some of the errors are reported in bios/IPMI, but most of them are only reported in the OS.

I've conducted some tests trying to find the root course but out of luck.
I can't seem to reproduce the errors in memtest86+. Has been running for +12 hours without errors, 2 full passes at least, both with and without multicore. No errors reported in either memtest or bios.
However, when running either vmware or Windows straight on the metal, errors are occurring, mostly under heavy load. Vmware at random and in Windows more or less straight away after starting ram test in passmark burnintest.
Also when idling in Windows, errors are produced at random (but not reported by bios/IPMI). Temperatures /voltages within thresholds at all time.
I don't have access to test either new ram or ram in different setup right now (might be able to test ram in a HP ProLiant later on).
UPDATE: Got access to 2x8gb hynix ddr4-rdimm, hma41gr7mfr4n-tf (on the QVL), this should rule out compatability.
System is stable, but the errors corrected seems excessive for a new system (and i guess, at some time the errors will go from corrected to uncorrected).
Any ideas? Bad ram, bad motherboard or something else? Are the memory and board simply not compatible, even though they should be according to specs?

Full setup:
Motherboard: Supermicro X10SDV-TLN4F (rev-b, d-1541), newest bios (1.1)
Ram: 2x32gb crucial Ddr4-2133 rdimm, CT32G4RFD4213
Disks: 240gb Ocz Trion, 120gb Corsair neutron, 1tb Wd black, 1tb Seagate 7200.10
Case: Corsair 250d
Psu: Be quiet! L8-300W

Test results:
DIMM1A(module1) + DIMM1B(module2)
21:28 Passmark started
21:29 several errors corrected in windows
21:31 ECC corrected in bios
multiple errors corrected in windows log, not logged in bios

DIMM1A(module1)
21:53 Passmark started
21:54 multiple errors corrected in windows
21:55 ECC corrected in bios
multiple errors corrected in windows log, not logged in bios

DIMM1A(module2)
22:08 passmark started
22:09 multiple errors corrected in windows
22:10 multiple ECC corrected in bios
multiple errors corrected in windows log, not logged in bios

DIMM1A(module2)+DIMM1B(module1)
22:21 passmark started
22:25 multiple errors corrected in windows
multiple errors corrected in windows log, not logged in bios

DIMM1A(module2)+DIMM1B(module1)
22:37 passmark started
no errors

DIMM1A(module2)+DIMM1B(module1)
22:45 passmark started
no errors

DIMM1A(module2)+DIMM1B(module1)
idling in windows
23:03 error corrected in windows, not logged in bios
01:20 error corrected in windows, not logged in bios
06:11 error corrected in windows, not logged in bios
06:12 error corrected in windows, not logged in bios

Further testing with 2x8gb hynix ddr4-rdimm

DIMM1A(hynix)+DIMM1B(hynix)
16:10 passmark started
no errors

DIMM1A(crucial2)+DIMM1B(crucial1)
16:20 passmark started
16:24 error corrected in windows, not logged in bios
16:28 multiple errors corrected in windows, not logged in bios
16:29 error corrected in windows, not logged in bios
16:30 error corrected in windows, not logged in bios

DIMM1A(hynix)+DIMM1B(hynix)
16:38 passmark started
17:00 3 passes, no errors

DIMM1A(hynix)+DIMM1B(hynix)+DIMM2A(crucial2)+DIMM2B(crucial1)
Hangs on post (PEI--IPMI init)
No log in bios

DIMM1A(crucial2)+DIMM1B(crucial1)+DIMM2A(hynix)+DIMM2B(hynix)
Hangs on post (PEI--IPMI init)
No log in bios

DIMM1A(hynix)+DIMM1B(hynix)
17:20 passmark started
17:40 3 passes, no errors

DIMM1A(hynix)
17:48 passmark started
18:15 3 passes, no errors

DIMM1A(crucial2)
18:22 passmark started
18:30 First run passed
18:40 Second run passed

DIMM1A(crucial1)
18:44 passmark started
18:45 multiple errors corrected in windows, not logged in bios

DIMM1A(crucial2)
18:50 passmark started
19:01 error corrected in windows, not logged in bios

DIMM1A(hynix)+DIMM1B(hynix)
19:08 passmark started
20:15 7 passes, no errors

DIMM1A(crucial1)+DIMM1B(crucial2)
20:16 passmark started
20:18 multiple errors corrected in windows, not logged in bios

DIMM1A(crucial2)
21:08 passmark started
21:45 No errors

 

[chalde]

My conclusion so far:
- The Hynix (from SM's QVL) does not introduce ECC corrections.
- When operating with both Crucial modules, errors occur on both modules more or less instantaneous
- When operating with Crucial module 1 alone, errors occur more or less instantaneous
- When operating with Crucial module 2 alone, errors does not occur or might take a long time before occuring

It doesn't seem to be related to the motherboard or any other hardware.
So are we talking bad memory or compatibility?
If compatability, would the errors be more systematic and/or the same rate for both Crucial modules?

 

[Terry Kennedy]

Any ideas? Bad ram, bad motherboard or something else? Are the memory and board simply not compatible, even though they should be according to specs?

I'd try swapping the power supply with a different one.

 

[chalde]

I'd try swapping the power supply with a different one.

I will test later today with a different PSU.

 

[chalde]

Picked up 2x32gb kingston branded hynix instead. KVR21R15D4/32 with partnumber HMA84GR7MFR4N-TF (is on the QVL).
Running a couple of tests. So far no errors.
Will return the Crucial memory.

 

[chalde]

Has been running stable for about 2 days now, so either the Crucial was faulty or incompatible.