Tweet
Hi all,
I work for an SI and we use Memtest on everything (literally, everything). It's not the be all and end all, obviously. But it gives us a good idea on system stability.
We've got four systems in progress at the moment with the new AMD Threadripper Pro CPUs (2x 3975WX, 2x 3995WX), all on the ASUS Pro WS WRX80E-SAGE SE WIFI mainboard.
The systems all work, fully populated with 16GB, 32GB and 64GB (3200MHz) ECC Registered parts (all Samsung original parts). However, one system needs 8x 128GB 2933MHz parts.
They're on the ASUS QVL list - M393AAG40M3B-CYFCQ - for 1, 2, 4 and 8 DIMM configurations.
If all eight modules are run together I'll get 1000+ errors and the big red FAIL. I broke them down to pairs and 3 of the four pairs report ECC Errors - but nothing for total errors, so I'm guessing the ECC nature of the DIMMS/system is correcting any problems at that point.
Ran individually, the DIMMs will all pass. I've taken four of the modules that tested for 60+ hours (no errors) and put them in a system together - and they produce ECC errors in <20 minutes.
I know it's not the same thing, but the built-in Windows tests takes two days - but completes with no errors.
Is my configuration just too new? could these be false positives?
I do feel I may have one faulty DIMM as it produces ECC errors on it's own (it's the only one)... but it's really odd.
Any ideas/thoughts would be very much appreciated.
I work for an SI and we use Memtest on everything (literally, everything). It's not the be all and end all, obviously. But it gives us a good idea on system stability.
We've got four systems in progress at the moment with the new AMD Threadripper Pro CPUs (2x 3975WX, 2x 3995WX), all on the ASUS Pro WS WRX80E-SAGE SE WIFI mainboard.
The systems all work, fully populated with 16GB, 32GB and 64GB (3200MHz) ECC Registered parts (all Samsung original parts). However, one system needs 8x 128GB 2933MHz parts.
They're on the ASUS QVL list - M393AAG40M3B-CYFCQ - for 1, 2, 4 and 8 DIMM configurations.
If all eight modules are run together I'll get 1000+ errors and the big red FAIL. I broke them down to pairs and 3 of the four pairs report ECC Errors - but nothing for total errors, so I'm guessing the ECC nature of the DIMMS/system is correcting any problems at that point.
Ran individually, the DIMMs will all pass. I've taken four of the modules that tested for 60+ hours (no errors) and put them in a system together - and they produce ECC errors in <20 minutes.
I know it's not the same thing, but the built-in Windows tests takes two days - but completes with no errors.
Is my configuration just too new? could these be false positives?
I do feel I may have one faulty DIMM as it produces ECC errors on it's own (it's the only one)... but it's really odd.
Any ideas/thoughts would be very much appreciated.
Comment