-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 01/29/2017 09:38 AM, Peter Stuge wrote:
Timothy Pearson wrote:
Populating all four slots closest to CPU1 with large ECC registered DIMMs is a surefire way to recreate the instability -- note a training failure is not common, the main issue is that the marginal routing causes severe memory corruption when the BKDG-recommended algorithms are used.
What does the vendor BIOS do?
Unknown. I would imagine there is a hack in the vendor BIOS to work around this board-specific issue, but that doesn't help coreboot very much. We had looked into it before, but decided it was not worth the cost to research and implement a fix.
For what it's worth, there are several threads online about the KGPE-D16 and memory corruption with the vendor BIOS. It seems to have taken ASUS some time to iterate and work around the problem.
- -- Timothy Pearson Raptor Engineering +1 (415) 727-8645 (direct line) +1 (512) 690-0200 (switchboard) https://www.raptorengineering.com