> The main finding of this paper is that none of the existing mitigations is fully adequate, not even ECC. I suspect AMD's new memory encryption would provide

> The main finding of this paper is that none of the existing mitigations is fully adequate, not even ECC. I suspect AMD's new memory encryption would provide significant protection as specific bit flips in rows that do not share the same key would be totally unpredictable.