https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87444
--- Comment #2 from George <studio at anchev dot net> --- Forgive me, I am not a developer and I am not aware how this was designed. But may I ask: why was it designed to be wrong and why only on particular CPUs? Also: what happens when a wrong l2-cache-size is used? Won't that casue issues with the final binary? (e.g. instability, crashing, I don't know...)
