Linus Torvalds writes: (Summary) wrote:
Oh it definitely is expected to be slower, because it does the IPI to all the cores and actually gets their frequency right. It was the old one that we had to revert (because it did so sequentially) that was really bad, and took something like 2+ seconds on Ingo's 160-core thing, iirc.
on Ingo's 160-core thing, iirc.
It sounds like the current patch is slower, but likely acceptable considering that you get the right results now ..
considering that you get the right results now ..
Linus
Linus
Linus
[...]
So that patch is actually slower.Oh it definitely is expected to be slower, because it does the IPI to all the cores and actually gets their frequency right. It was the old one that we had to revert (because it did so sequentially) that was really bad, and took something like 2+ seconds on Ingo's 160-core thing, iirc.
on Ingo's 160-core thing, iirc.
It sounds like the current patch is slower, but likely acceptable considering that you get the right results now ..
considering that you get the right results now ..
Linus
Linus
Linus