>> Anyway the 2.3.36 SMP scheduler here seems to trashes caches too much so
>> I'll probably check all the scheduler heuristics before doing any bench.
>> At least it's not obviously buggy as the 2.2.x one ;).
>
>I'm just about to send Linus on the IBM patch for the goodness() cache
>behaviour. It cuts goodness() down to referencing one cache line per task
Whith "caches" I was not talking about the kernel but about userspace.
About the task_struct reordering to take the goodness checked fields in
the same cachelines it's an obviously right idea (and also an obviously
right patch) of course :). I have not checked the implementation
carefully yet.
Andrea
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/