Why a special implementation for SMP? I gather that the load balancing should make swapping out pages scale linearly with the number of drives and swapping in pages should be slightly improved.
At least I'd hope this was the case - since I've put swap partitions at the start of both disks in my machine.
I don't see that giving a swap file to an individual processor would help, unless you could give each CPU its own disk controller.
But I imagine what might happen in the case where one CPU is being monopolised by a fairly numerically intensive task while the other is handling more mundane task switching and swapping. The first CPU may never need to swap and so it's disk r/w activity will have bandwidth to spare, while the second CPU may find itself limited by it's ability to swap pages to it's single disk rather than utilising the spare disk throughput on the second disk.
OK... am I wrong? Stupid? Uninformed?
:-)
Scott Manley (aka Szyzyg) /------ _@/ Mail -----\
___ _ _ __ __ _ | Armagh Observatory |
/ __| __ ___| |_| |_ | \/ |__ _ _ _ | |___ _ _ | Armagh |
\__ \/ _/ _ \ _| _| | |\/| / _` | ' \| / -_) || | | Northern Ireland |
|___/\__\___/\__|\__| |_| |_\__,_|_||_|_\___|\_, | | BT61 9DG. |
http://star.arm.ac.uk/~spm/welcome.html |__/ \=====================/
-
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/