size (Kb) | number of procs | time (microseconds)
threads | processes
-------------------------------------------------
1 | 2 | 4 | 5
128 | 2 | 8 | 205
1 | 4 | 5 | 6
64 | 4 | 11 | 107
128 | 4 | 13 | 415
1 | 64 | 7 | 9
128 | 64 | 21 | 1063
Column 3 is explained by the fact that the same mapped buffer for the
pipe bytes are being used by all threads in your implementation (if
this were not the case, these numbers make no sense at all), thus it
always hits the cache, the increasing numbers show cache misses for
the task_struct's in the kernel for each thread.
Column 4 is explained by the fact we lack any L2 page coloring in the
kernel still.
Later,
David S. Miller
davem@dm.cobaltmicro.com
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu