These numbers show a 30% reduction in rmap overhead with my patch,
close to what I originally reported:
((35.904 + 35.690 + 35.921) - (27.814 + 27.763 + 27.861)) /
((39.316 + 39.405 + 39.062) - (27.814 + 27.763 + 27.861)) ~= .70
But they also show that rmap overhead is around 29% on your box,
even with my patch:
(35.904 + 35.690 + 35.921) / (27.814 + 27.763 + 27.861) ~= 1.29
Granted, it's still way too high, and we are still in search of the
'dark cycles'.
Did we do an apples-to-apples comparison of 2.4 to 2.5? Because if
we did, then going by your numbers, 2.5.26 is already considerably
worse than 2.4.19-pre7:
((30.260 + 29.642)/2) / ((27.814 + 27.763 + 27.861)/3) ~= 1.08
Is it fair to compare your 2.4 vs 2.5 numbers?
-- Daniel - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/