readahead window size benchmarking

[yan cui]

Dear all:

    I am new to the Linux readahead system. After reading the recent
kernel code (3.8.8), I find there is an initial window size decided
based on the read size. And it will increase the window size to a
maximal value (32 pages) if the detected reading is sequential. To me,
the window size seems an important parameter to performance. However,
after testing using some home-made micro-benchmarks, I find that
performance is not sensible to this parameter. So, I wonder to know
what benchmarks (or applications) do you use to test the performance
difference when using different window size? Why not just use the
fixed window size (e.g., 32 pages)?


Thanks,
Yan

--
Think big; Dream impossible; Make it happen.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/