Re: [PATCH 00 of 66] Transparent Hugepage Support #32

[caiqian]

There were some changes of behaviours with THP and KSM statistics demonstrated by this program, http://people.redhat.com/qcai/ksm01.c. 

There are 3 programs (A, B ,C) to allocate 128M memory each using KSM.
A has memory content = 'c'.
B has memory content = 'a'.
C has memory content = 'a'.
Then without THP,
pages_shared = 2
pages_sharing = 98285
pages_sharing = 98292
pages_unshared = 0
pages_volatile = 17
pages_to_scan = 98304
sleep_millisecs = 0
with THP,
pages_shared is 2.
pages_sharing is 18422.
pages_unshared is 0.
pages_volatile is 8.

Later,
A has memory content = 'c'
B has memory content = 'b'
C has memory content = 'a'.
Then without THP,
pages_shared = 3
pages_sharing = 98296
pages_unshared = 0
pages_volatile = 5
with THP,
pages_shared = 3
pages_sharing = 16358
pages_unshared = 0
pages_volatile = 23

Later,
A has memory content = 'd'
B has memory content = 'd'
C has memory content = 'd'
Then without THP,
pages_shared = 1
pages_sharing = 98274
pages_unshared = 0
pages_volatile = 29
with THP,
pages_shared = 1
pages_sharing = 8668
pages_unshared = 0
pages_volatile = 35

Finally,
A changes one page to 'e'
Then without THP,
pages_shared = 1
pages_sharing = 98274
pages_unshared = 1
pages_volatile = 28
with THP,
pages_shared = 1
pages_sharing = 8163
pages_unshared = 1
pages_volatile = 27

Are those differences for pages_sharing between with and without THP are expected?

CAI Qian
--
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/