>>>>> "andrea" == Andrea Arcangeli <andrea@suse.de> writes:
Hi
andrea> It gives me smoother swap behaviour since the swap cache hardly pollutes
andrea> the lru_cache now.
Andrea I have run here last night your patch (classzone-18) against
pre7-3 in one machine and vanilla pre7-3 in other machine. The test
was from my memtest suite:
while (true); do time ./mmap002; done
suited to the memory of my system (Uniprocessor 96MB)
The results are very good for your patch:
Vanilla pre7-3 pre7-3+classzone-18
real 3m29.926s real 2m10.210s
user 0m15.280s real 2m10.210s
sys 0m20.500s real 2m10.210s
That are the normal times. classzone patches variations are very low
(all the iterations are between 2m08 and 2m10). But in vanilla
pre7-3, the variations are higher: between 3m4 and 4m20, and the worst
part, when the kswapd problem appear, the program takes until 36m20
(yes 36, ten times more, is not a typo). Furthermore, vanilla pre7-3
kill the process after 2 hours and a half, classzone works for more
than 12 hours without a problem.
That is the description of the situation. Andrea, why do you reverse
the patch of filemap.c:truncate_inode_pages() of using TryLockPage()?
That change was proposed by Rik due to an Oops that I get here. It
was one of the non-easily reproducible ones that I get.
Later, Juan.
-- In theory, practice and theory are the same, but in practice they are different -- Larry McVoy- 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/
This archive was generated by hypermail 2b29 : Sun May 07 2000 - 21:00:14 EST