2.1.115 is not slower than 2.0.34

Andries.Brouwer@cwi.nl
Sat, 15 Aug 1998 10:25:00 +0200 (MET DST)

After I sent a patch improving the speed of ext2, I hoped
for people telling me about the effects it had. However,
it seems all activity on the list is about killing X.

So, I just did some timings myself.
This is on a 166MHz Pentium with 64MB - rather quiet at the time.
2.1.116a is a kernel with my ext2 patch.

Test 1: time du -s $HOME (a 700 MB tree, nothing unusual,
with directories containing dozens but not thousands of files).

kernel user system elapsed %CPU
2.0.34 0.57 9.49 1:04.62 15
2.0.34 0.34 9.68 1:02.90 15
2.1.114 0.48 8.05 1:02.16 13
2.1.116a 0.37 6.13 0:59.67 10
2.1.116a 0.38 5.95 1:00.34 10

Test 2: time du -s /tmp/tst (a directory containing 8000 empty files)

kernel user system elapsed %CPU
2.0.34 0.10 22.23 0:22.92 97
2.0.34 0.07 22.31 0:22.38 99
2.1.114 0.11 20.43 0:21.13 97
2.1.116a 0.06 10.43 0:11.01 95

All this is on first attempt. Doing this again with cache filled
eliminates most disk I/O. Also the system time decreases in the
cases where dcache helps.

Test 1:
kernel user system elapsed %CPU
2.0.34 0.40 9.31 0:11.75 82
2.1.114 0.37 7.09 0:07.45 97
2.1.116a 0.33 5.62 0:05.94 100
2.1.116a 0.31 4.99 0:05.31 99

Test 2:
kernel user system elapsed %CPU
2.0.34 0.06 22.12 0:22.18 99
2.0.34 0.07 22.31 0:22.38 99
2.1.114 0.05 3.60 0:03.64 100
2.1.116a 0.06 10.37 0:10.42 100
2.1.116a 0.03 2.80 0:02.82 100
2.1.116a 0.02 2.64 0:02.66 99

Conclusion:
1. In these tests there is no indication that 2.1.11* would
be slower than 2.0.34. It is systematically a fraction faster.
2. For 2.1.11* with my patch the system time is 20-50% smaller,
and this has a noticeable effect on the elapsed time.

Andries

-
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.altern.org/andrebalsa/doc/lkml-faq.html