[BENCHMARK] 2.5.40{-mm1} with contest 0.42

Con Kolivas (conman@kolivas.net)
Tue, 1 Oct 2002 23:38:57 +1000


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Here follow the latest benchmarks for 2.5.40 and 2.5.40-mm1:

noload:
Kernel Time CPU% Ratio
2.4.19 67.7 98 1.00
2.5.38 72.4 94 1.07
2.5.38-mm3 73.0 93 1.08
2.5.39 73.2 93 1.08
2.5.39-mm1 73.0 94 1.08
2.5.40 73.2 94 1.08
2.5.40-mm1 73.4 93 1.08

process_load:
Kernel Time CPU% Ratio
2.4.19 110.8 57 1.64
2.5.38 85.7 79 1.27
2.5.38-mm3 96.3 72 1.42
2.5.39 88.9 76 1.31
2.5.39-mm1 99.0 69 1.46
2.5.40 96.4 71 1.43
2.5.40-mm1 94.5 72 1.40

io_load:
Kernel Time CPU% Ratio
2.4.19 216.1 33 3.19
2.5.38 887.8 8 13.11
2.5.38-mm3 105.2 70 1.55
2.5.39 229.4 34 3.39
2.5.39-mm1 239.5 32 3.54
2.5.40 230.3 33 3.40
2.5.40-mm1 117.5 62 1.74

mem_load:
Kernel Time CPU% Ratio
2.4.19 105.4 70 1.56
2.5.38 107.9 73 1.59
2.5.38-mm3 117.1 63 1.73
2.5.39 103.7 72 1.53
2.5.39-mm1 104.6 73 1.54
2.5.40 103.0 73 1.52
2.5.40-mm1 111.3 67 1.64

Summary:
2.5.40 shows no significant difference from 2.5.39 (except in process_load;
now equivalent to 2.5.39-mm1 - the significance of which is not clear of it's
importance)

2.5.40-mm1 shows a substantial improvement under IO load (consistent with the
change in fifo_batch as discussed under the thread "[BENCHMARK] 2.5.39-mm1")
2.5.40-mm1 also shows a slight drop in performance under mem_load. Note of
interest that both 2.5.38-mm3 and 2.5.40-mm1 which have fifo_batch set to 16
instead of 32 have both better IO load performance, and worse mem_load
performance - related?

Con
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQE9maV0F6dfvkL3i1gRAjQZAJ9Ue60e1pe9A/WKigGW3UiMfshngACfT/FW
cpLOyqDqFq7LkaTCMJo51l8=
=yrf+
-----END PGP SIGNATURE-----
-
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/