Messages in this thread | | | From | Con Kolivas <> | Date | Sat, 23 Nov 2002 23:30:27 +1100 | Subject | [BENCHMARK] 2.4.19-rmap15 with contest |
| |
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Here is a set of contest (http://contest.kolivas.net) benchmarks for 2.4.19-rmap15
noload: Kernel [runs] Time CPU% Loads LCPU% Ratio 2.4.18 [5] 71.7 93 0 0 0.99 2.4.19 [5] 69.0 97 0 0 0.95 2.4.19-rmap15 [3] 73.1 92 0 0 1.01
cacherun: Kernel [runs] Time CPU% Loads LCPU% Ratio 2.4.18 [2] 66.6 99 0 0 0.92 2.4.19 [2] 68.0 99 0 0 0.94 2.4.19-rmap15 [3] 67.7 99 0 0 0.94
process_load: Kernel [runs] Time CPU% Loads LCPU% Ratio 2.4.18 [3] 109.5 57 119 44 1.51 2.4.19 [3] 106.5 59 112 43 1.47 2.4.19-rmap15 [3] 112.7 56 123 44 1.56
Marginally slower here.
ctar_load: Kernel [runs] Time CPU% Loads LCPU% Ratio 2.4.18 [3] 117.4 63 1 7 1.62 2.4.19 [2] 106.5 70 1 8 1.47 2.4.19-rmap15 [3] 98.4 77 1 6 1.36
First significant difference. Notably faster while creating tars.
xtar_load: Kernel [runs] Time CPU% Loads LCPU% Ratio 2.4.18 [3] 150.8 49 2 8 2.09 2.4.19 [1] 132.4 55 2 9 1.83 2.4.19-rmap15 [3] 130.2 55 1 19 1.80
io_load: Kernel [runs] Time CPU% Loads LCPU% Ratio 2.4.18 [3] 474.1 15 36 10 6.56 2.4.19 [3] 492.6 14 38 10 6.81 2.4.19-rmap15 [3] 222.9 33 13 9 3.08
Well this is nice. io_load is the most felt of the slowdowns and rmap manages to blunt the effect io load has (note io load uses quite a bit of ram to do the io load).
read_load: Kernel [runs] Time CPU% Loads LCPU% Ratio 2.4.18 [3] 102.3 70 6 3 1.41 2.4.19 [2] 134.1 54 14 5 1.85 2.4.19-rmap15 [3] 129.5 56 20 5 1.79
list_load: Kernel [runs] Time CPU% Loads LCPU% Ratio 2.4.18 [3] 90.2 76 1 17 1.25 2.4.19 [1] 89.8 77 1 20 1.24 2.4.19-rmap15 [3] 90.4 76 0 12 1.25
mem_load: Kernel [runs] Time CPU% Loads LCPU% Ratio 2.4.18 [3] 103.3 70 32 3 1.43 2.4.19 [3] 100.0 72 33 3 1.38 2.4.19-rmap15 [3] 105.3 72 37 5 1.46
Note that while mem_load is marginally slower with rmap, the machine was in a much more usable state after running the mem_load. Much less was coming back from swap upon using it after the benchmark, and contest cant show this effect.
Looking good Rik.
Con -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.0 (GNU/Linux)
iD8DBQE933TmF6dfvkL3i1gRAsPGAJsGcgRoKaMUtIVebjIHlFeyuzBgeACePG7N GUhW6kzYg7amphsfe4W5GQ4= =+NY1 -----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/
| |