lkml.org 
[lkml]   [2003]   [Jan]   [16]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: [BENCHMARK] 2.5.58-mm1 with contest
Date
On Friday 17 Jan 2003 12:28 am, Nick Piggin wrote:
> Con Kolivas wrote:
> >-----BEGIN PGP SIGNED MESSAGE-----
> >Hash: SHA1
> >
> >Contest (0.61pre) benchmark results for 2.5.58-mm1
> >
> >no_load:
> >Kernel [runs] Time CPU% Loads LCPU% Ratio
> >2.5.55 2 78 96.2 0 0.0 1.00
> >2.5.56 3 79 96.2 0 0.0 1.00
> >2.5.57 3 79 96.2 0 0.0 1.00
> >2.5.58 2 79 96.2 0 0.0 1.00
> >2.5.58-mm1 4 79 96.2 0 0.0 1.00
> >cacherun:
> >Kernel [runs] Time CPU% Loads LCPU% Ratio
> >2.5.55 2 76 98.7 0 0.0 0.97
> >2.5.56 3 76 100.0 0 0.0 0.96
> >2.5.57 3 76 100.0 0 0.0 0.96
> >2.5.58 2 76 100.0 0 0.0 0.96
> >2.5.58-mm1 4 77 97.4 0 0.0 0.97
> >process_load:
> >Kernel [runs] Time CPU% Loads LCPU% Ratio
> >2.5.55 2 94 79.8 30 17.0 1.21
> >2.5.56 3 93 80.6 29 16.1 1.18
> >2.5.57 3 93 81.7 28 16.1 1.18
> >2.5.58 2 92 81.5 27 15.2 1.16
> >2.5.58-mm1 3 94 80.9 29 16.0 1.19
> >ctar_load:
> >Kernel [runs] Time CPU% Loads LCPU% Ratio
> >2.5.55 3 133 78.9 1 3.8 1.71
> >2.5.56 3 152 75.7 1 4.6 1.92
> >2.5.57 3 132 79.5 1 3.8 1.67
> >2.5.58 3 117 82.1 1 6.0 1.48
> >2.5.58-mm1 3 109 81.7 1 4.6 1.38
> >xtar_load:
> >Kernel [runs] Time CPU% Loads LCPU% Ratio
> >2.5.55 3 113 79.6 1 6.2 1.45
> >2.5.56 3 111 78.4 1 6.3 1.41
> >2.5.57 3 107 80.4 1 5.6 1.35
> >2.5.58 3 122 80.3 1 6.6 1.54
> >2.5.58-mm1 3 121 76.0 1 6.6 1.53
> >io_load:
> >Kernel [runs] Time CPU% Loads LCPU% Ratio
> >2.5.55 3 126 63.5 6 12.7 1.62
> >2.5.56 3 131 59.5 7 13.0 1.66
> >2.5.57 5 124 64.5 5 11.3 1.57
> >2.5.58 3 153 54.9 8 14.3 1.94
> >2.5.58-mm1 6 156 51.3 9 14.7 1.97
> >read_load:
> >Kernel [runs] Time CPU% Loads LCPU% Ratio
> >2.5.55 3 95 82.1 9 5.3 1.22
> >2.5.56 3 99 80.8 6 6.1 1.25
> >2.5.57 3 100 80.0 6 7.0 1.27
> >2.5.58 3 96 82.3 9 5.2 1.22
> >2.5.58-mm1 4 96 83.3 176189 6.2 1.22
>
> Something seems to have gone wrong with "Loads".

Thanks for the heads up. Must be falling asleep to miss that one. The run
seems to have progressed ok but something died in the output (Aggelos?).
After deleting that result (just in case) the results look like this:

2.5.58 3 96 82.3 9 5.2 1.22
2.5.58-mm1 3 100 81.0 6 6.0 1.27

>
> >list_load:
> >Kernel [runs] Time CPU% Loads LCPU% Ratio
> >2.5.55 3 91 84.6 0 8.8 1.17
> >2.5.56 3 91 84.6 0 8.8 1.15
> >2.5.57 3 91 84.6 0 8.8 1.15
> >2.5.58 3 91 85.7 0 8.8 1.15
> >2.5.58-mm1 2 92 83.7 0 9.8 1.16
> >mem_load:
> >Kernel [runs] Time CPU% Loads LCPU% Ratio
> >2.5.55 3 116 73.3 66 1.7 1.49
> >2.5.56 3 107 80.4 45 0.9 1.35
> >2.5.57 3 110 80.0 47 0.9 1.39
> >2.5.58 3 107 73.8 66 1.9 1.35
> >2.5.58-mm1 3 104 75.0 50 1.0 1.32
> >dbench_load:
> >Kernel [runs] Time CPU% Loads LCPU% Ratio
> >2.5.55 3 117 68.4 2 16.2 1.50
> >2.5.56 2 124 62.9 3 25.8 1.57
> >2.5.57 3 121 64.5 3 22.3 1.53
> >2.5.58 3 122 64.8 3 24.6 1.54
> >2.5.58-mm1 3 118 66.9 3 22.0 1.49
> >
> >Only statistically significant diff b/w 2.5.58 and 2.5.58-mm1 is in
> > ctar_load which is faster in mm1. Trend towards being faster in mem_load.
> >
> >Con
>
> Looks good, thanks.

No problem.

Will have to keep an eye out on the code for more bugs. So far the contest
code changeover is showing far less variability in the results (good job
Aggelos!) which should make for more meaningful results.

Con
-
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/

\
 
 \ /
  Last update: 2005-03-22 13:32    [W:0.081 / U:0.608 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site