Hi All,
I am running lmbench version-3.0-a9 on two embedded platforms.
P1 - Xscale processor with 2.6.21.1 linux kernel
P2 - OMAP3430 processor with 2.6.24.7 linux kernel
|
File size |
P1 |
P2 |
|
1MB |
10.4805 |
8.9601 microsec |
|
3 |
11.3464 |
17.9124 |
|
4 |
11.7725 |
22.4209 |
|
6 |
12.6081 |
31.5996 |
|
8 |
13.4312 |
40.4351 |
|
10 |
14.2652 |
49.4035 |
|
12 MB |
15.0599 |
58.4115 |
|
|
|
|
P1 |
free -m
total used free shared buffers cached
Mem: 109 41 67 0 2 17
-/+ buffers/cache: 21 88
Swap: 0 0 0
|
|
P2 |
free -m
total used free shared buffers cached
Mem: 122 43 79 0 5 22
-/+ buffers/cache: 15 107
Swap: 0 0 0
|
I want to know why the cost of page fault in case of P2 is more as the file size increases.
But I dont know where to start. Please give directions
Thanks for your time
Regards
Basavaraj
