Brian Williams wrote:
> Hi.
> Make sure you are comparing apples with apples.
>
> 3 consecutive runs with parameter c = 0 gives:
>
> /dev/hda:
> Timing buffered disk reads: 64 MB in 34.29 seconds = 1.87 MB/sec
>
> /dev/hda:
> Timing buffered disk reads: 64 MB in 8.86 seconds = 7.22 MB/sec
>
> /dev/hda:
> Timing buffered disk reads: 64 MB in 9.73 seconds = 6.58 MB/sec
>
> The difference is for the second and third tests, data is cached in the
> disk. After making sure the cache would have been cleared, I rerun with
> c =3 and get:
>
> /dev/hda:
> Timing buffered disk reads: 64 MB in 24.30 seconds = 2.63 MB/sec
>
> /dev/hda:
> Timing buffered disk reads: 64 MB in 6.16 seconds = 10.39 MB/sec
>
> /dev/hda:
> Timing buffered disk reads: 64 MB in 5.73 seconds = 11.17 MB/sec
>
> It is clear that I always get a performance improvement. However, if you
> run the test, change the value then rerun the test, you will not get a
> valid comparison.
It would seem that for a valid comparison, cacheing should be temporarily
turned off. Is there a way to do that?
Paul
_______________________________________________
Win4Lin-users mailing list
[EMAIL PROTECTED]
https://lists.netraverse.com/mailman/listinfo/win4lin-users
