|
At 07:02 01/10/02 +1000, Tim Churches wrote (in part):
>On a machine with sufficient memory (and Paul's machine has 1GB of RAM),
>the entire 100MB test file will be cached in the in-memory disc buffer
>maintained by the operating system (yes, even Windows does this). So no
>disc access is involved at all. In which case 0.81 secs is OK, but hardly
>amazing. Remember that CPU-to-memory bandwidth of modern PCs is measured
>in GB/sec.
Indeed - and the test of that is obviously to store the test file and do
whatever is necessary to make sure it is well and truely flushed out of the
cache (the power switch would be an effective method!) prior to timing the
read operation.
Kind Regards
John
----------------------------------------------------------------
Dr John Whittington, Voice: +44 (0) 1296 730225
Mediscience Services Fax: +44 (0) 1296 738893
Twyford Manor, Twyford, E-mail: John.W@mediscience.co.uk
Buckingham MK18 4EL, UK mediscience@compuserve.com
----------------------------------------------------------------
| 
