Bryan.C.Breen.1 at gsfc.nasa.gov
Thu Nov 14 17:39:02 PST 2002
At 20:19 11/14/02 +0000, you wrote:
>I wrote a little program to read 1MB then print time in microseconds,
>then read another MB, etc. and found that the disk read was averaging
>about 2MB/sec for the most part, with an extended period of 0.8MB/sec,
>and picking up to a spotty average of 8-10MB/sec (with peaks of 20) for
>about the last 1/3 of the file.
>I know this is a non-ideal case, but 0.8MB/sec! really?! with an 8MB
>cache on the drive?
The big cache is handy for keeping a lot of recently accessed data ready to
be re-read without accessing the actual data on the drive's disks. To see
the affects of this, run your test on a small file, say one that's about a
meg or two in size, and then run the test again on the same small file. The
second (and any additional) times will be much faster than the first since
the data is then in the cache.
If you are reading in a very large file, it will completely fill the cache
(and more). So any additional reads of the same file will still have to
come off the disks again, thus negating the performance benefits of the
Don't expect drives that are identical, besides the size of their cache, to
perform any differently when reading large files.
As for your overall poor performance... ya' got me. :(
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe blfs-support' in the subject header of the message
More information about the blfs-support