Esteban Ribičić wrote:
> is it mounted sync or async?
>   
Reading through the man page of hdparm, I gather that the benchmark
performed using the -t flag will flush the IO-buffers and then read from
the device. And since hdparm talks directly with the device, not via the
file system, I think the result is probably quite accurate. (Not sure if
it is peak or average or if there is much difference on flash card)

The -T option will test reading from the IO-buffers, so it is really a
measurement of processor, cache and memory bus.

Hope I haven't missed any details.

Regards,
Pontus.
> On 12/18/07, Bent <[EMAIL PROTECTED]> wrote:
>   
>> For what it is worth, here is a measurement done on a 5501 running Gentoo 
>> Linux:
>>
>> leopold ~ # hdparm -tT /dev/hda
>>
>> /dev/hda:
>>  Timing cached reads:   390 MB in  2.01 seconds = 194.14 MB/sec
>>  Timing buffered disk reads:   34 MB in  3.14 seconds =  10.82 MB/sec
>> leopold ~ # hdparm -tT /dev/hdb
>>
>> /dev/hdb:
>>  Timing cached reads:   388 MB in  2.00 seconds = 194.03 MB/sec
>>  Timing buffered disk reads:   72 MB in  3.05 seconds =  23.60 MB/sec
>> leopold ~ #
>>
>> hda is a 8 GB SanDisk Ultra II. It is spec'ed to 10 MB/s on the package.
>> hdb is a 80 GB Seagate SATA disk.
>>
>> 73,  Bent
>> _______________________________________________
>> Soekris-tech mailing list
>> [email protected]
>> http://lists.soekris.com/mailman/listinfo/soekris-tech
>>
>>     
> _______________________________________________
> Soekris-tech mailing list
> [email protected]
> http://lists.soekris.com/mailman/listinfo/soekris-tech
>   
_______________________________________________
Soekris-tech mailing list
[email protected]
http://lists.soekris.com/mailman/listinfo/soekris-tech

Reply via email to