| 1 |
Simon <turner25 <at> gmail.com> writes: |
| 2 |
|
| 3 |
> I am another newbie trying to answer a question, here on my eeepc |
| 4 |
> 701, hdparm gives even more limited information. This is a 2g |
| 5 |
> kingston compact flash. (Note that even though last line reports i/o |
| 6 |
> error, nothing was written in /var/log/messages about it). I suspect |
| 7 |
> device drivers to yield a block device asap but let optimizations and |
| 8 |
> other extras aside. |
| 9 |
|
| 10 |
|
| 11 |
Well I was looking for one of the wise (you7ng)old masters to answer. |
| 12 |
Since that did not happen, I just 'brute forced' a solution, that is |
| 13 |
optimal (hopefully) for my hardware. Here's what I did, so it may or |
| 14 |
may not help you. |
| 15 |
|
| 16 |
|
| 17 |
You can make adjustments to hdparm dynamically at the command line like so |
| 18 |
|
| 19 |
hdparm -c1 /dev/hda < your /dev/drive may differ so adjust accordingly> |
| 20 |
|
| 21 |
hdparm -tT /dev/hda < this perform a simple throughput test> |
| 22 |
|
| 23 |
hdparm -a 256 -m16 -c1 /dev/hda < make some adjustments and retest > |
| 24 |
|
| 25 |
hdparm -m4 -c1 /dev/hda < gave me the best results for my hardware> |
| 26 |
|
| 27 |
|
| 28 |
Thus my /etc/conf.d/hdparm file looks like this: |
| 29 |
|
| 30 |
# These setting for the 4 gig CF Sandisc card seem to be optimal |
| 31 |
hda_args="-m4 -c1 " |
| 32 |
|
| 33 |
|
| 34 |
|
| 35 |
If you really wanted to go all out testing, use bonnie or bonnie++, |
| 36 |
both are in portage. |
| 37 |
|
| 38 |
|
| 39 |
hth, |
| 40 |
James |
Archives