| 1 |
On Thu, Aug 11, 2011 at 14:27, Nikos Chantziaras <realnc@×××××.de> wrote: |
| 2 |
> On 08/11/2011 09:56 AM, Pandu Poluan wrote: |
| 3 |
>> |
| 4 |
>> Just wondering, is it possible to specify a custom timer frequency? |
| 5 |
>> E.g., HZ=500 instead of one of the canned values (100, 250, 300, |
| 6 |
>> 1000). |
| 7 |
> |
| 8 |
> It is possible, but it's a bad idea because non-standard values can result |
| 9 |
> in driver breakage. Some code assumes specific timer granularities (100Hz = |
| 10 |
> 10ms, 250Hz = 4ms, etc). This usually happens with values above 1000Hz, so |
| 11 |
> it might be possible to experiment with non-standard sub-1000Hz values. |
| 12 |
> |
| 13 |
> But why do you want a custom value anyway? |
| 14 |
> |
| 15 |
|
| 16 |
Well, for a firewall, I've calculated (gathered and extrapolated from |
| 17 |
a lot of sources), the latency per-packet is usually less than 1 ms, |
| 18 |
at worst still less than 2 ms. |
| 19 |
|
| 20 |
Thus, setting the timer freq to 100 Hz (as suggested for 'normal' |
| 21 |
server load) means the timeslice is way too long (10 ms per time |
| 22 |
slice). |
| 23 |
|
| 24 |
So, I speculate that better -- and uniform -- performance will be |
| 25 |
achieved with a timer freq of 500 Hz. |
| 26 |
|
| 27 |
Of course, this is a wild speculation/guess from me. I never quite |
| 28 |
understand netfilter/xtables' relation to the timeslices, so I may be |
| 29 |
talking nonsense :-) |
| 30 |
|
| 31 |
Rgds, |
| 32 |
-- |
| 33 |
FdS Pandu E Poluan |
| 34 |
~ IT Optimizer ~ |
| 35 |
|
| 36 |
• Blog : http://pepoluan.tumblr.com |
| 37 |
• Linked-In : http://id.linkedin.com/in/pepoluan |
Archives