| 1 |
Hi Daniel, |
| 2 |
on Saturday, 2007-02-10 at 12:49:14, you wrote: |
| 3 |
> I will give short overview what i have tried so far. |
| 4 |
> |
| 5 |
> 1. Trying different I/O Scheduler ( cfq anticipatory and deadline) |
| 6 |
> 2. Enabling Low latency kernel and Preemptible kernel |
| 7 |
> 3. Setting 1000 HZ for timer frequency |
| 8 |
> 4. Tried the new kernel 2.6.19-gentoo-r6 and even the testing version |
| 9 |
> 2.6.20-gentoo with core 2 enabled in processor type |
| 10 |
|
| 11 |
Oh, so it is a multicore CPU---sorry if you mentioned it already, I |
| 12 |
had deleted the start of the thread already when I read Benno's advice. |
| 13 |
In that case, try 100 Hz scheduling period as well. I've had very bad |
| 14 |
experiences with I/O and 250 Hz or higher on a dual Xeon. My guess is |
| 15 |
that it was a cache effect and therefore shouldn't happen on the |
| 16 |
Core2Duo, but it might still be worth a try. |
| 17 |
|
| 18 |
> As i am using Xfce i installed the diskperf-plugin which monitors disk |
| 19 |
> I/O. The monitoring is divided in disk-read and disk-write. |
| 20 |
> I recognized that every time when reading stops writing starts. So is |
| 21 |
> this staggering of writing to disk normal as the programs have to read |
| 22 |
> data they want to write to disk? On my previous machine i didn't |
| 23 |
> recognize such a behaviour. |
| 24 |
|
| 25 |
So you're reading and writing from/to the same disk? I'd expect that |
| 26 |
behavior then, because the I/O scheduler tries to satisfy requests with |
| 27 |
as little thrashing as possible. So if there are enough write requests |
| 28 |
queued up it may keep the HD busy writing for a while before reading the |
| 29 |
next chunk from somewhere else. |
| 30 |
|
| 31 |
cheers! |
| 32 |
Matthias |
| 33 |
-- |
| 34 |
I prefer encrypted and signed messages. KeyID: FAC37665 |
| 35 |
Fingerprint: 8C16 3F0A A6FC DF0D 19B0 8DEF 48D9 1700 FAC3 7665 |
Archives