1 |
On 09/06/2009 01:48 AM, Volker Armin Hemmann wrote: |
2 |
> On Samstag 05 September 2009, Nikos Chantziaras wrote: |
3 |
>> On 09/05/2009 05:59 PM, Volker Armin Hemmann wrote: |
4 |
>>>> 1000 Hz timer freq |
5 |
>>> |
6 |
>>> change that to 300 |
7 |
>>> |
8 |
>>>> Did you mean tickless system with "noticks"? I have this enabled ATM |
9 |
>>> |
10 |
>>> deactivate that. |
11 |
>> |
12 |
>> Even with that, there still are problems. With composite enabled, move |
13 |
>> an mplayer window around and see how the video starts to skip big |
14 |
>> amounts of frames at the moment you start moving and when you "drop" it |
15 |
>> again. The compositor takes away CPU time and mplayer starves for a |
16 |
>> short time. BFS solves this. |
17 |
> |
18 |
> nope. No drops. vlc, xine, mplayer. At least no visible drop - and none of the |
19 |
> three is complaining. |
20 |
|
21 |
Well, not here. |
22 |
|
23 |
|
24 |
> But there is a pro-tip: use a non-broken X. aka one with |
25 |
> |
26 |
> fedora_dont_backfill_bg_none.patch |
27 |
|
28 |
I do use that. |
29 |
|
30 |
|
31 |
>> Also, have you considered that you got it all backwards? The kernel |
32 |
>> configuration tells you that for lower latencies, you should use 1000Hz |
33 |
>> and PREEMPT. It even says "Desktop" right there. Why should I take |
34 |
>> your word over that of the kernel devs who actually wrote that code? |
35 |
> |
36 |
> low latency means bad throughput and that hurts IO. |
37 |
|
38 |
GUI stalls still happen with high latency settings. Doesn't seem to |
39 |
matter; 300Hz, 1000Hz, tickless or not, PREEMPT or not, multicore |
40 |
scheduler or not, all the same. GUI stalls during load. It only goes |
41 |
away with Con's scheduler. |