[mythtv-users] Kernel options, Preemption model and Timer frequency

[Hika van den Hoven]

Hoi Rajil,

Sunday, May 13, 2018, 12:32:57 AM, you wrote:

> On 05/12/2018 05:07 PM, Hika van den Hoven wrote:
>>
>> Is your user member of both the audio and the usb group?
> Yes, it is. The USB sound actually doesnt expose that prealloc setting.
> # tree /proc/asound/card1
> /proc/asound/card1
> ├── id
> ├── oss_mixer
> ├── pcm0c
> │   ├── info
> │   ├── oss
> │   └── sub0
> │       ├── hw_params
> │       ├── info
> │       ├── status
> │       └── sw_params
> ├── pcm0p
> │   ├── info
> │   ├── oss
> │   └── sub0
> │       ├── hw_params
> │       ├── info
> │       ├── status
> │       └── sw_params
> ├── stream0
> ├── usbbus
> ├── usbid
> └── usbmixer

> 4 directories, 18 files
>> For the backend I, as it is dedicated I would use default server
>> settings. Other choices are mostly aimed at a reactive desktop. So
>> No forced preemption and a low timer frequency.
> What about a dedicated combined BE/FE?
> _______________________________________________

It still mostly is a single purpose machine.

As said a lower timer frequency means that any running proces is less
interrupted, so runs slightly more efficient. A higher rate means
faster reactions to external interrupts, mostly from the user on the
desktop.

Forced preemption means that no matter what a process will be
interrupted to give time to other processes, which for a timing
sensitive process like recording can have a negative effect. However
it means that you as a user get a more reactive machine.

So I would choose intermediate timer frequency. Say 250/300 and keep
the "No forced preemption" in place. If you find the frontend to slow
in reaction, you can try "Voluntary preemption" or raise the
frequency.

Tot mails,
  Hika                            mailto:hikavdh at gmail.com

"Zonder hoop kun je niet leven
Zonder leven is er geen hoop
Het eeuwige dilemma
Zeker als je hoop moet vernietigen om te kunnen overleven!"

De lerende Mens