Project

General

Profile

Feature #3829

How is the descrambling threaded ?

Added by danny skjodt over 8 years ago. Updated over 7 years ago.

Status:
Rejected
Priority:
Normal
Assignee:
-
Category:
-
Target version:
-
Start date:
2016-05-27
Due date:
% Done:

0%

Estimated time:

Description

When i descramble 7 channels from the same transponder on a pi3 only 1 core is used to 100% the last 3 sit arround 10% ?

History

#1

Updated by Jaroslav Kysela over 8 years ago

The data from one mux are processed in this way:

- hw input thread - dst: buffer1
- input thread - src: buffer1 - [descrambling] - TS packet split to queues
- queue threads (DVR, streaming)

So yes, the 'input thread' is the bottleneck in this case. But if you like to descramble 7 channels from one mux, I would buy more appropriate hw. ARM CPUs are a bit slow for this.

#2

Updated by danny skjodt over 8 years ago

Yeah i thought so, didnt have any problems with 6 channels for a long time, when i put on the 7th things started to have lagging.

The arm cpu should be good for descr 6x4 channels if all cores were used. So I dont think its that slow again =)

So what I need todo is to split the input into spts'es before tvheadend recieve them that way each channels decode will have its own thread ?

#3

Updated by Jaroslav Kysela over 8 years ago

Do you also use optimized dvbcsa library for ARM? It might be much faster on ARM.

https://github.com/glenvt18/libdvbcsa

#4

Updated by danny skjodt over 8 years ago

Yes ofcause :)

#5

Updated by Jaroslav Kysela over 7 years ago

  • Status changed from New to Rejected

Also available in: Atom PDF