[erlang-questions] Erlang, dirfy nifs, scheduler threads

Max Lapshin <>
Wed Mar 18 15:34:18 CET 2015


Hi.

Some time ago I've put ffmpeg inside erlang and made transcoding inside
erlang.

To make it working I had to start thread per each instance of
encoder/decoder and make my own thread pool, because R17 could get locked
after some time.

Now I want to update to R17.4 and use nif flags:

  {"decode0_nif", 4, avcodec_async_decode, ERL_NIF_DIRTY_JOB_CPU_BOUND},


what is the proper way to understand that erlang got stuck due to bad
calculation of nif timing?
Maybe I should write to logs some metric periodically?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20150318/dfa28a6f/attachment.html>


More information about the erlang-questions mailing list