I want to stream the line input to other devices using icecast. Both ices and darkice work to do this, but top reports ~50% CPU on the pi zero wireless. I installed google profiler and built the ices software on the pi, and it reports that most of the execution time is taken by the GIselect() and __GI___poll() functions called by the alsa_read() function. These calls seem to be wrappers to syscall requests. I would rather not deploy a system that consumes CPU and energy like this, if top can be believed.
It seems that the select and poll functions are not working, perhaps because of an invalid handle. I am new to linux development, and my head is 'spinning' at the number of layers potentially involved. I don't know where to go. Could this be a problem for the alsa project? Or a problem with the kernel module?
If you know anything about this issue, I would appreciate some suggestions. Thanks
I want to stream the line input to other devices using icecast. Both ices and darkice work to do this, but top reports ~50% CPU on the pi zero wireless. I installed google profiler and built the ices software on the pi, and it reports that most of the execution time is taken by the GIselect() and __GI___poll() functions called by the alsa_read() function. These calls seem to be wrappers to syscall requests. I would rather not deploy a system that consumes CPU and energy like this, if top can be believed.
It seems that the select and poll functions are not working, perhaps because of an invalid handle. I am new to linux development, and my head is 'spinning' at the number of layers potentially involved. I don't know where to go. Could this be a problem for the alsa project? Or a problem with the kernel module?
If you know anything about this issue, I would appreciate some suggestions. Thanks