Re: [linux-audio-dev] audio-disk thread interaction Was: Re: discussion about development overlap

[Paul Barton-Davis]

Benno writes:

>I thought the same a few months ago, but after running benchmarks, I
>discovered that one single thread with reasonable read write sizes
>(256KB), is at part or faster than a multithreaded approach when
>reading/writi ng files.

The person who should know the most about this is probably Stephen
Tweedie. When I asked him about this issue, he said that it was really
hard to tell, and that only experiments with a specific disk load
would resolve the issue. 

It will be interesting to see how much O_DIRECT behaviour may have on
this issue - its not entirely clear how much of the performance
difference is related to disk scheduling and how much is due to VM
issues with the buffer cache. Alas, O_DIRECT won't be here till 2.5.

--p