MUD-Dev
mailing list archive

[ Other Periods  | Other mailing lists  | Search  ]

Date:  [ Previous  | Next  ]      Thread:  [ Previous  | Next  ]      Index:  [ Author  | Date  | Thread  ]

[MUD-Dev] Re: lurker emerges



In message:  <35CE7173.7DEEBAEA#freehold,crocodile.org>
             Vadim Tkachenko <vt#freehold,crocodile.org> writes:

[ Re: double-buffering and non-blocking I/O ]

>I could be wrong, but I wish it rest in piece - it was a performance
>tuning nightmare. Either your I/O theread hogs all the resources looping
>idle, waiting for the input, or it lags because of the wrong priority
>set or bad timing.
>
>For me, the blocking I/O and a possibility to interrupt it (exceptions)
>is a blessing.
>
>Comments, anyone?

Well, my comment would be that you've apparently only seen very
bad implementations of a perfectly reasonable technique.  When
the I/O thread has no more work to do, it should go to sleep,
to be awakened by the next I/O completion.  With proper signal
support, no busywait is needed, and bad timing is irrelevant
if you're acually getting the device notifications.

- Alex




[ Other Periods  | Other mailing lists  | Search  ]