[Home] [Groups] - Message: [Prev in Group] [Next in Group]

nu.kanga.list.mud-dev

6798: [MUD-Dev] Re: lurker emerges

[Full Header] [Plain Text]
From: "T. Alexander Popiel" <popiel@snugharbor.com>
Newsgroups: nu.kanga.list.mud-dev
Date: Sun, 09 Aug 1998 22:05:41 -0600
References: [1]
Organization: Kanga.Nu
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