lkml.org 
[lkml]   [2001]   [Mar]   [26]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectSlow... pipe/socket on 2.4.1

Hello,
The attached program attempts to show what happens when a
server tries to talk to a forked child on a UP machine.

The response time using a UNIX socket is awful.

When the machine uses software compiled for a SMP machine, the
data-rate is about 13 bytes / microsecond. This is not too
bad, but less than I would expect.

When the same machine uses software compiled for a UP machine,
therefore not using the second CPU, the data-rate is about
0.4 bytes / microsecond.

Running this same program on a Sun shows about 20 bytes / microsecond.
This is more like what I would expect.

It seems as though the problem is that, even though poll is called
(which should give up the CPU), the child waiting for data is not
given the CPU for a long time (perhaps the next timer-tick??).

I have not tried to use a named pipe for communications. Maybe
this is faster?

If anybody has some suggestions, please answer by the end-of-the
day. My machine will be off-the-air until I convince (@&$^@$)!)
that Linux is not a security hazard. This may take several
days.

Cheers,
Dick Johnson

Penguin : Linux version 2.4.1 on an i686 machine (799.53 BogoMips).

"Memory is like gasoline. You use it up when you are running. Of
course you get it all back when you reboot..."; Actual explanation
obtained from the Micro$oft help desk.

[unhandled content-type:application/octet-stream]
\
 
 \ /
  Last update: 2005-03-22 13:17    [W:0.048 / U:0.052 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site