Asri-unix.570
net.unix-wizards
utcsrgv!utzoo!decvax!ucbvax!hpvax!sri-unix!mike@BRL
Mon Jan 18 21:09:31 1982
Sticky Programs.  /tmp in memory
Mike -
       Our system has lots of stuff sticky, and it DOES make a big
difference.  Especially because swapping is on a separate controller
from the D-space fetch.  Being sticky saves CPU time (1 DMA -vs- many,
no block copying) and I/O time (1 continuous operation -vs- many).
They payoffs are less substantial, but still noticable, if you swap
on the same controller as /usr is on.

       Putting /tmp out in memory is a win, but putting it out on
the UNIBUS as a real controller has certain advantages.  Most importantly,
the xfer from the /tmp memory to the buffer cache is DMA for a real
controller, but it is a software copy the other way.  Of course, there
is the interupt overhead to pay when emulating a real disk.  Either
way is a big win over doing nothing.
                                               Cheers,
                                                -Mike

-----------------------------------------------------------------
gopher://quux.org/ conversion by John Goerzen <[email protected]>
of http://communication.ucsd.edu/A-News/


This Usenet Oldnews Archive
article may be copied and distributed freely, provided:

1. There is no money collected for the text(s) of the articles.

2. The following notice remains appended to each copy:

The Usenet Oldnews Archive: Compilation Copyright (C) 1981, 1996
Bruce Jones, Henry Spencer, David Wiseman.

You are viewing proxied material from infinitelyremote.com. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.