Article View: comp.sys.sequent
Article #475m_set_procs wierdness?
From: Tim Russell
Date: Mon, 18 Feb 1991 03:45
Date: Mon, 18 Feb 1991 03:45
25 lines
1376 bytes
1376 bytes
I'm working on a program for my parallel computing class that uses (of
course) the standard parallel library. However, I've come across an in-
consistency between the man page for m_set_procs() and the way it actually
functions.
The manual says that the argument, nprocs, declares the total number of
processes that will run in parallel, including the main. Fine. But then
it states that if nprocs is zero, the program creates no child processes
but acts as if it were going to.
Is this a typo? Whenever I use zero as an argument, m_numprocs gets
set to 4, which is the default of the number of processors / 2. Using 1
causes one process to be used.
Also, can someone tell me why a program can only use (number of processors
- 1) processes? Why can't I make maximum use of the system?
If it makes any difference, this is on a Balance 8000 with 8 processors
running Dynix 3.0.14. Thanks for any help!
--
____
\TR/ Tim Russell
\/ russell@ursa-major.spdcc.com
Technical Details
Message-ID:
<6518@spdcc.SPDCC.COM>Path:
arc1-ca.pugleaf.net!utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!think.com!paperboy!hsdndev!spdcc!russellShow Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!think.com!paperboy!hsdndev!spdcc!russell From: russell@spdcc.COM (Tim Russell) Newsgroups: comp.sys.sequent Subject: m_set_procs wierdness? Message-ID: <6518@spdcc.SPDCC.COM> Date: 18 Feb 91 03:45:55 GMT Organization: S.P. Dyer Computer Consulting, Cambridge MA Lines: 25