Child process execution

New Message Reply Date view Thread view Subject view Author view

Richard M. Spencer (rspencer++at++rti.org)
Wed, 25 Jun 1997 18:44:57 -0400

My Vega app has two serial device readers that require several
milliseconds of app time. I therefore spawned these two read functions
with sproc and trigger them with semaphores. I hoped to take advantage
of the time apparently available just before/during vgSyncFrame/pfSync I
assumed was freeing about 10% of the CPU time on my single R10000 Max
Impact.

However, despite where I trigger the read semaphores relative to
vfSyncFrame/pfSync and vgFrame/pfFrame, my Performer statistics now show
a lower gray line after app and before draw corresponding to the
processing time for the readers. My overall frame time has increased
about 2.5 ms due to process and semaphore overhead.

I assume (and will confirm) this will work properly on a multiprocessor
machine (I'll eventually make it multithreaded) but what can I do to
force the drivers to run during the single CPU idle time and not hinder
the main simulation process app time? Any suggestions?

Thanks

-- 
Richard M. Spencer           Research Triangle Institute - CDSE
Research Computer Scientist  3040 Cornwallis Road, Herbert Bldg, Rm 226
rspencer++at++rti.org             P.O. Box 12194; RTP, NC 27709-2194  USA
http://www.rti.org/vr/       919-541-6733 (voice), 919-541-6515 (fax)
=======================================================================
List Archives, FAQ, FTP:  http://www.sgi.com/Technology/Performer/
            Submissions:  info-performer++at++sgi.com
        Admin. requests:  info-performer-request++at++sgi.com

New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:55:30 PDT

This message has been cleansed for anti-spam protection. Replace '++at++' in any mail addresses with the '@' symbol.