Linux offers no assurances whatsoever about timing. Try running some loads at the same time and see what happens.
On Sat, Feb 02, 2002 at 09:53:22PM -0200, Cassia Yuri Tatibana wrote: > > > Hello > > I've been running real time threads under Linux (kernel 2.4.0-test10), > using SCHED_FIFO policy in order to try to describe temporal behavior of > these tasks under Linux (not real time linux). The threads use the highest > priority allowed for the scheduller (5 threads, 5 different priorities) > and each of the five threads implemented use the same code but different > periods. > > Periods are determined by nanosleep call, which is called using a > different time amount parameter according to each thread's chosen period. > > My idea was to see how these tasks were treated by Linux, through the > analysis of the time interval caught in each instances of the tasks > executed (response time: instant of task's execution conclusion - > instant of task arrival). > > Well after adjusting the nanosleep time amount paramenter (in order to > get a more precise sleep period), the response times got were closer to > the expected. > > But sometimes, like coming from nowhere, one instance of the > experiment returns a time value much higher than the rest. > (ex: when most of the values are 67us, this particular instance returns > 20ms). And I just can't imagine what it could be!! All I know is that it > is caused by the nanosleep call, which, despite being called with a time > parameter very close to the rest of the instances, takes much more time in > sleep.. > It has happens once in each experiment (first task with 3000 instances / > period = 50ms), and sometimes it just don't happen... > Would anyone have a clue of what could it be? > > The code of my little program can be downloaded in > http:\\www.lcmi.ufsc.br\~cytatiba > the file is tr5.c > > Thank you so much! > Cassia Yuri Tatibana > <[EMAIL PROTECTED]> > > "Life is hard, > after all it kills you." > Katherine Hepburn > > > > > > -- [rtl] --- > To unsubscribe: > echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR > echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED] > -- > For more information on Real-Time Linux see: > http://www.rtlinux.org/ -- --------------------------------------------------------- Victor Yodaiken Finite State Machine Labs: The RTLinux Company. www.fsmlabs.com www.rtlinux.com -- [rtl] --- To unsubscribe: echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED] -- For more information on Real-Time Linux see: http://www.rtlinux.org/
