threads
- number of stages in the pipeline
stimuli
- the number of messages to send
cpus
- the number of cpus to switch between
g
- indicates that the example should use group scheduling
c
- indicates taht the example should use ccsm
main thread
- forks off the pipeline threads
- sends messages to the first stage of the pipeline
- not placed under group scheduling control
- read a message
- does some work
- send out the same message
- uses the group scheduling api
- placed under exclusive group scheduling control under the sequential sdf
thread priorities are set in ascending order
b. lowest priorities are better a. ie: thread1 = 1, thread2 = 2, etc.
- dsui
gap interval
- the time it takes for a message to travel from one stage of the pipeline to an adjacent stage
- histogram is generated
pipe interval
- the time it takes for a message to travel through the entire pipeline
- histogram is generated
- dski
- dski events are recorded using the task filter
- child added to a ccsm set created by dski via child_alias
- cpu time
- the amount of time that each thread spends on the cpu throughout the program
- measured using switch_to/switch_from
- group scheduling narrative
- sequential sdf evetns