In this paper, we focus on real-time systems programming. Designs
of such systems often consist of several hundreds of highlevel
functionalities (or computational nodes) with timing constraints.
For example, the number of nodes ranges from 500 to 1000 in the
flight control system of an aircraft or of a space vehicle [9, 17].
When implementing such systems, real-time programmers can not
directly implement each of the nodes as a thread (or task) because
real-time operating system usually do not support such a high number
of threads. This limitation stems from the fact that having a
huge number of tasks in a system induces important overheads,
such as time overhead due to context switches [33, 21] and a bigger
memory footprint (e.g. task control block, size of the stack,
etc.) Thus, to cope with this limitation, real-time developers have
to group several nodes together into the same thread. This work is
generally performed manually and may be tedious and error prone
regarding the number of nodes involved. We are concerned here
with the automation of this process.
In this paper, we focus on real-time systems programming. Designsof such systems often consist of several hundreds of highlevelfunctionalities (or computational nodes) with timing constraints.For example, the number of nodes ranges from 500 to 1000 in theflight control system of an aircraft or of a space vehicle [9, 17].When implementing such systems, real-time programmers can notdirectly implement each of the nodes as a thread (or task) becausereal-time operating system usually do not support such a high numberof threads. This limitation stems from the fact that having ahuge number of tasks in a system induces important overheads,such as time overhead due to context switches [33, 21] and a biggermemory footprint (e.g. task control block, size of the stack,etc.) Thus, to cope with this limitation, real-time developers haveto group several nodes together into the same thread. This work isgenerally performed manually and may be tedious and error proneregarding the number of nodes involved. We are concerned herewith the automation of this process.
การแปล กรุณารอสักครู่..

In this paper, we focus on real-time systems programming. Designs
of such systems often consist of several hundreds of highlevel
functionalities (or computational nodes) with timing constraints.
For example, the number of nodes ranges from 500 to 1000 in the
flight control system of an aircraft or of a space vehicle [9, 17].
When implementing such systems, real-time programmers can not
directly implement each of the nodes as a thread (or task) because
real-time operating system usually do not support such a high number
of threads. This limitation stems from the fact that having a
huge number of tasks in a system induces important overheads,
such as time overhead due to context switches [33, 21] and a bigger
memory footprint (e.g. task control block, size of the stack,
etc.) Thus, to cope with this limitation, real-time developers have
to group several nodes together into the same thread. This work is
generally performed manually and may be tedious and error prone
regarding the number of nodes involved. We are concerned here
with the automation of this process.
การแปล กรุณารอสักครู่..
