Express Logic Innovation Reduces Real-Time System Overhead
2/20/2012 3:00:00 AM
Embedded World class details benefits of Preemption-Threshold Scheduling
SAN DIEGO--Express Logic, Inc., the worldwide leader in royalty-free, real-time operating systems (RTOS), is scheduled to present details of its Preemption-Threshold Scheduling (PTS) in a class at Embedded World 2012 in Nuremberg, Germany. Although preemption is how real-time systems guarantee responsiveness to time-critical events, Express Logic has devised a way to minimize the number of preemptions, while preserving the systems' schedulability. Academic research has shown that systems implementing PTS can gain a 20-30% increase in performance and a 30â35% reduction in stack size, while still meeting schedulability demands.
"Preemption-thresholdscheduling ensures the application can meet those real-time deadlines using the least number of context switches possible. Academic research has proven that such refinements boost system performance while reducing memorytwo features that may explain the enormous success the ThreadX RTOS enjoys."
Technical Session Details:
- Paper titleMeet Hard Real-Time Deadlines with Reduced Overhead Through Preemption-Threshold Scheduling (PTS)
- PresenterJohn Carbone, VP of Marketing
- TimeWednesday, February 29th from 11:00â11:45 A.M. (session 11)
- Description: Real-time embedded systems typically use a collectionof independent tasks or threads that must complete their work before a certain deadline. Hard real-time systems demand that all deadlines be met with absolute certainty, even under worst-case conditions. To do this, real-time systems use preemptive scheduling which guarantees that the most critical threads get immediate attention, enabling them to meet their deadlines. However, preemptive scheduling can result in significant context-switch overhead. Preemption-Threshold Scheduling (PTS) reduces the number of context switches in a real-time system by inhibiting preemption for a thread-specific range of priorities. Developers can elect where and when to use PTS, and otherwise use traditional, fully preemptive scheduling. Research findings on PTS will demonstrate how PTS is optimal among all preemption-limiting methods for boosting application performance, minimizing a system's total stack memory requirements and avoiding priority inversion.