Initial values for online response time calculations

Abstract

1.1. HVE consumer products +9( FRQVXPHU SURGXFWV DUH KHDYLO\ FRVW FRQVWUDLQHG 7KHUHIRUH WKH DYDLODEOH UHVRXUFHV PXVW EH XVHG YHU\ FRVW HIIHFWLYHO\ ZKLOH SUHVHUYLQJ W\SLFDO TXDOLWLHV RI +9( FRQVXPHU WHUPLQDOV VXFK DV UREXVWQHVV DQG PHHWLQJ VWULQJHQW WLPLQJ UHTXLUHPHQWV LPSRVHG E\ KLJKTXDOLW\ GLJLWDO DXGLR DQG YLGHR SURFHVVLQJ 0RUHRYHU perception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any real-time systems needing an on-line schedulability test require exact schedulability analysis. In this paper we evaluate standard initial values for the iterative procedure to calculate worst-case response times of periodic tasks under fixed priority preemptive scheduling and arbitrary phasing. For discrete scheduling, we show that the number of iterations needed to determine the worst-case response time of a task using standard initial values increases logarithmically for an increasing worst-case computation time of that task. We present a new initial value, and prove that the number of iterations for that value is bounded. The costs of using the standard and new initial values are compared by means of an experiment. We briefly discuss the applicability of the initial value in other contexts, such as best-case response time analysis and jitter analysis.

Period-Dependent Initial Values for Exact Schedulability Test of Rate Monotonic Systems

“…We show in this paper that the task set is schedulable as long as some TDS point exists in the critical interval of i . We propose the initial value as max{p i -p i-1 , p i /2, c i /(1-(u 1 +u 2 + ... +u i-1 ))}.The performance of the proposed initial value has been tested and compared to earlier works [1,3].…”

“…Although exact tests provide a better schedulability testing for RM tasks, they require the task response time to be calculated iteratively [1,2,3,7,11]. In recent years, researchers have proposed several methods to improve the run time of RMA tests [2,3,11].…”

“…In recent years, researchers have proposed several methods to improve the run time of RMA tests [2,3,11]. Bini and Buttazzo [2] proposed a way to balance the required run time of the time demand analysis method and the false-identification rate of schedulable tasks.…”

“…Bini and Buttazzo [2] proposed a way to balance the required run time of the time demand analysis method and the false-identification rate of schedulable tasks. Bril et al [3] proposed a new initial value WR (0) = max{WR i-1 +c i , c i /(1-(u 1 +u 2 + ... +u i-1 ))} used by RTA. More recently, Lu et al [11] proposed an RTA iterative formula to derive WR (l+1) for l 0.…”

“…This paper studies the initial values used in RTA and proposes new initial values with a much better performance. Previous works [1,3] try to find the first TDS point in the critical interval of i (see Def. 4).…”