Gustav Pospischil, Peter Puschner, Alexander Vrchoticky, and Ralph Zainlinger. Developing real-time tasks with predictable timing. IEEE Software, pages 35--44, September 1992.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....knowledge, it has not been used in the context of WCET analysis. The idea of showing execution time characteristics in a graphical environment and enable users to change between different abstraction levels has been presented by Pospischil, Puschner, Vrchoticky and Zainlinger in the Mars project [PPVZ92]. The main difference to the approach proposed in this paper is our need to handle unstructured code, a necessity in the class of systems we target. This research was partly funded by Ericsson AB and the Swedish Knowledge Foundation. The prototype tool components are implemented for the PLEX ....

G. Pospischil, P. Puschner, A. Vrchoticky and R. Zainlinger. Developing Real-Time Tasks with Predictable Timing. IEEE Software, 9(5), 1992.

....nor has it yet modelled the timing behaviour of a real machine. Augmented high level language compilers are an obvious way of automating the discharge of timing obligations, relieving programmers from the burden of cycle counting . A number of compilers and tools that can predict timing behaviour [13] have been proposed, some of which rearrange code in order to improve performance [9] Indeed, performance enhancing heuristics for guiding real time software development are a valuable adjunct to the raw Quartz development rules. There is still a great deal of cre17 ativity involved in ....

G. Pospischil, P. Puschner, A. Vrchoticky, and R. Zainlinger. Developing real-time tasks with predictable timing. IEEE Software, 9(5):35--44, September 1992. 18

....which indicate those pipeline stages used by that instruction at each cycle. Sequential composition of these matrices allows them to overlap, as long as there is no conflict for access to a pipeline stage. This overlapping yields much less pessimistic timing estimates than previous approaches [27, 3, 23, 16, 24]. In particular, their predictions become better over longer instruction sequences, where more contextual information is available. A similar approach is then proposed to model access to cache blocks [21] Furthermore, tools are becoming available that allow the timing behaviour of RISC code to ....

....of observable outputs, a feature entirely compatible with the Topaz approach. The well established MARS model uses a fully integrated programming environment for construction of real time systems, including a programming language, development tools, compiler, timing analyser and hardware [24, 25]. The compiler creates a timing graph that acts as a common data structure used to communicate between tools. By relying on timing constraints being reduced to simple straight line paths, Topaz instead uses the annotated assembler code in this role. 6 Conclusion The Topaz methodology for ....

G. Pospischil, P. Puschner, A. Vrchoticky, and R. Zainlinger. Developing real-time tasks with predictable timing. IEEE Software, 9(5):35--44, September 1992.

....or parallel real time programs to determine one or more program points at which monitoring instrumentation can be placed so that it will be non intrusively absorbed by the application. Extensive research has been carried out in the area of real time languages and their analysis for scheduling [4, 14, 18]. Previous work on timing analysis has mainly concentrated on execution timing analysis. This includes compiler support for computing WET estimates [9, 10, 12] and the run time refinement of WET estimates based upon a combination of compile time information [15, 6] and run time information [3, 5] ....

G. Pospischil, P. Puschner, A. Vrchoticky, and R. Zainlinger, "Developing Real-Time Tasks with Predictable Timing," IEEE Software, special issue on Real-time systems, pp.35-44, September 1992.

....the timing behavior of programs and allow these proofs to be connected to a model of the hardware. The use of semantic information about the application could be applied in a sound manner rather than in an ad hoc fashion during development of timing routines. The Mars design environment system [40] allows the user to do timing calculations similar to that of TAL. That is, the system can calculate upper bounds of program execution time, mostly automatically from the syntax. The user can add timing assertions presumably that involve some higher level understanding of the program in ....

Gustav Pospischil, Peter Puschner, Alexander Vrchoticky, and Ralph Zainlinger. Developing real-time tasks with predictable timing. IEEE Software, pages 35--44, September 1992.

....of the i th job; C is the worst case contention free execution time of the job; T is the period of the task; oe and are nonnegative real constants; we call oe the maximum deadline laxity and the maximum interval laxity. There has been extensive research in the area of real time scheduling [2, 4, 5, 7, 8, 12, 13, 16, 17, 18, 21, 22]; also see related work section. For example, for independent preemptive tasks, the RM (rate monotonic) priority assignment and the EDF (earliest deadline first) have been shown to be optimal scheduling algorithms in the static and dynamic cases, respectively [12] However, most real time ....

G. Pospischil, et al., "Developing real-time tasks with predictable timing," IEEE Software, 1992, pp. 35-44.

....and the nature of the support provided by the system. The programming environment of the MARS project takes a slightly different approach, expressing all computations in terms of modules which receive messages as they begin execution, send messages as they complete, and execute periodically [67]. This has aspects of object orientation and data flow about it, although the language, Modula R, is an extension of Modula 2 and is not not explicitly object oriented. Their approach has the advantage of regularity and simplicity, is well coordinated with the properties of the supporting ....

....but not all computations can be described in monotonic form. The MARS system approaches the problem from a different direction, supporting process based programming and run time models with synchronous communication, but limited to communication at the beginning and end of each process [67]. MARS places the responsibility for decomposing the computation into a set of communicating processes on the developer. The programming model imposes other restrictions as well, including that every process be periodic, and that information exchange between processes is limited to messages. MARS ....

[Article contains additional citation context not shown here]

Pospischil, G., Puschner, P., Vrchoticky, A., and Zainlinger, R. Developing RealTime Tasks with Predictable Timing. IEEE Software, 9(5):35--44, September 1992.

....path is one that can be derived from the static program structure but can never be executed in practice. To obtain a tighter WCET bound, we must utilize additional information to eliminate infeasible paths. Many methods use additional information to tighten the WCET bound. Puschner and Koza [38, 39, 40] introduced several new language constructs with which programmers can describe the timing behavior of their programs. Their experiments showed that, with this valuable information, the gap between the estimated WCET bound and the real WCET can be reduced significantly. Gupta and Gopinath [7] ....

Gustav Pospischil, Peter Puschner, Alexander Vrchoticky, and Ralph Zainlinger. Developing real-time tasks with predictable timing. IEEE Software, pages 35--44, September 1992.

....5. Related Work Extensive research has been carried out in the area of real time languages (Kenny and Lin, 1991, Lee and Gehlot, 1985, Nirkhe, Tripathi, and Agrawala, 1990, Stoyenko, 1987) and their analysis for scheduling (Lin, Natarajan, and Liu, 1987, Lin and Natarajan, 1987, Mok et al. 1989, Pospischil, Puschner, Vrchoticky, and Zainlinger, 1992, Shaw, 1989, Stoyenko, 1987, Xu and Parnas, 1990) Previous work on timing analysis has mainly concentrated on execution timing analysis. This includes compiler support for computing WCET estimates (Neihaus, 1991, Harmon, Baker, and Whalley, 1992, Hong and Greber, 1993, Nirkhe and Pugh, 1992) and ....

Pospischil, G., P. Puschner, A. Vrchoticky, and R. Zainlinger, "Developing Real-Time Tasks with Predictable Timing," IEEE Software, September, 1992, pp. 35--44.

....descriptions of how to perform the phases. The tools do often lack support for functional and or timing verification (analysis) Even more rare is support for generating the designs (synthesis) given input requirements. Typical design methods are HRT HOOD (cf. 5, 6] and the MARS method (cf. [11, 14]) HRT HOOD requires a method to estimate (or determine) WCETs and a decision on scheduling strategy. HRT HOOD is more general than MARS, since MARS has the hardware architecture and programming environment defined together with the method. On the other hand it is easier to incorporate synthesis ....

Pospischil, G., Puschner, P., Vrchoticky, A., and Zainlinger, R. Developing Real-Time Tasks with Predictable Timing. IEEE Software, 35--44, September 1992.

No context found.

G. Pospischil, P. Puschner, A. Vrchoticky,and R. Zainlinger. Developing real-time tasks with predictable timing. IEEE Software, 9(5), Sep. 1992.

....The real time programmer has to develop software that meets the execution time constraints imposed by the earlier phases of system design. Therefore, detailed information about the timing behavior of a process is desirable during the implementation phase. The MARS programming environment [20, 18] is built such that it not only gives the programmer a very detailed documentation of a processes timing, but also allows an immediate observation of the effects of code changes on its overall execution time. These facilities are especially desirable when developing critical processes with tight ....

G. Pospischil, P. Puschner, A. Vrchoticky, and R. Zainlinger. Developing real-time tasks with predictable timing. IEEE Software, 9(5):35 -- 44, Sep. 1992.

....When the authors experimented with the tool they observed that the tool did not find the correct mapping between the two representations for all input programs. To avoid this problem, a new task timing analysis tool was developed [14] It is based on a single data structure, called timing tree [11], which contains all information needed to calculate MAXT bounds. This tool not only allows its users to compute worst case execution time bounds of high quality; it also produces detailed information about the contribution of every statement to this bound and allows programmers to experiment with ....

G. Pospischil, P. Puschner, A. Vrchoticky, and R. Zainlinger. Developing real-time tasks with predictable timing. IEEE Software, 9(5):35--44, Sep. 1992.

....have not yet addressed the problems posed by code improving transformations: They either require manual intervention by the programmer (i.e. specification of execution frequency bounds at the assembler level) or rule out such transformations altogether. We have designed and implemented a system [9] that calculates source level execution time bounds of programs written in a high level language [13] and presents the calculated bounds to the programmer as annotations to the source code state This work has been supported by the Austrian Science Foundation (Fonds zur Forderung der ....

....generates this representation. The impact of code optimizations is discussed in Section 4. Finally, Section 5 assesses our approach, identifies some of the problems, and gives directions for future research. 2 Timing Tree In addition to generating assembler code, our compiler creates timing trees [9, 14] for later perusal by the execution time analysis tool. A timing tree is a hierarchical representation of the block structure of a procedure, vaguely resembling a heavily pruned abstract syntax tree. Figure 1 shows the timing tree of a procedure implementing Euclid s algorithm for computing the ....

[Article contains additional citation context not shown here]

G. Pospischil, P. Puschner, A. Vrchoticky, and R. Zainlinger. Developing Real-Time Tasks with Predictable Timing. IEEE Software, 9(5):35--44, Sep. 1992.

No context found.

Gustav Pospischil, Peter Puschner, Alexander Vrchoticky, and Ralph Zainlinger. Developing real-time tasks with predictable timing. IEEE Software, pages 35--44, September 1992.

No context found.

G. Pospischil, et al., "Developing Real-Time Tasks with Predictable Timing," IEEE Software, September 1992, pp. 35-44.

No context found.

G. Pospischil, P. Puschner, A. Vrchoticky, and R. Zainlinger. Developing real-time tasks with predictable timing. IEEE Software, 9(5):35--44, September 1992.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC