R. Leupers, Code Optimization Techniques for Embedded Processors Methods, Algorithms, and Tools. New York: Kluwer, 2000.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....and is directly integrated in the target code (see low level access functions under Linux [95] In this chapter, the application of known compiler techniques in the special field of device driver programming is briefly discussed. As reference for compiler techniques, mainly [66] and additionally [54, 68, 81, 120, 32] have been used. 8.1.1 Techniques and Pre conditions State of the art compiler techniques are code hoisting, constant folding, control flow analysis, dead code elimination, and code re ordering, just to name a few. The main question is, which optimisations can safely be done in combination with ....

Rainer Leupers. Code Optimization Techniques for Embedded Processors. Kluwer Academic Publishers, 2000.

....aware code generator is described in section V. We demonstrate the effectiveness of our approach in section VI and conclude the paper with a summary. II Compiler framework An overview of the compilation process and our compiler framework is given in figure 1. First, a front end (here: Lance2 [4]) reads a given C source program and transforms it into a machine independent intermediate representation (IR) After running a set of standard optimizations, the Lance2 IR is mapped to our generic low level IR (GeLIR) 5] which serves Lance2to GeLIR Lance2 frontend Lance2IR Asm ....

Rainer Leupers. Code Optimization Techniques for Embedded Processors. Kluwer Academic Publishers, Boston u.a., 2000.

....performed if such a tail duplication will not exceed the code size constraint. Although it may be possible to find the real optimal solution (i.e. tail duplications with best IPC) with an exhaustive search algorithm, like what used in determining best function inlining under a code size limit [18], the complexity of such a search approach is further increased by the fact that one tail duplication may change the efficiency of other candidates and increase the number of the possible tail duplications. Figure 4: The algorithm for best tail duplication for global scheduling under code size ....

Rainer Leupers, "Code Optimization Techniques for Embedded Processors", Kluwer Academic Publishers, 2000.

.... has two different meanings: a) if there exists a limit on code size, the optimal solution is maximizing the IPC while satisfying the code size constraint (i.e. find the best average code size efficiency for a given code size) Although code size constraints are more common in embedded processors [18] than high performance EPIC processors, it is useful when we want to limit the whole or working program size (i.e. the part of the code with execution frequency larger than zero) below the level 1 I cache size. The solution to it can be represented using a curve showing the best possible IPC for ....

....performed if such a tail duplication will not exceed the code size constraint. Although it may be possible to find the real optimal solution (i.e. tail duplications with best IPC) with an exhaustive search algorithm, like what used in determining best function inlining under a code size limit [18], the complexity of such a search approach is further increased by the fact that one tail duplication may change the efficiency of other candidates and increase the number of the possible tail duplications. a) b) c) d) Parent tree Parent tree Parent tree1 Parent tree1 Parent tree2 ....

Rainer Leupers, "Code Optimization Techniques for Embedded Processors", Kluwer Academic Publishers, 2000.

....for source code analysis, generation of an intermediate representation (IR) and machineindependent optimizations, while the backend maps the machine independent IR into machine speci c assembly code. As a frontend, we use the LANCE compiler system developed at the University of Dortmund [28]. LANCE is a machine independent, optimizing ANSI C frontend. There is no support for automatic retargeting, but LANCE comprises a backend interface that allows for direct coupling between the generated three address code IR and the data ow tree format used by contemporary code generator ....

R. Leupers, \Code Optimization Techniques for Embedded Processors," Kluwer Academic Publishers, 2000. Software: http://LS12-www.cs.uni-dortmund.de/leupers.

No context found.

R. Leupers, Code Optimization Techniques for Embedded Processors Methods, Algorithms, and Tools. New York: Kluwer, 2000.

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