S. Richardson and M. Ganapathi, `Interprocedural optimization: experimental results', Software--- Practice and Experience, 19, 149--169 (1989).  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... of procedure calls [TIF86] LY88b] LY88a] HK91] resulting in significant execution time performance improvements [MS91] For scalar compilation, research on the effectiveness of interprocedural information has produced mixed results, ranging from moderate [Con83] to marginal improvement [RG89b] This dissertation does not address the issue of making interprocedural analysis effective. The focus of this research is to determine when interprocedural transformations such as inline substitution and cloning can be more effective than global optimization based on interprocedural ....

....with a combination of inlining and interprocedural information, based on empirical results on program characteristics for the PQCC multi language compiler backend [Con83] Richardson and Ganapathi observed an average of 1. 5 percent using only mod and ref information to optimize Pascal programs [RG89b] Richardson and Ganapathi s results may not reflect what can be expected from scientific fortran code. Since Pascal allows the programmer to declare a parameter as call by value or call by reference, the programmer can convey a certain amount of the interprocedural information to the compiler ....

S. Richardson and M. Ganapathi. Interprocedural optimization: Experimental results. Software---Practice and Experience, 19(2):149--169, February 1989.

....are used or modified by functions, or what variables are aliased to other variables. In the absence of this information, optimizers must make worstcase assumptions at function calls, assuming all global or possibly aliased variables are used or changed by the functions. Richardson and Ganapathi [43] found that for Pascal programs this information did not improve optimization. This kind of information is not useful for most functional languages languages, excluding Scheme. In most functional languages, variables cannot be altered once they are bound. Thus, compilers are guaranteed that ....

Stephen Richardson and Mahadevan Ganapathi. Interprocedural optimization: experimental results. Software --- Practice and Experience, 19(2):149--168, February 1989.

.... compilers, optimization Additional Keywords and Phrases: inline substitution, interprocedural analysis, interprocedural optimization 1 Introduction In recent years, many articles dealing with issues of interprocedural analysis and interprocedural optimization have appeared in the literature [1, 2, 5, 6, 7, 10, 12, 14, 17, 18, 19, 20, 21]. Several of these articles have attempted to assess the practical value of interprocedural data flow information or of specific cross procedural transformations. Ganapathi and Richardson point out that inline substitution can be viewed as an upper limit on the improvement available through use of ....

S. Richardson and M. Ganapathi. Interprocedural optimization: Experimental results. Software--- Practice and Experience, 19(2):149--169, Feb. 1989.

....example, Ball [Bal82] spent considerable effort just calculating the degree inlining procedures with constant parameters enables additional optimization. However, for the MIPS X architecture [HCe87] the removal of call and return code tends to dominate the other factors. Richardson and Ganapathi [RG89b] along with Davidson and Holler [DH88] found similar results for other architectures. Thus, a simple estimate can be used based on the typical number of instructions eliminated for each removed call. More complex estimates would produce better results, but this simple method is sufficient to ....

S. Richardson and M. Ganapathi. Interprocedural optimization: Experimental results. Software - Practice and Experience, 19(2):149--169, February 1989.

....dependence of compile time optimizations on the somewhat arbitrary structure of program code. In particular, procedure boundaries inhibit the effectiveness of many optimizations. While studies have shown that there are significant benefits to be gained from optimizing across procedure boundaries [23,24], finding and exploiting interprocedural opportunities can be quite challenging. Aggressive function inlining can remove many procedure boundaries entirely, but comes at the cost of increased code size, which, among its other drawbacks, can greatly 2 increase cache misses. Consistently effective ....

S. Richardson and M. Ganapathi, "Interprocedural Optimization: Experimental Results," Software---Practice and Experience, Vol. 10, No. 2, Feb. 1989, p. 149-169.

....speedup was 1.381. There had been other studies dealing with other aspects of optimization. Arnold [Arno83] reports on the effectiveness of the CYBER 205 vectorizing compiler in producing either vector or scalar versions of a loop as a function of the number of iterations. Richarson and Ganapathi [Rich89] have shown that certain types of interprocedural data flow analysis provides only marginal improvement on most of the programs in their suite. Callahan, Dongarra, and Levine have collected a large suite of tests for vectorizing compilers and have evaluated a large number of compilers [Call88] ....

Richardson, S. and Ganapathi, M., "Interprocedural Optimization: Experimental Results", Software-Practice and Experience, Vol.19, No.2, February 1989, pp. 149-170.

....or the ability of the programmers. A performance study based on the HP Precision Architecture global optimizer [11] found that on the same programs used by Chow the average speedup was 1.381. There have been other studies dealing with other aspects of optimization. Richardson and Ganapathi [19] have shown that certain types of interprocedural data flow analysis provide only marginal improvement on most of the programs in their suite. Callahan, Dongarra, and Levine have collected a large suite of tests for vectorizing compilers and have evaluated a large number of compilers [4] Most ....

Richardson, S. and Ganapathi, M., "Interprocedural Optimization: Experimental Results", Software-Practice and Experience, Vol.19, No.2, February 1989, pp. 149-170.

....summarize array side effects. In general, summary analysis for loop parallelization is less precise than the analysis of inlined code. On the other hand, inlining can yield an increase in code size which may disastrously increase compile time and seriously inhibit separate compilation [CHT91, RG89] Furthermore, inlining may cause a loss of precision in dependence analysis due to the complexity of subscripts that result from array parameter reshapes. For example, when the dimension size of a formal array parameter is also passed as a parameter, translating references of the formal to the ....

S. Richardson and M. Ganapathi. Interprocedural optimization: Experimental results. Software---Practice and Experience, 19(2), February 1989.

No context found.

S. Richardson and M. Ganapathi, `Interprocedural optimization: experimental results', Software--- Practice and Experience, 19, 149--169 (1989).

No context found.

S. Richardson and M. Ganapathi. Interprocedural optimization: Experimental results. Software---Practice & Experience, 19(2):149--169, 1989.

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