Ball, J. E. Predicting the effects of optimization on a procedure body. ACM SIGPLAN Notices, 14(8):214 220, 1979.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....too big and making the analysis impractically slow or memory intensive. 2.2. 2 Background on Data Flow Analysis Our interprocedural data flow analysis framework draws from previous work on single procedure, global data flow analysis frameworks [70] flow insensitive interprocedural analysis [8, 27], and interval analysis [3] which we summarize here. Comprehen sive descriptions of global and interprocedural data flow analysis can be found else where [80, 26] Included in this description are the program representations used by analysis. In this section, we also introduce notation that we ....

Ball, J. E. Predicting the effects of optimization on a procedure body. ACM SIGPLAN Notices, 14(8):214 220, 1979.

....by inlining are not given. However, other work suggests that improvements will arise from (1) propagating constant valued parameters through the body of the called procedure, 2) enabling code motion across the former call site, and (3) exposing more information to the register allocator [Bal79] RG89a] WZ85] Unfortunately, very little experimental evidence has been published to prove these assumptions. Prior to this study, only two others have analyzed through experimentation the impact of inlining on optimization. Richardson and Ganapathi s study on 5 Pascal programs demonstrated ....

....are not substantial. We were interested in determining which of these was the case. If secondary effects were masking improvements, then other interprocedural techniques might still be effective. Improved constant propagation has been suggested as the most important effect of inlining [Bal79] WZ89] To test this, we performed an experiment to isolate the effects of interprocedural constant propagation. We also studied improvements when constants information is further refined by cloning. The experiment, described in the next section, proved that interprocedural constants are ....

[Article contains additional citation context not shown here]

J. E. Ball. Predicting the effects of optimization on a procedure body. In Proceedings of the SIGPLAN 79 Symposium on Compiler Construction. ACM, August 1979.

....approach using static analysis to estimate the opportunity cost. For example, it can be the number of positions that differ between a pair of cloning vectors. This stategy can be improved by taking into account execution frequency estimates and weighting the effects of each piece of information [1]. Given a method to compute the opportunity cost of merging two cloning vectors, the compiler can adopt a relatively simple rationing scheme. Assume that we set a quota for the total number of cloning vectors allowed during compilation and a quota for each procedure. The overall quota should be ....

Ball, J.E. Predicting the effects of optimization on a procedure body. ACM SIGPLAN Notices, 14(8):214--220, 1979.

....by using static analysis to estimate the merging cost. For example, the merging cost can be the number of positions that differ between a pair of cloning vectors. This can be improved by taking into account execution frequency estimates and weighting the effects of each piece of information [1]. Given a method to compute the cost of merging two cloning vectors, the compiler can adopt a relatively simple rationing scheme. Assume that we set a quota for the total number of cloning vectors allowed during compilation and a quota for each procedure. The overall quota should be proportional ....

J. E. Ball. Predicting the effects of optimization on a procedure body. In Proceedings of the SIGPLAN 79 Symposium on Compiler Construction, SIGPLAN Notices 14(8), pages 214--220. ACM, August 1979.

....limiting code growth, and achieving termination in the presence of recursion. Their algorithm is not online, polyvariant, or context sensitive in the sense we have described. Ball describes an analysis that determines which parameters contribute to the value of the expressions in a procedure body [5]. When constant parameters are available at a call site he uses the parameter dependency information to guide inlining decisions with an estimate of code savings and performance gain that is based on predictions about the impact of subsequent optimizations. We suspect the speed of our algorithm ....

J. Eugene Ball. Predicting the effects of optimization on a procedure body. SIGPLAN Notices, 14(8):214--220, August 1979. Proceedings of the ACM SIGPLAN '79 Symposium on Compiler Construction.

....limiting code growth, and achieving termination in the presence of recursion. Their algorithm is not online, polyvariant, or context sensitive in the sense we have described. Ball describes an analysis that determines which parameters contribute to the value of the expressions in a procedure body [5]. When constant parameters are available at a call site he uses the parameter dependency information to guide inlining decisions with an estimate of code savings and performance gain that is based on predictions about the impact of subsequent optimizations. We suspect the speed of our algorithm ....

J. Eugene Ball. Predicting the effects of optimization on a procedure body. SIGPLAN Notices, 14(8):214--220, August 1979. Proceedings of the ACM SIGPLAN '79 Symposium on Compiler Construction.

No context found.

J. E. Ball, `Predicting the effects of optimization on a procedure body', Proceedings of the SIGPLAN 79 Symposium on Compiler Construction, SIGPLAN Notices, 14, (8), 214-220 (1979).

No context found.

Ball, J. Eugene. "Predicting the Effects of Optimization on a Procedure Body". Proc.Sigplan '79 Symp. on Compiler Constr., also Sigplan Not. 14,8 (Aug. 1979),214-220.

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