Using Inter-Procedural Side-Effect Information in JIT Optimizations

JPure: A Modular Purity System for Java

“…In [18], an interprocedural pointer analysis is used to infer side-effect information for use in the Jikes RVM. This enabled upto a 20% improvement in performance for a range of benchmarks.…”

“…Knowing which methods are pure in a program has a variety of applications, such as: specification languages [19,1,3,10], model checking [31], compiler optimisations [9,18,36], atomicity [13], query systems [21,34] and memoisation of function calls [15].…”

Processor virtualization and split compilation for heterogeneous multicore embedded systems

“…On the other hand, LLVM [31] is a widely used framework which demonstrated that lifelong program analysis and transformation can be made available to arbitrary software, and in a manner that is transparent to programmers. For deferred compilation to be effective, high-level information must be propagated while lowering the program representation: it can take the form of annotations in Java class file for register allocation [2,18,27], array bound checks removal [38] or side effect analysis [29,32]. Split compilation generalizes these approaches: it uses annotations and coding conventions in the intermediate language to coordinate the optimization process over the entire lifetime of the program.…”

Automated Generation of Context-Aware Tests

“…To identify capps corresponding to statements that depend on reading or writing context data objects, we utilize side-effect analysis; more specifically we have instantiated the approach given by Le et al [7] to just focus on the contextual data types. For our example, this results in all the statements that may be affected by ctxData, that is, 12, 14, 21, 27, and 29 (capps #1, 2, 3, 5, 6 respectively).…”