Sensitivity Analysis (SA) is a novel compiler technique that complements, and integrates with, static automatic parallelization analysis for the cases when relevant program behavior is input sensitive. In this paper we show how SA can extract all the input dependent, statically unavailable, conditions for which loops can be dynamically parallelized. SA generates a sequence of sufficient conditions which, when evaluated dynamically in order of their complexity, can each validate the dynamic parallel execution of the corresponding loop. For example, SA can first attempt to validate parallelization by checking simple conditions related to loop bounds. If such simple conditions cannot be met, then validating dynamic parallelization may require evaluating conditions related to the entire memory reference trace of a loop, thus decreasing the benefits of parallel execution.

Program analysis and program optimization of Java programs require reference information that estimates the instances of classes that may be accessed through dereferences. Recent work has presented several approaches for adapting Andersen's algorithm [1]---the most precise flow-insensitive and context-insensitive points-to analysis algorithm developed for C--- for analyzing Java programs (e.g., [5, 9, 12]). Studies in our previous work [6] indicate that this algorithm may compute very imprecise reference information for Java programs.

Program analyses and optimizations of Java programs require reference information that determines the instances that may be accessed through dereferences. Reference information can be computed using reference analysis. This paper presents a set of studies that evaluate the precision of two existing approaches for identifying instances and one approach for computing reference information in a reference analysis. The studies use dynamic reference information collected during run-time as a lower bound approximation to the precise reference information. The studies measure the precision of an existing approach by comparing the information computed using the approach with the lower bound approximation. The paper also presents case studies that attempt to identify the cases under which an existing approach is not effective. The presented studies provide information that may guide the usage of existing reference-analysis techniques and the development of new reference analysis techniques.

Regression testing is applied to modified software to provide confidence that the changed parts behave as intended and that the unchanged parts have not been adversely affected by the modifications. To reduce the cost of regression testing, test cases are selected from the test suite that was used to test the original version of the software---this process is called regression test selection. A safe regressiontest -selection algorithm selects every test case in the test suite that may reveal a fault in the modified software. Safe regression-test-selection techniques can help to reduce the time required to perform regression testing because they select only a portion of the test suite for use in the testing but guarantee that the faults revealed by this subset will be the same as those revealed by running the entire test suite. This paper presents the first safe regression-test-selection technique that, based on the use of a suitable representation, handles the features of the Java language. Unlike other safe regression test selection techniques, the presented technique also handles incomplete programs. The technique can thus be safely applied in the (very common) case of Java software that uses external libraries or components

This paper presents extensions to Steensgaard's and Andersen's algorithms to handle Java features. Without careful consideration, the handling of these features may affect the correctness, precision, and efficiency of these algorithms. The paper also presents the results of empirical studies. These studies compare the precision and efficiency of these two algorithms and evaluate the effectiveness of handling Java features using alternative approaches. The studies also evaluate the impact of the points-to information provided by these two algorithms on client analyses that use the information.