COpt: A High Level Domain-Specific Language to Generate Compiler Optimizers

Venkat, Sruthi; Kanwal, Preet

doi:10.1109/icacat.2018.8933593

Cited by 2 publications

(1 citation statement)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, CAnDL requires developers to write constraints to specify optimizations, which is more complicated than high-level expressions JOG takes; also the generated compiler pass is independent to existing code structure in LLVM and thus more difficult to be integrated. COpt [47] is a high-level DSL that allows compiler developers to specify a set of ten high-level optimizations. In contrast to JOG that applies to peephole optimizations, COpt applies on high-level optimizations such as global value numbering, common subexpression elimination, function inlining, etc.…”

Section: Dsls For Optimizationsmentioning

confidence: 99%

Pattern-Based Peephole Optimizations with Java JIT Tests

Zang

Thimmaiah

Gligoric

2023

Proceedings of the 32nd ACM SIGSOFT International Symposium on Software Testing and Analysis

View full text Add to dashboard Cite

We present JOG, a framework that facilitates developing Java JIT peephole optimizations alongside JIT tests. JOG enables developers to write a pattern, in Java itself, that specifies desired code transformations by writing code before and after the optimization, as well as any necessary preconditions. Such patterns can be written in the same way that tests of the optimization are already written in OpenJDK. JOG translates each pattern into C/C++ code that can be integrated as a JIT optimization pass. JOG also generates Java tests for optimizations from patterns. Furthermore, JOG can automatically detect possible shadow relation between a pair of optimizations where the effect of the shadowed optimization is overridden by another. Our evaluation shows that JOG makes it easier to write readable JIT optimizations alongside tests without decreasing the effectiveness of JIT optimizations. We wrote 162 patterns, including 68 existing optimizations in OpenJDK, 92 new optimizations adapted from LLVM, and two new optimizations that we proposed. We opened eight pull requests (PRs) for OpenJDK, including six for new optimizations, one on removing shadowed optimizations, and one for newly generated JIT tests; seven PRs have already been integrated into the master branch of OpenJDK. CCS CONCEPTS• Software and its engineering → Just-in-time compilers; Domain specific languages; Software testing and debugging; Formal software verification; Source code generation.

show abstract

Section: Dsls For Optimizationsmentioning

confidence: 99%