Thu 26 Aug 2021 05:20 - 05:25 - Analysis—Static Analysis and Symbolic Execution Chair(s): Akond Rahman
We present LLSC, a prototype compiler for nondeterministic parallel symbolic execution of the LLVM intermediate representation (IR). Given an LLVM IR program, LLSC generates code preserving the symbolic execution semantics and orchestrating solver invocations. The generated code runs efficiently, since the code has eliminated the interpretation overhead and explores multiple paths in parallel. To the best of our knowledge, LLSC is the first compiler for fork-based symbolic execution semantics that can generate parallel execution code.
In this demonstration paper, we present the current development and preliminary evaluation of LLSC. The principle behind LLSC is to automatically specialize a symbolic interpreter via the 1st Futamura projection, a fundamental connection between interpreters and compilers. The symbolic interpreter is written in an expressive high-level language equipped with a multi-stage programming facility. We demonstrate the run time performance through a set of benchmark programs, showing that LLSC outperforms interpretation-based symbolic execution engines in significant ways.
Wed 25 AugDisplayed time zone: Athens change
17:00 - 18:00 | Analysis—Static Analysis and Symbolic ExecutionIdeas, Visions and Reflections / Research Papers / Demonstrations +12h Chair(s): Vaibhav Sharma Amazon Web Services | ||
17:00 10mPaper | IDE Support for Cloud-Based Static Analyses Research Papers Linghui Luo Paderborn University, Germany, Martin Schäf Amazon Web Services, Daniel J Sanchez Amazon Alexa, Eric Bodden University of Paderborn; Fraunhofer IEM DOI Pre-print | ||
17:10 10mPaper | A Bounded Symbolic-Size Model for Symbolic Execution Research Papers DOI Media Attached | ||
17:20 5mPaper | LLSC: A Parallel Symbolic Execution Compiler for LLVM IR Demonstrations Guannan Wei Purdue University, Shangyin Tan Purdue University, Oliver Bračevac Purdue University, Tiark Rompf Purdue University DOI Pre-print | ||
17:25 5mPaper | Learning Type Annotation: Is Big Data Enough? Ideas, Visions and Reflections Kevin Jesse University of California at Davis, Prem Devanbu University of California at Davis, Toufique Ahmed University of California at Davis DOI | ||
17:30 30mLive Q&A | Q&A (Analysis—Static Analysis and Symbolic Execution) Research Papers |
Thu 26 AugDisplayed time zone: Athens change
05:00 - 06:00 | Analysis—Static Analysis and Symbolic ExecutionIdeas, Visions and Reflections / Research Papers / Demonstrations Chair(s): Akond Rahman Tennessee Tech University | ||
05:00 10mPaper | IDE Support for Cloud-Based Static Analyses Research Papers Linghui Luo Paderborn University, Germany, Martin Schäf Amazon Web Services, Daniel J Sanchez Amazon Alexa, Eric Bodden University of Paderborn; Fraunhofer IEM DOI Pre-print | ||
05:10 10mPaper | A Bounded Symbolic-Size Model for Symbolic Execution Research Papers DOI Media Attached | ||
05:20 5mPaper | LLSC: A Parallel Symbolic Execution Compiler for LLVM IR Demonstrations Guannan Wei Purdue University, Shangyin Tan Purdue University, Oliver Bračevac Purdue University, Tiark Rompf Purdue University DOI Pre-print | ||
05:25 5mPaper | Learning Type Annotation: Is Big Data Enough? Ideas, Visions and Reflections Kevin Jesse University of California at Davis, Prem Devanbu University of California at Davis, Toufique Ahmed University of California at Davis DOI | ||
05:30 30mLive Q&A | Q&A (Analysis—Static Analysis and Symbolic Execution) Research Papers |