[ Contact Info ]

Welcome

The Paradyn project develops technology that aids tool and application developers in their pursuit of high-performance, scalable, parallel and distributed software. The primary project, Paradyn, leverages a technique called dynamic instrumentation to efficiently obtain performance profiles of unmodified executables. This dynamic binary instrumentation technology is independently available to researchers via the Dyninst API.

Other research by the Paradyn project includes dynamic instrumentation of running operating system kernels, the Kerninst project, and the development of middleware for scalable, efficient, robust applications in the MRNet multicast/reduction network.

 

News Items

The 15th Annual Scalable Tools Workshop is to be held at the Granlibakken Resort in Lake Tahoe, California, August 11 to August 15, 2024.

• February 2024: Dyninst 13.0.0 has been released.

• February 2023: Dyninst 12.3.0 has been released.

• July 2022: Dyninst 12.2.0 has been released.

• March 2022: Dyninst 12.1.0 has been released.

• November 2021: Dyninst 12.0.0 has been released.

• June 2021: Dyninst 11.0.1 has been released.

• April 2021: Dyninst 11.0.0 has been released.

• September 2020: Dyninst 10.2.1 has been released.

• July 2020: Dyninst 10.2.0 has been released.

Recent Papers

On identifying remedies to performance issues and automatically applying them to CPU-GPU applications "Identifying and (Automatically) Remedying PerformanceProblems in CPU/GPU Applications" by Welton and Miller, Under Submission

On detecting and estimating benefit of fixing performance issues in CPU-GPU applications "Diogenes: Looking For An Honest CPU/GPU Performance Measurement Tool" by Welton and Miller, Supercomputing 2019 (SC19). Denver, CO, November 2019.

On generating adversarial binaries for misleading authorship identification: "Adversarial Binaries for Authorship Identification" by Meng and Miller, Under Submission.

On the impacts of compilation toolchains on identifying multiple authors in binary programs: "Binary Code Multi-Author Identification in Multi-Toolchain Scenarios" by Meng and Miller, Under Submission.

[ Contact Info ]