Sciunit is a suite of tools for creating lightweight containers from reference executions of an application. It supports command-line, notebook, and data-intensive applications. Each tool enables efficient containerization of multiple executions through content-based deduplication. The resulting containers can be re-executed in compatible x86-based Linux kernel environments, with support for modifying input parameters or datasets to reproduce results. Sciunit simplifies sharing self-contained applications among collaborators and leverages audited provenance to ensure repeatability and reproducibility.

Funding: ICER, ICER, ICER, CISE-CSR, NASA

ProvScope is an accurate and efficient tool for comparing extremely large execution provenance traces generated during reference runs of long-running applications. It uses an LLVM-based specification of the application to identify precise differences between traces, reporting the exact lines and functions where executions diverge. For MPI-based applications, ProvScope-MPI enables detection of differences across runs that exhibit both communication and input non-determinism. ProvScope reports all divergences and convergences.

Funding: CISE-CSR

Floability enables rapid and portable deployment of notebooks expressing complex scientific workflows across a wide range of cyberinfrastructure. It addresses the challenge of incomplete workflows by capturing software dependencies, required datasets, and cluster hardware capabilities. It also targets translation between notebooks and workflows. Floability is collaboratively developed by the University of Notre Dame, the University of Missouri-Columbia, and the University of Illinois.

Funding: NSF

FlexiFlow is a system designed to improve both performance and accuracy of end-to-end machine learning workflows during inference time. The project is under active development and will be made open-source.

Funding: MU Seed Funding