PPoSS

A Full-stack Approach to Declarative Analytics at Scale

Sidharth Kumar, PI.

Grant

  • Collaborative Research: PPoSS: Large: A Full-stack Approach to Declarative Analytics at Scale
  • UIC PI: Sidharth Kumar
  • Lead PI: Thomas Gilray. Co-PIs: Matthew Might, Sidharth Kumar
  • Sponsor: National Science Foundation, Principles and Practice of Scalable Systems (PPoSS) program, Division of Computing and Communication Foundations
  • NSF Award #2316157, with collaborating awards at Washington State University, Syracuse University, The Ohio State University, and the University of Texas at Austin
  • Start Date: August 1, 2023
  • End Date: July 31, 2028
  • Award Amount: $2.6M for the lead award, with $960K to UIC

Overview

The project investigates full-stack implementation methodologies for expressive programming systems that bridge the gap between human-level specification and high-performance implementation of complex reasoning tasks at scale. Declarative languages let a programmer provide high-level rules and declarations that define a sought-after solution, which the computer then materializes automatically. The project scales this vision of high-performance declarative reasoning both to structured, higher-order, and probabilistic formulations and to the next generation of supercomputers and cloud-based clusters, with impacts on application designers and programmers in precision medicine, stochastic modeling, software verification, graph analytics, and security.

Next-generation exascale systems pose key challenges: increased parallelism, heterogeneity, GPUs, deep memory hierarchies, and performance tuning across the full software stack. As hardware grows more complex and diverse, it becomes difficult to write maintainable and scalable applications by hand. Modern chain-forward reasoning systems are meanwhile being extended with structured, higher-order data, probabilistic semantics, lattice orderings, recursive aggregation, and first-order theories, each posing implementation challenges in a parallel setting. The investigators are developing a unified, tunable, full-stack foundation for highly expressive chain-forward programming to be deployed at scale, released as open-source tools, programming languages, and frameworks for users across disciplines.