Skip to content

Latest commit

 

History

History
35 lines (26 loc) · 1.48 KB

paper.md

File metadata and controls

35 lines (26 loc) · 1.48 KB

Wasp: Foundation Model Training System for the Edge

Paper companion. Canonical place for paper metadata, the abstract, and figure-reproduction notes. The README links here from its Citation section.

System

Wasp is a coordinator-driven runtime for exact synchronous foundation-model training over decentralized edge resources. It decomposes GEMM-dominated training into row/column sub-GEMM shards and elevates each shard to a first-class runtime object with explicit placement (selective hybrid tensor parallelism), backpressure-aware dispatch, shard caching, and shard-granular recovery, preserving cloud-style synchronous update semantics.

Authors

  • Leyang Xue (The University of Edinburgh)
  • Yufeng Xia (The University of Edinburgh)
  • Myungjin Lee (Cisco Research)
  • Mahesh K. Marina (The University of Edinburgh)

Built on Morphling

Large-scale results use the measurement-calibrated Morphling emulator (EdgeSys '26, https://github.com/netsys-edinburgh/morphling), consumed through the public morphling.api surface. Please cite Morphling when using the emulated-scale results.

Reproduction scope

This repository ships a reference implementation and a runnable OPT-125M emulated example (examples/opt125m_emulated/). Baseline reimplementations and full paper-figure reproduction are seeded from the original artifact but are not a gated artifact-evaluation pipeline. The placement solver defaults to the license-free heuristic backend; the exact Gurobi MIP backend is optional.