AI Assistants to Enhance and Exploit the PETSc Knowledge Base

Barry Smith, Junchao Zhang, Hong Zhang, Lois Curfman McInnes, Murat Keçeli, Archit Vasan, Satish Balay, Toby Isaac, Le Chen, Venkatram Vishwanath

ICPP Workshops '25: Workshop Proceedings of the 54th International Conference on Parallel Processing, 9–17 (2025), arXiv:2506.20608

Summary

Describes the PETSc team's LLM-powered system—RAG, reranking, and chatbots over source code, documentation, mailing lists, and issues—to make three decades of fragmented PETSc knowledge accessible to users and developers.

Keywords

PETSc knowledge base retrieval-augmented generation chatbots scientific software LLM assistants

Cite

@article{2025_smith_petsc_ai_assistants,
  title = {AI Assistants to Enhance and Exploit the PETSc Knowledge Base},
  author = {Barry Smith and Junchao Zhang and Hong Zhang and Lois Curfman McInnes and Murat Keçeli and Archit Vasan and Satish Balay and Toby Isaac and Le Chen and Venkatram Vishwanath},
  journal = {ICPP Workshops '25: Workshop Proceedings of the 54th International Conference on Parallel Processing, 9–17 (2025), arXiv:2506.20608},
  year = {2025},
  doi = {10.1145/3750720.3757281},
  eprint = {2506.20608},
  archivePrefix = {arXiv}
}