Loading…
Loading grant details…
| Funder | National Science Foundation (US) |
|---|---|
| Recipient Organization | Massachusetts Institute of Technology |
| Country | United States |
| Start Date | Aug 01, 2024 |
| End Date | Jul 31, 2026 |
| Duration | 729 days |
| Number of Grantees | 3 |
| Roles | Principal Investigator; Co-Principal Investigator |
| Data Source | National Science Foundation (US) |
| Grant ID | 2346223 |
This POSE Phase II project will establish an open-source ecosystem (OSE) for OpenCilk, called Fastcode. OpenCilk is a state-of-the-art, completely open, task-parallel programming platform. Fastcode will encompass research and education in software performance engineering: making software run fast or otherwise consume few resources, such as time, storage, energy, network bandwidth, etc.
The FastCode OSE for OpenCilk and the academic-first strategy will lead to a significant increase in computer science and engineering graduates with experience in software performance engineering (SPE). Expanding the SPE workforce will, in turn, lead to significant efficiency gains in academic research, industry, and government across many domains. Fastcode will provide conduits for research in parallel programming and software performance engineering to find its way to application developers, impacting diverse domains of society, including business, science, technology, etc.
By providing software infrastructure to support novel research and teaching in software performance engineering, Fastcode will enable the much larger community of applications researchers to leverage the performance of modern computers. The benefits of computing, including faster and better scientific discoveries, will be magnified, and the environmental impact of computing will be mitigated due to reduced energy consumption (if you can compute something in half the time with the same resources, it saves about half the energy).
For application developers to cope with the end of Moore's Law, they must embrace software performance engineering (SPE) and all its constituent technologies: parallel programming, vectorization, caching, algorithms, compiler optimization, etc. The OpenCilk task-parallel platform simplifies parallel programming, arguably the most difficult of these technologies and the one with the greatest potential.
But without a more general knowledge of SPE, programmers cannot effectively exploit the full capabilities of modern multicore computers. The Fastcode OSE for OpenCilk will enable researchers to advance their understanding of SPE and parallel programming, providing the next generation of researchers and software developers with principled and scientific foundations for obtaining application performance in the post-Moore era.
This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
Massachusetts Institute of Technology
Complete our application form to express your interest and we'll guide you through the process.
Apply for This Grant