Scientific software is a critical component of the scientific discovery process. The development and use of scientific software is a complex and challenging endeavor that requires a diverse set of skills and expertise. The PESO project is a community-driven effort to promote the development and use of scientific software as an ecosystem with stakeholders and partners who engage with the US Department of Energy (DOE) in developing and using advanced high-performance computing.
Software ecosystems offer economies of scales in a variety of ways. They can provide a platform for sharing and reusing software components, tools, and libraries. They can also provide a platform for collaboration and coordination among software developers, users, and stakeholders. By promoting the development and use of scientific software as an ecosystem, PESO aims to create a sustainable and vibrant community of software developers, users, and stakeholders who work together to advance the state of the art in scientific computing.
The following links describes the promise of how investing in scientific software as an ecosystem can lead to a more productive and sustainable scientific discovery process for your organization:
| Topic | Description | Resources |
|---|---|---|
| Spack | Spack: What and Why | Spack |
| E4S | E4S: What and Why | E4S |
| HPSF | HPSF: What and Why | HPSF |
Powers of ten
A summary of the benefits of the PESO-sponsored software ecosystem, at every scale:
100,000+ — Lines of code replaced with high-quality libs & tools
Using robust open-source libraries and tools helps you eliminate native source code that is less capable, more fragile, and harder to maintain.
10,000+ — Community members via ecosystem collaborations
Shared experiences using the same ecosystem increases community knowledge base, improved software capabilities, and cross-teaming.
1,000+ — Code teams share ecosystem costs & benefits
Pooling investments and making the software ecosystem available everywhere optimizes cost and benefit sharing across many users.
100+ — Speedup using advanced devices like GPUs
ECP-sponsored applications realize a factor of 100 or more science-speedup by reformulating algorithms and software to exploit GPUs.
10+ — Reduction in build times via Spack build caches
Spack build caches reduce re-build times by a factor of 10 or more, greatly reducing staff wait times and accelerating debugging and reconfiguration activities.
1 — Source code base for all computing systems
Your single source code base that achieves performance portability, now and in the future, using libraries, tools, and compilers available via Spack and E4S.