Practical SQE on a Large Multi-Disciplinary HPC Development Team
In this paper we will discuss several software engineering practices that have proven useful in a large multidisciplinary physics code development project at Lawrence Livermore National Laboratory. In the project discussed here, as with many large scale efforts in HPC scientific computing, we have had to balance the competing demands of being a stable ''production'' code that our user base can rely on with being a platform for research into new physics, models, and software architectures. Much of this has been learned through necessity and experience. Likewise, much of it has been learned through interactions with other similar projects and hearing of their successes, and tailoring their ideas to our own requirements. The ideas presented here are not meant to necessarily transfer to other environments with different needs. It is our belief that projects need to be given large latitude in defining their own software engineering process versus a prescribed a solution. However, the ideas presented here are hopefully high level and general enough that we hope other projects might find some inspiration and adopt similar methods if it is to their benefit, much as we have done through the years.
- Research Organization:
- Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
- Sponsoring Organization:
- USDOE
- DOE Contract Number:
- W-7405-ENG-48
- OSTI ID:
- 15014047
- Report Number(s):
- UCRL-CONF-203192; TRN: US200803%%1017
- Resource Relation:
- Conference: Presented at: International Workshop on Software Engineering for High Performance Computing System (HPCS) Applications, Edinborough, United Kingdom, May 23 - May 28, 2004
- Country of Publication:
- United States
- Language:
- English
Similar Records
Development of the Breakthrough Vorcat Technologyfor Cloud-Based Complex Energy Simulations
Ten million and one penguins, or, lessons learned from booting millions of virtual machines on HPC systems.