NESAP for Workflows Call for Proposals
The National Energy Research Scientific Computing (NERSC) Center is now accepting applications from NERSC projects for the NERSC Science Acceleration Program (NESAP) for Workflows. Chosen science teams will work with NERSC for one year to prepare for and better utilize advanced workflow capabilities, such as hardware acceleration, reconfigurable storage, advanced scheduling, and integration with edge services. These project goals are in alignment with the DOE's Integrated Research Infrastructure (IRI) initiative.
Accepted teams will be partnered with resources at NERSC and NERSC’s technology vendors, including access to:
- NERSC Workflow Readiness staff assistance with deploying code on new hardware, profiling, optimization, integrating with workflow managers, using cloud-native technologies where possible (containers, k3s orchestration, etc.)
- Collaboration with and assistance from vendor engineers
- Training sessions and hack-a-thons
- Computational time on NERSC systems
- Opportunity for a postdoctoral or early career staff to be placed within your team
- Depending on the scope of the engagement, NERSC will provide support ranging from a named contact to active contributions to your code or project.
Who Should Apply
Active NERSC users on projects that seek to better utilize advanced workflow capabilities, such as hardware acceleration, reconfigurable storage, advanced scheduling, and integration with edge services, should apply. We encourage applications with any or no levels of GPU readiness.
Workflows include but are not limited to, the following areas:
- High-performance simulation & modeling workflow (e.g., large-scale multi-physics applications with checkpoint/restart, data post-processing, and visualization)
- High-performance AI (HPAI) workflow (e.g., data integration-intensive science patterns such as training, inference, and hyperparameter optimization)
- Cross-facility workflow Rapid data analysis and real-time steering (e.g., time-sensitive science patterns such as Superfacility, edge, and hybrid cloud)
- Hybrid HPC-HPAI-HPDA workflow For seamless coupling of high-performance simulation & modeling, machine learning, and data analysis (HPDA) (e.g. long-term campaign science patterns, AI-in-the-loop, AI-around-the-loop)
- Scientific data lifecycle workflow Interactive, data analytics and visualization (e.g., data integration-intensive science patterns such as Jupyter, scientific databases, VSCode)
- External event-triggered and API-driven workflow (e.g., time-sensitive science patterns such as function-as-a-service, and microservices)
To apply, teams must
- Propose a defined project scope (please refer to the examples below, to help gauge the scope of your project).
- Articulate clear, specific, and measurable objective(s).
- Make specific resource commitments.
Project Scope and Goals
Applicants are expected to define performance improvement goals as well as integration of advanced technological capability in their workflows.
Performance Improvement Goals
Examples of targeted performance improvement goals are as follows:
- GPU enablement/optimization
- Reduced precision/Tensor cores
- Portability to Frontier/Aurora machines
- I/O compression/optimization
- Optimization of AI/ML-based workflows (including HPC + AI-in-the-loop)
- HPC + AI-in-the-loop
- New algorithms demonstrating a performance improvement at scale
Integration of Advanced Technology Capability
Examples of integration of advanced technology capability are as follows:
- IRI/Multi-site workflows
- Superfacility based API tools usage
- Cloud native/containerization
- Compression/Compute on NICs
- Persistent applications/workflows
- Quality-of-Service for storage
- Quality-of-Service for network/scheduling
- HPC + AI-in-the-loop
- Enabling new workflows on HPC (e.g., applying HPC to a new science area)
Selected teams will be expected for the duration of the award to commit to the following:
- Designate and maintain a point of contact for their code team.
- Appropriately staff the team for this work. (A minimum 0.5 FTE for the duration of your participation is strongly recommended.)
- Produce intermediate and final reports detailing science results and performance improvements.
Submissions will be evaluated on the following criteria:
- Computing resource usage within the DOE Office of Science
- Representation across the Office of Science programs.
- Potential for scientific advancements
- Potential for transfer of improvements to the broader community
- Resources available to augment NERSC/vendor capabilities.
- Commitment from the team (FTEs, etc.)
- Technical feasibility of the proposed work
- Clarity in scope, timeline, objectives, track record of previous engagements
To apply, please use the Application form for NESAP Term Projects.
This call for proposals is open from Monday, September 11, through Monday, October 2, 2023.
Selections will be announced by the end of October.