Scalable Parallelism in the Extreme | 17 600

Opportunity Information: Apply for 17 600

The National Science Foundation (NSF) grant opportunity titled Scalable Parallelism in the Extreme (SPX) focuses on a major turning point in computing: performance gains can no longer rely on faster single processors the way they did during the peak years of Moore's Law. Modern computing has shifted from single-core machines to systems dominated by multicore and many-core processors, specialized accelerators, and massive cloud and warehouse-scale platforms. At the same time, semiconductor scaling is running into hard physical and economic limits, and single-processor performance improvements have largely plateaued. SPX is framed as a response to that reality, aligned with the National Strategic Computing Initiative (NSCI), which calls for a long-term path forward for high-performance computing (HPC) in the "post-Moore's Law" era.

The core idea behind SPX is that the most promising way to keep improving performance is to exploit parallelism much more effectively than we do today. Parallel computing is not new, but SPX emphasizes that future systems will operate at unprecedented scale, and that scale changes everything. Small inefficiencies that might be tolerable now can become catastrophic when multiplied across huge numbers of cores, nodes, and services. Even short serial sections of code can severely limit speedups, reflecting classic limits like Amdahl's Law but in a more punishing real-world setting. The program highlights how heterogeneity is becoming the norm, with CPUs, GPUs, and other accelerators working together, each with different performance characteristics and complex memory hierarchies. Managing these mixed environments efficiently is a major challenge, especially when developers need predictable performance and portability rather than one-off tuning for each platform.

SPX also targets system-level issues that emerge as architectures become more complex. Extreme-scale parallel systems can suffer from unacceptable latency, coordination overhead, and reduced reliability, including shorter mean time between failures simply because there are so many components involved. Even if each individual part is highly reliable, the overall system can fail more often at scale. Another major theme is the difficulty of understanding and guaranteeing performance behavior in highly interconnected distributed environments, where communication, synchronization, contention, and topology can strongly influence outcomes. The opportunity notes that these are only a few examples of the kinds of problems that become dominant when parallelism is pushed to extremes.

In terms of what NSF is trying to fund, SPX aims to support research that raises performance in this new era through coordinated advances across the stack. The solicitation explicitly calls for collaborative work spanning multiple research areas, from services and applications down through runtime systems, operating systems, compilers, programming models, and even micro-architecture. The intent is not only to produce isolated ideas, but to build approaches that scale, remain programmable, and can be reused as platforms evolve. The program also encourages coordination with industry efforts pursuing similar goals, reflecting that real-world impact and technology transition matter in this space.

SPX is positioned to support the full set of NSCI strategic objectives, tying the research to national priorities for future computing. These include foundational research in architectures and software that can drive performance gains beyond traditional semiconductor scaling; development and deployment of programmable, scalable, reusable platforms within national HPC and scientific cyberinfrastructure; stronger integration between data analytics and modeling and simulation workflows; and progress toward capable extreme-scale computing. Overall, the opportunity is meant to push the field toward practical, scalable parallel computing methods that can sustain performance growth even when hardware improvements alone are no longer enough.

Administrative details from the source listing include: the agency is NSF; the funding instrument is a grant; the activity category is science and technology and other R and D; the CFDA number is 47.070; the opportunity number is 17-600; it anticipated around 25 awards; and it listed an award ceiling of 0 (typically meaning the ceiling is not specified in that field rather than indicating no funding). The original closing date shown was January 9, 2018, and eligibility is listed as "Others" with clarification referenced in an additional eligibility section.

• The National Science Foundation in the science and technology and other research and development sector is offering a public funding opportunity titled "Scalable Parallelism in the Extreme" and is now available to receive applicants.
• Interested and eligible applicants and submit their applications by referencing the CFDA number(s): 47.070.
• This funding opportunity was created on Sep 23, 2017.
• Applicants must submit their applications by Jan 09, 2018. (Agency may still review applications by suitable applicants for the remaining/unused allocated funding in 2026.)
• The number of recipients for this funding is limited to 25 candidate(s).
• Eligible applicants include: Others (see text field entitled Additional Information on Eligibility for clarification).

Apply for 17 600

[Watch] Creating a grant proposal using the step-by-step wizard inside the applicant portal: