Skip to navigation Skip to content
Careers | Phone Book | A - Z Index
Performance and Algorithms Research



CSPACER is a lightweight communication runtime for application-specific optimized communication patterns. It is composed of two layers.  The lowest layer is a sub-microsecond communication layer that implements the consistent space abstraction primitives. On top of this layer is a communication pattern layer that is used as integration skeletons with the application.

 The main focus of this effort is communication-bound applications, especially with irregular communication patterns. The target programming abstraction for the runtime is the space consistency model, which defines consistency guarantees at the granularity of memory spaces. This model has relaxed consistency semantics that enables a wide range of runtime optimizations. The runtime leverages threading to accelerate communication primitives, especially collective operations. It also allows efficient pipelining of communication operations and enable constructing a consistent state of multiple unordered communication activities targeting a memory space. The runtime uses a reduced API design that decomposes complex communication primitives in traditional general-purpose runtime into a sequence of simpler steps. To improve the productivity of using this runtime, we provide communication patterns commonly used for regular scientific computing applications and irregular data analytics. These communication patterns offer skeletons for the integration with application computation to allow efficient overlap.


Khaled Ibrahim

Roel Van Beeumen


 [1] Khaled Z. Ibrahim. 2021. CSPACER: A Reduced API Set Runtime for the Space Consistency Model. In The International Conference on High Performance Computing in Asia-Pacific Region (HPC Asia 2021). Association for Computing Machinery, New York, NY, USA, 58–68.

[2] K. Ibrahim, "Optimizing Breadth-First Search at Scale Using Hardware-Accelerated Space Consistency," 2019 IEEE 26th International Conference on High Performance Computing, Data, and Analytics (HiPC), 2019, pp. 23-33, doi:10.1109/HiPC.2019.00015.

[3] Van Beeumen R, Ibrahim KZ, Kahanamoku–Meyer GD, Yao NY, Yang C. Enhancing scalability of a matrix-free eigensolver for studying many-body localization. The International Journal of High Performance Computing Applications. 2022;36(3):307-319. doi:10.1177/10943420211060365