Heterogenous Computing Projects
Hybrid Advanced Workflows (HAW)
Principal Investigator: V.G. Castellana
Start Date: 3/22/2019
End Date: 9/30/2021
Address converged application challenges, such as the size of the data to process, different sparsity, and real-time power and trust constraints, by developing methodologies and tools for the design-space exploration of hardware-software architectures.

Software-Defined Architecture for Data Analysis (SO(DA)2)
Principal Investigator: A. Tumeo
Start Date: 3/22/2019
End Date: 3/21/2022
Provide required breakthroughs at various levels of the hardware-software stack to enable productive use of reconfigurable hardware.
Visit the SODA website.
Winner of the Best Poster Award at the Associate for Computing Machinery (ACM) Computing Frontiers 2022 Conference.

Computation-Flow Architecture (CFA): Designing Non-Von-Neumann Architecture for Future Data-Centric Computing
Principal Investigator: A. Li
Start Date: 8/6/2019
End Date: 8/5/2022
Demonstrate CFA effectiveness on key DMC kernels derived from machine learning and application domains.

Application-Algorithm-Architecture Co-Design for Large-Scale, Sparse Tensor/Matrix Methods (HiParTI)
Principal Investigator: J. Li
Start Date: 08/09/2019
End Date: 08/08/2022
Accelerate graph analytics (matrices), machine learning (tensor representations), quantum chemistry (tensor contractions), and protein dynamics (matrix-tensor methods).

Segmented Global Address Space (SGAS): A Memory-Centric Programming Model and Runtime
Principal Investigator: A. Lumsdaine
Start Date: 08/09/2019
End Date: 09/30/2021
Balance the tension between communication bandwidth and memory bandwidth requirements for fine-grained, data-driven applications.

Fixing Amdahl's Law within the Limits of Accelerated Systems (FALLACY)
Principal Investigator: A. Marquez
Start Date: 8/15/2019
End Date: 8/14/2022
Use memory-centric tools to inform hardware-software stack of application memory behavior and identification of memory costs and bottlenecks.

A Compiler Infrastructure for Data-Model Convergence (DuoMO)
Principal Investigator: G. Kestor
Start Date: 10/01/2019
End Date: 9/30/2022
Bridge the gap between application domains and fast-changing architectures and enable both domain scientists and hardware developers to focus on implementing efficient algorithms and hardware accelerators.

Navier: Dataflow Architectures for Computational Chemistry
Principal Investigator: R. Gioiosa
Start Date: 12/01/2021
End Date: 9/30/2022
Advance the capability of extreme scale computational chemistry by harnessing novel dataflow architectures and generalizing a dataflow-based software eco-system for integrating a heterogeneous pool of tightly coupled hardware resources and co-design novel dataflow accelerators to efficiently support sparse and irregular scientific workflows.