Search papers, labs, and topics across Lattice.
This paper addresses the challenge of mapping unstructured multiphysics codes onto Spatial Dataflow Architectures, which are optimized for structured grid kernels. By employing a model-centric approach that combines communication and memory modeling, the authors propose a joint decomposition strategy that minimizes communication overhead and optimizes memory usage. The results demonstrate that the Livermore Unstructured Lagrangian Explicit Shock Hydrodynamics (LULESH) application can be effectively mapped to the Cerebras Wafer-Scale Engine, achieving superior performance compared to traditional GPU implementations.
Unstructured grid codes can outperform GPUs on Spatial Dataflow Architectures, challenging conventional assumptions about hardware limitations.
Spatial Dataflow Architectures are an emerging hardware pattern in high-performance computing, whose mesh-connected fixed-memory processing elements are tailored for structured grid kernels with two-dimensional neighborhoods. However, practical multiphysics codes are often computed on unstructured grids, which induce indirect memory accesses and high-dimensional communication patterns, making them infeasible to directly map onto said architectures. This work takes a principled, model-centric approach to partitioning unstructured problems onto spatial dataflow architectures. Through communication and memory modeling, we propose a joint decomposition that considers both the size of the application's fields and its subroutines. In particular, we automate the analysis process of the original code, define a high-dimensional decomposition that minimizes communication via space-filling curves, and apply memory optimization techniques, crucial in this memory-limited environment. We demonstrate mapping the Livermore Unstructured Lagrangian Explicit Shock Hydrodynamics (LULESH) application to the Cerebras Wafer-Scale Engine, showing that larger, unstructured grid codes can still outperform GPUs.
