SparseAdapt: Runtime control for sparse linear algebra on a reconfigurable accelerator

Abstract

Dynamic adaptation is a post-silicon optimization technique that adapts the hardware to workload phases. However, current adaptive approaches are oblivious to implicit phases that arise from operating on irregular data, such as sparse linear algebra operations. Implicit phases are short-lived and do not exhibit consistent behavior throughout execution. This calls for a high-accuracy, low overhead runtime mechanism for adaptation at a fine granularity. Moreover, adopting such techniques for reconfigurable manycore hardware, such as coarse-grained reconfigurable architectures (CGRAs), adds complexity due to synchronization and resource contention. We propose a lightweight machine learning-based adaptive framework called SparseAdapt. It enables low-overhead control of configuration parameters to tailor the hardware to both implicit (datadriven) and explicit (code-driven) phase changes. SparseAdapt is implemented within the runtime of a recently-proposed CGRA called Transmuter, which has been shown to deliver high performance for irregular sparse operations. SparseAdapt can adapt configuration parameters such as resource sharing, cache capacities, prefetcher aggressiveness, and dynamic voltage-frequency scaling (DVFS). Moreover, it can operate under the constraints of either (i) high energy-efficiency (maximal GFLOPS/W), or (ii) high powerperformance (maximal GFLOPS3/W). We evaluate SparseAdapt with sparse matrix-matrix and matrixvector multiplication (SpMSpM and SpMSpV) routines across a suite of uniform random, power-law and real-world matrices, in addition to end-to-end evaluation on two graph algorithms. SparseAdapt achieves similar performance on SpMSpM as the largest static configuration, with 5.3× better energy-efficiency. Furthermore, on both performance and efficiency, SparseAdapt is at most within 13% of an Oracle that adapts the configuration of each phase with global knowledge of the entire program execution. Finally, SparseAdapt is able to outperform the state-of-the-art approach for runtime reconfiguration by up to 2.9× in terms of energy-efficiency.