Abstract | ||
---|---|---|
Sparse solvers are heavily used in computational fluid dynamics (CFD), computer-aided design (CAD), and other important application domains. These solvers remain challenging to execute on massively parallel architectures, due to the sequential dependencies between the fine-grained application tasks. In particular, parallel sparse solvers typically suffer from substantial scheduling and dependency-management overheads relative to the compute operations. We propose adaptive task aggregation (ATA) to efficiently execute such irregular computations on GPU architectures via hierarchical dependency management and low-latency task scheduling. On a gamut of representative problems with different data-dependency structures, ATA significantly outperforms existing GPU task-execution approaches, achieving a geometric mean speedup of 2.2X to 3.7X across different sparse kernels (with speedups of up to two orders of magnitude). |
Year | DOI | Venue |
---|---|---|
2019 | 10.1109/PACT.2019.00033 | 2019 28th International Conference on Parallel Architectures and Compilation Techniques (PACT) |
Keywords | Field | DocType |
data dependency, fine-grained parallelism, GPUs, runtime adaptation, scheduling, sparse linear algebra, task parallel execution | CAD,Kernel (linear algebra),Gamut,Task analysis,Scheduling (computing),Massively parallel,Computer science,Parallel computing,Computation,Speedup | Conference |
ISSN | ISBN | Citations |
1089-795X | 978-1-7281-3614-1 | 0 |
PageRank | References | Authors |
0.34 | 35 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ahmed E. Helal | 1 | 14 | 3.39 |
Ashwin M. Aji | 2 | 9 | 1.55 |
Michael L. Chu | 3 | 0 | 0.34 |
Bradford M. Beckmann | 4 | 0 | 0.34 |
Wu-chun Feng | 5 | 2812 | 232.50 |