A Compiler Framework for Optimizing Dynamic Parallelism on GPUs

TL;DR

This paper introduces a compiler framework with three key optimizations—thresholding, coarsening, and aggregation—to improve the performance of dynamic parallelism on GPUs, especially for irregular nested workloads.

Contribution

It presents a novel compiler framework that optimizes dynamic parallelism on GPUs through thresholding, coarsening, and aggregation techniques, reducing performance penalties.

Findings

Achieves a 43.0x geometric mean speedup over non-optimized dynamic parallelism.

Attains an 8.7x speedup over applications without dynamic parallelism.

Provides a 3.6x improvement over prior dynamic parallelism aggregation methods.

Abstract

Dynamic parallelism on GPUs allows GPU threads to dynamically launch other GPU threads. It is useful in applications with nested parallelism, particularly where the amount of nested parallelism is irregular and cannot be predicted beforehand. However, prior works have shown that dynamic parallelism may impose a high performance penalty when a large number of small grids are launched. The large number of launches results in high launch latency due to congestion, and the small grid sizes result in hardware underutilization. To address this issue, we propose a compiler framework for optimizing the use of dynamic parallelism in applications with nested parallelism. The framework features three key optimizations: thresholding, coarsening, and aggregation. Thresholding involves launching a grid dynamically only if the number of child threads exceeds some threshold, and serializing the child threads in the parent thread otherwise. Coarsening involves executing the work of multiple thread blocks by a single coarsened block to amortize the common work across them. Aggregation involves combining multiple child grids into a single aggregated grid. Our evaluation shows that our compiler framework improves the performance of applications with nested parallelism by a geometric mean of 43.0x over applications that use dynamic parallelism, 8.7x over applications that do not use dynamic parallelism, and 3.6x over applications that use dynamic parallelism with aggregation alone as proposed in prior work.