Efficient Irregular Wavefront Propagation Algorithms on Hybrid CPU-GPU Machines

TL;DR

This paper presents optimized algorithms for the irregular wavefront propagation pattern on hybrid CPU-GPU systems, achieving significant speedups in image processing tasks by improving data access and parallelization strategies.

Contribution

It introduces a multi-level queue structure and tile-based parallelization for efficient wavefront propagation on hybrid systems, enabling high performance in image processing applications.

Findings

Achieved up to 85x speedup over single-core CPU implementations.

Developed a multi-level queue structure to improve GPU memory utilization.

Implemented a tile-based parallelization strategy for multi-CPU and GPU execution.

Abstract

In this paper, we address the problem of efficient execution of a computation pattern, referred to here as the irregular wavefront propagation pattern (IWPP), on hybrid systems with multiple CPUs and GPUs. The IWPP is common in several image processing operations. In the IWPP, data elements in the wavefront propagate waves to their neighboring elements on a grid if a propagation condition is satisfied. Elements receiving the propagated waves become part of the wavefront. This pattern results in irregular data accesses and computations. We develop and evaluate strategies for efficient computation and propagation of wavefronts using a multi-level queue structure. This queue structure improves the utilization of fast memories in a GPU and reduces synchronization overheads. We also develop a tile-based parallelization strategy to support execution on multiple CPUs and GPUs. We evaluate our approaches on a state-of-the-art GPU accelerated machine (equipped with 3 GPUs and 2 multicore CPUs) using the IWPP implementations of two widely used image processing operations: morphological reconstruction and euclidean distance transform. Our results show significant performance improvements on GPUs. The use of multiple CPUs and GPUs cooperatively attains speedups of 50x and 85x with respect to single core CPU executions for morphological reconstruction and euclidean distance transform, respectively.