GPU‐based branch‐and‐bound method to solve large 0‐1 knapsack problems with data‐centric strategies

Abstract

An out-of-core branch-and-bound (B&B) method to solve large 0-1 knapsack problems on a graphics processing unit (GPU) is proposed. Given a large problem that produces many subproblems, the proposed method dynamically swaps subproblems to CPU memory. Because such a CPU-centric subproblem management scheme increases CPU-GPU data transfer, we adopt three data-centric strategies to eliminate this side effect. The first is an out-of-order search (O3S) strategy that reduces the data transfer overhead by adaptively transferring subproblems between the CPU and GPU. The second is an explicitly-managed pipelining strategy that hides the data transfer overhead by overlapping data transfer with GPU-based B&B operations. The third is a GPU-based stream compaction strategy that reduces the sparseness of arrays to be transferred. Experimental results demonstrate that the proposed out-of-core method stored 41 times as many subproblems as a previous in-core method that manages subproblems in GPU memory, solving approximately twice as many problem instances on the GPU. In addition, compared to a previous breadth-first search (BFS) strategy, the proposed O3S strategy achieved an average speedup of 7.5 times. KEYWORDSbranch and bound, data-centric method, GPU, knapsack, out-of-core computation Concurrency Computat Pract Exper. 2019;31:e4954.wileyonlinelibrary.com/journal/cpe FIGURE 3 Array-based subproblem management scheme. Pruned subproblems (ie, passive; X) make the array sparse as the depth of the search tree increases. Note that the branching, bounding, and pruning operations are performed on the GPU SUPPORTING INFORMATIONAdditional supporting information may be found online in the Supporting Information section at the end of the article.How to cite this article: Shen J, Shigeoka K, Ino F, Hagihara K. GPU-based branch-and-bound method to solve large 0-1 knapsack problems with data-centric strategies. Concurrency Computat Pract Exper. 2019;31:e4954. https://doi.

Accelerating the Held-Karp Algorithm for the Symmetric Traveling Salesman Problem

“…Future work is to deal with large problem sizes that cannot be naively stored in the GPU memory due to memory exhaustion. An out-of-core processing scheme [18], [19] may be useful for realizing this large-scale computation; the scheme divides data into small pieces and iteratively processes the pieces with overlapping GPU computation with CPU-GPU data transfer.…”

Novel Route Planning System for Machinery Selection. Case: Slurry Application

“…Therefore, this challenge has led to the need to develop algorithms that can produce near-optimal solutions in a reasonable amount of time [12]. During the years, different solution approaches have been developed including exact algorithms (such as branch-and-bound [13,14] and branch-and-cut [15]), heuristic algorithms (such as the Clarke-Wright savings algorithm [16]), and metaheuristic algorithms (such as simulated annealing [17,18], genetic algorithms [19], tabu search [20], and ant algorithms [21]. Earlier, conventional heuristic algorithms were designed as a response to limited computer processing power.…”

A Data-Centric Directive-Based Framework to Accelerate Out-of-Core Stencil Computation on a GPU

“…Currently, the graphics processing unit (GPU) is considered to be the most efficient architecture for parallel stencil code [7]. Armed with thousands of cores and 5-10 times higher memory bandwidth than CPUs, GPUs provide powerful solutions for both compute-and memory-intensive scientific prob-lems [8]- [11]. However, there are two main challenges in implementing GPU-accelerated stencil code: limited capacity of device (i.e., GPU) memory and considerable programming effort to implement GPU-accelerated code.…”