SUMMARY This paper proposes a parallel scheme for accelerating parameter sweep applications on a graphics processing unit. By using hundreds of cores on the graphics processing unit, we found that our scheme simultaneously processes multiple parameters rather than a single parameter. The simultaneous sweeps exploit the similarity of computing behaviors shared by different parameters, thus allowing memory accesses to be coalesced into a single access if similar irregularities appear among the parameters’ computational tasks. In addition, our scheme reduces the amount of off‐chip memory access by unifying the data that are commonly referenced by multiple parameters and by placing the unified data in the fast on‐chip memory. In several experiments, we applied our scheme to practical applications and found that our scheme can perform up to 8.5 times faster than a naive scheme that processes a single parameter at a time. We also include a discussion on application characteristics that are required for our scheme to outperform the naive scheme. Copyright © 2013 John Wiley & Sons, Ltd.

Flint is a simulator that numerically integrates heterogeneous biophysical models described by a large set of ordinary differential equations. It uses an internal bytecode representation of simulation-related expressions to handle various biophysical models built for general purposes. We propose two acceleration methods for Flint using a graphics processing unit (GPU). The first method interprets multiple bytecodes in parallel on the GPU. It automatically parallelizes the simulation using a level scheduling algorithm. We implement an interpreter of the Flint bytecode that is suited for running on the GPU, which reduces both the number of memory accesses and divergent branches to achieve higher performance. The second method translates a model into a source code for both the CPU and the GPU through the internal bytecode, which speeds up the compilation of the generated source codes, because the code size is diminished because of bytecode unification. With large models such that tens of thousands or more expressions can be evaluated simultaneously, the translated code running on the GPU achieves computational performance of up to 2.7 higher than that running on a CPU. Otherwise, with small models, the CPU is faster than the GPU. Therefore, the translated code dynamically determines on which to run either the CPU or the GPU by profiling initial few iterations of the simulation.

This paper proposes a fast method for computing the costs of all-pairs shortest paths (APSPs) on the graphics processing unit (GPU). The proposed method is implemented using compute unified device architecture (CUDA), which offers us a development environment for performing general-purpose computation on the GPU. Our method is based on Harish's iterative algorithm that computes the cost of the single-source shortest path (SSSP) for every source vertex. We present that exploiting task parallelism in the APSP problem allows us to efficiently use on-chip memory in the GPU, reducing the amount of data being transferred from relatively slower off-chip memory. Furthermore, our task parallel scheme is useful to exploit a higher parallelism, increasing the efficiency with highly threaded code. As a result, our method is 3.4--15 times faster than the prior method. Using on-chip memory, our method eliminates approximately 20% of data loads from off-chip memory.