Faster File Matching using GPGPUs
Department of Computer and Information Sciences, University of Delaware
Symposium on Application Accelerators in High Performance Computing, 2010
@article{mohan2010faster,
title={Faster File Matching Using GPGPUs},
author={Mohan, D. and Cavazos, J.},
booktitle={Application Accelerators in High Performance Computing, 2010 Symposium, Papers},
year={2010}
}
We address the problem of file matching by modifying the MD6 algorithm that is best suited to take advantage of GPU computing. MD6 is a cryptographic hash function that is tree-based and highly parallelizable. When the message M is available initially, the hashing operations can be initiated at different starting points within the message and their results can be aggregated as the final step. In the parallel implementation, the MD6 program was partitioned and effectively parallelized across the GPU using CUDA. To demonstrate the performance of the CUDA version of MD6, we performed various experiments with inputs of different MD6 buffer sizes and varying file sizes. CUDA MD6 achieves real time speedup of more than 250X over the sequential version when executed on larger files. CUDA MD6 is a fast and effective solution for identifying similar files.
February 19, 2011 by hgpu
