| Name | Modified | Size | Downloads / Week |
|---|---|---|---|
| OCM.tgz | 2012-06-05 | 63.4 kB | |
| CCM.tgz | 2012-06-05 | 32.9 kB | |
| README.txt | 2011-06-28 | 2.1 kB | |
| Totals: 3 Items | 98.4 kB | 0 |
******************************************************************************** This project contains the code tu run and evaluate PCM, a parallel content-based matching algorithm designed to efficiently run on multi-core CPUs and on CUDA GPUs. ******************************************************************************** Two separate versions of PCM can be downloaded: ******************************************************************************** 1) OCM (OpenMP Content-Based Matching) runs on multi-core CPUs using OpenMP. It requires GCC version 4.2 or greater to compile. Download and expand OCM.tar.gz make all to compile make clean to remove all compiled objects ./OCM to run ******************************************************************************** 2) CCM (CUDA Content-Based Matching) runs on CUDA GPUs. It requires requires CUDA drivers, CUDA toolkit (version 3.0 or greater), and a cuda-capable GPU to run. If you are running on a 32 bit machine, remove the -m64 option in Makefile to compile correctly Check the compute capabilities of your GPU and modify the -arch option in Makefile accordingly. Use -arch sm_11 if you are not sure of the capabilities of your hardware. Download and expand CCM.tar.gz make all to compile make clean to remove all compiled objects ./CCM to run ******************************************************************************** Detailed descriptions and evaluations of CCM and OCM can be found in the following papers: "High Performance Content-Based Matching Using GPUs" by Alessandro Margara and Gianpaolo Cugola http://home.dei.polimi.it/margara/Papers/cudaFF.pdf "High Performance Content-Based Matching Using Off-The-Shelves Parallel Hardware" by Alessandro Margara and Gianpaolo Cugola http://home.dei.polimi.it/margara/Papers/PCM.pdf ******************************************************************************** For further information, please refer to my personal page at home.dei.polimi.it/margara ********************************************************************************