Analysis of OpenCL on CUDA Platform

[manojthedear1]

Analysis of OpenCL on CUDA Platform
Submitted by
Abhijeet Kumar Singh
INDIAN INSTITUTE OF INFORMATION TECHNOLOGY
ALLAHABAD

ABSTRACT:
In this project, the newly released GPGPU platform OpenCL has been studied. The field of High Scale Computation (HSC) has seen a lot of development in the recent times, especially the GPGU (General Purpose Graphical Processing Unit). the previously existing technology of CUDA is expected to be outperformed by the OpenCL as it is platform independent. The CUDA can perform only in the CUDA compatible NVIDIA graphic cards. The options of Opencl and CUDA is available for the CUDA platform. In this project work, the the performance of the two is compared on a Nvidia CUDA enabled system. The performance is analysed and the conclusion of the better one for the high performance computing is reached.

CUDA ( Compute Unified Device Architecture)
CUDA is a parallel computing architecture owned by NVIDIA. NVIDIA s GPUs feature this technology and it is made available to the developers through industry standard programming languages. CUDA C which is C with CUDA extensions is used by the developers and a Open64 C compiler is used to compile the code to run on the GPU.

Full report:
[attachment=8278]