• Dr. Paresh M Dholakia
• Pankti Dholakia

Over the last decade, GPUs have been extensively used to meet the computational needs for big data and machine learning applications. In this study we propose to study and review NVIDIA’s CUDA (Compute Unified Device Architecture) – a parallel computing library that provides APIs that allows varied applications to utilize the underlying GPU. This paper will discuss the CUDA model and improvements since CUDA 4.0 and how it compares to OpenCL. This graphic library enables compute intensive applications to leverage GPUs for parallel processing and this research explores the dominant performance of CUDA with GPUs using exploratory deep learning applications.

GPU, NVIDIA, CUDA

"A Deep Dive into CUDA Architecture", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.4, Issue 5, page no.11-17, November-2017, Available :http://www.jetir.org/papers/JETIR1711228.pdf

"A Deep Dive into CUDA Architecture", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.4, Issue 5, page no. pp11-17, November-2017, Available at : http://www.jetir.org/papers/JETIR1711228.pdf