CUDA Enabled for Android Tablets through DS-CUDA
CUDA Enabled for Android Tablets through DS-CUDA
复制标题
通过 DS-CUDA 为 Android 平板电脑启用 CUDA
DOI：
--
复制
发表时间：
2013
期刊：
影响因子：
--
通讯作者：
Narumi Tetsu
中科院分区：
文献类型：
--
作者：
EdgarJosafatMartinez;Kawai Atsushi;Yoshikawa Kazuyuki;Yasuoka Kenji;Narumi Tetsu
使用GPU（图形处理器单元）进行并行计算，GPGPU（通用计算GPU），已成为一种自然的方式，以显着提高计算机应用程序和模拟的性能。前十大超级计算机中的一些已经配备了许多GPU加速器[1]。自2006年首次亮相以来，NVIDIA的CUDA（计算统一设备架构）已成为使用GPU进行并行计算的最完整和最常见的框架之一。CUDA已经成功地加速了众多领域的许多应用[3-4]。另一方面，自从智能手机和平板电脑出现以来，我们与计算机交互的方式已经改变。这些相对较新的设备提供了一种新的方式来表示数据并与其交互（即触摸显示器），从不同的来源（即陀螺仪，加速度计，相机等）获取输入。然而，与PC相比，它们的计算能力仍然非常有限。在本文中，我们建议在Android平板电脑上使用DS-CUDA（分布式共享CUDA），以合并这两个世界，移动性和高性能计算。
The usage of GPUs (Graphics Processor Units) for parallel computing, GPGPU (General-Purpose Computing on GPU), has become a natural way to increase dramatically the performance of computer applications and simulations. Some of the top 10 super computers are already equipped with many GPU accelerators [1]. Since its first appearance in 2006, NVIDIA’s CUDA [2] (Compute Unified Device Architecture) has become one of the most complete and common frameworks for parallel computing using GPUs. CUDA has already successfully accelerated many applications in an enormous variety of fields [3-4]. On the other hand, the way we interact with computers has been changed since the appearance of smartphones and tablets. These relative new devices offer a new way to represent data and interact (i.e. touch displays) with it taking input from different sources (i.e. gyro, accelerometer, camera etc.). However their computing power is still very limited compared with PCs. In this paper, we propose to use DS-CUDA (Distributed-Shared CUDA), on Android tablet in order to merge these two worlds, mobility and high performance computing.