統一計算架構GPU高效能計算
|
課程名稱 |
統一計算架構GPU高效能計算 |
1學分 1小時 |
|
英文課程名稱 |
GPU high-performance computing using Compute Unified Device Architecture (CUDA) |
|
|
中文課程概要 |
隨著科技的不斷發展與進步，數據亦越來越龐大，因而造成資料處理、分析與應用的困難。過去十年來，應用統一計算架構(Compute Unified Device Architecture, CUDA) 之繪圖處理單元（Graphics Processing Unit , GPU）的快速進步，已經讓它的應用範圍不只侷限於繪圖處理平台，而是擴大應用到高效能運算上，其多處理器核心的平行運算架構以及高記憶體頻寬，同時提供了高運算能力及低成本兩大優點。因此如何融入高效能運算技術，藉以提升整體運算效率，已逐漸成為此一領域重要議題之一。 本課程將對巨量資料高效能運算技術做一系列介紹，內容包含如下： 1. 巨量資料背景介紹 2. 平行計算概論 3. GPU-CUDA介面介紹 4. CUDA程式撰寫 5. GPU-CUDA程式撰寫最佳化 |
|
|
英文課程概要 |
With modern technology developing, the data is increasingly huge. It caused difficulties in data processing, analysis and applications. Over the decades, graphics processing unit (GPU) has not only expanded the scope of application of graphics processing platforms, but also the application of the high performance computing. Its many-core computing architecture and high memory bandwidth advance the high performance computing with low cost. Thus how to integrate a modern GPU architecture with NVIDIA’s compute unified device architecture (CUDA) technology to enhance overall operational efficiency has become one of the important issues. In order to discuss this important topic, this course is aimed to cover the basis as follows： 1. Introduction to the big data 2. Introduction to the principles of parallel programming 3. Introduction to GPU-CUDA interfaces 4. GPU programming using CUDA 5. Optimization of GPU-CUDA programming |
|