Graphics Processing Units, an overview

by Patrick H. Stakem

Overview

This book discusses the topic of Graphics Processing Units, which are specialized units found in most modern computer architectures. Although we can do operations of graphics data in regular arithmetic logic units (ALU's), the hardware approach is much faster, Just like for floating pount arithmetic, specialized units speed up the process. We will discuss the applications for GPU's, the data format, and the operations they perform. These specialized units are the backbone to video, and to a large extent audio processing in modern computer architectures.

The GPU is a specialized computer architecture, focused on image data manipulation for graphics displays and picture processing. It has applications far that. The normal ALU, Arithmetic-Logic Unit, in a computer does the four basic math operations, and logical operations on integers. These integers are usually 32 or 64 bits at this time. The GPU greatly enhances the spped of 3D graphics.

GPU's find application in arcade machines, games consoles, pc's, tablets, phones, car dashboards, tv's and entertainment systems.

GPU's can process integer and floating point data much faster than a cpu, if it is presented in the right format. They don't have all the general purpose features of ALU's, but they can contain 100 cores or more. This has lead to the employment of large numbers of GPU's as the basis for the current generation of Supercomputers.

Product Details

BN ID:	2940157399757
Publisher:	PRRB Publishing
Publication date:	05/22/2017
Sold by:	Barnes & Noble
Format:	eBook
File size:	42 KB

About the Author

The author's first computer experience was on a Bendix G-20 mainframe, a 32-bit machine, using the Algol language. His first assembly language was on an IBM S/360 mainframe, specifically the Model 67 variant, with virtual memory. He went on to program the Univac 1108 series mainframe , the DEC PDP-8, 9, 10, and 11, the Bendix G-15, the Athena Missile Guidance Computer by Sperry Rand and many more. The concept of a personal computer was, at the time, ludicrous.
The author is fairly conversant with IBM Fortran -G and -H, S/360 assembly language, S/360 JCL, Univac Fortran-V, and Bendix G-20 Algol.
In spite of all the interesting computers, the author managed to graduate from Carnegie Mellon University with a BSEE. He went on to get Masters degrees from the Johns Hopkins University in Physics and Computer Science. He teaches for Johns Hopkins University, Engineering for Professionals Program.