Linux High Performance Clusters Paperback
An in-depth look at creating clustered high-performance computing systems with Linux and off the shelf computer hardware.
This book covers both distributed systems as well as shared memory super-computing clusters.
The author is the founder of Distributed Systems Research Lab, and is one of the leading authorities in this field.
This book provides real-world examples with source code while also providing enough theory and overview to aid readers in getting the most out of their systems Based on "Linux Cluster Architecture", the highly regarded book on Linux clustering, "High Performance Linux Cluster Architectures" takes the unique approach of focusing on building high-performance Linux clusters.
Building from the ground up, this book guides the reader through the nuts and bolts of creating a tuned distributed cluster, and a shared memory super-computing cluster along with a set of performance monitoring tools. This book shows the reader all the steps involved, including how to interconnect the hardware, how to configure the software, how to set up a user home directory for software development, how to write distributed system software, how to instrument the source code, analyze the performance results, and how to tune the system.
Readers who actually build such a system and fiddle with the source code should gain modern, marketable skills that are becoming even more relevant as the mobile Internet makes its way onto our belts.
The reader should be familiar with the C programming language, have an idea about how computers can be interconnected via a network, and have some experience developing software in a UNIX environment, preferably Linux.
Two very basic chapters are included on these issues and a chapter on the C software skills necessary to develop distributed systems.
Beyond that is an important concept of how the same network of PCs can become a data server or a supercomputer, by simply running different software.