Chapter12Rev08 - Chapter 12(Revision number 8 Multithreaded...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
1 Chapter 12 (Revision number 8) Multithreaded Programming and Multiprocessors What is multithreading ? It is a technique for a program to do multiple tasks, perhaps concurrently. Since the early stages of computing, exploiting parallelism has been a quest of computer scientists. As early as the 70’s, languages such as Concurrent Pascal and Ada have proposed features for expressing program-level concurrency. Humans think and do things in parallel all the time. For example, you may be reading this book while you are listening to some favorite music in the background. Often, we may be having an intense conversation with someone on some important topic (such as world peace!), while working on something with our hands, may be fixing a car or folding our laundry. Given that computers extend the human capacity to compute, it is only natural to provide the opportunity for the human to express concurrency in the tasks that they want the computer to do on their behalf. Sequential programming forces us to express our computing needs in a sequential manner. This is unfortunate since humans think in parallel but end up coding up their thoughts sequentially! For example, let us consider an application such as video surveillance. We want the computer to gather images from ten different cameras continuously, analyze each of them individually for any suspicious activity, and raise an alarm in case of a threat. There is nothing sequential in this description. If anything, the opposite is true. Yet if we want to write a computer program in C to carry out this task, we end up coding it sequentially. The intent of this chapter is to introduce concepts in developing multithreaded programs, the operating system support needed for these concepts, and the architectural support needed to realize the operating system mechanisms. The important point we want to convey in this chapter is that the threading concept and the system support for threading are simple and straightforward. 12.1 Why Multithreading? Multithreading allows expressing the inherent parallelism in the algorithm. A thread is similar to a process in that it represents an active unit of processing. Later in this chapter, we will discuss the semantic differences between a thread and a process. Suffice it to say at this point that a user level process may comprise multiple threads. How does multithreading help at the programming level? First, it allows the user program to express its intent for concurrent activities in a modular fashion, just as the procedure abstraction helps organize a sequential program in a modular fashion. Second, it helps in overlapping computation with inherent I/O activities in the program. We know from Chapter 10, that DMA allows a high-speed I/O device to communicate directly with the memory without the intervention of the processor. Figure 12.1 shows this situation for a program that periodically does I/O but does not need the result of the I/O immediately. Expressing the I/O activity as a separate thread helps in overlapping computation with communication.
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 11/25/2010 for the course CENG 100 taught by Professor Ceng during the Spring '10 term at Universidad Europea de Madrid.

Page1 / 51

Chapter12Rev08 - Chapter 12(Revision number 8 Multithreaded...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online