- Using Multiple Processes to Improve System Productivity
- Multiple Users Utilizing a Single System
- Improving Machine Efficiency Through Consolidation
- Using Parallelism to Improve the Performance of a Single Task
- Parallelization Patterns
- How Dependencies Influence the Ability Run Code in Parallel
- Identifying Parallelization Opportunities
- Summary
Summary
This chapter has discussed the various strategies that can be used to utilize systems more efficiently. These range from virtualization, which increases the productivity of the system through increasing the number of active applications, to the use of parallelization techniques that enable developers to improve the throughput or speed of applications.
It is important to be aware of how the amount of code that is made to run in parallel impacts the scaling of the application as the number of threads increases. Consideration of this will enable you to estimate the possible performance gains that might be attained from parallelization and determine what constraints need to be met for the parallelization to be profitable.
The chapter introduces various parallelization strategies, and these should provide you with insights into the appropriate strategy for the situations you encounter. Successful parallelization of applications requires identification of the dependencies present in code. This chapter demonstrates ways that the codes can be made parallel even in the presence of dependencies.
This chapter has focused on the strategies that might be employed in producing parallel applications. There is another aspect to this, and that is the handling of data in parallel applications. The individual threads need to coordinate work and share information. The appropriate method of sharing information or synchronizing will depend on the implementation of the parallelization strategy. The next chapter will discuss the various mechanisms that are available to support sharing data between threads and the ways that threads can be synchronized.