Traditionally, computer science (CS) curricula have done little to teach practical aspects of the discipline of computing. More specifically, CS education does precious little to make students aware of the resource costs of their work, and there is ample indication that the cheap availability of computing power (in comparison with the scarcity and hardships of earlier decades) has engendered wastefulness: “People knew how to write small, efficient programs in those days, a skill that has subsequently been lost” [1]. Parkinson’s second law doubtless holds true of computing as much as it does of other things.

While all computer scientists have been educated to understand computational complexity in a mathematical sense, and can analyze the complexities of abstract algorithms in terms of big-oh and similar, there is little understanding of performance issues with whole computing systems doing practical tasks, and thereby almost none of the resource costs (including environmental damage) are due to computing. Even leading researchers in the world are well short of being able to analytically predict computing systems’ performance accurately from the ground up. There is good understanding at the complementary metal-oxide semiconductor (CMOS) level, but not so much even at the chip level and much less for larger hardware or software systems.

Perhaps, soon enough we will need to learn such skills, even more particularly with energy consumption (that is, predicting how much energy will be consumed to run a particular code fragment on a certain reference architecture), but the state of the theory and practice are well short of what is needed.

In recent years, particularly within the past decade, there has been a great burgeoning of interest in such issues, in particular toward reducing the energy consumption of computing systems. This is in part due to increased awareness in society of environmental issues, global warming, and energy scarcity, but also because energy cost (that is, the electricity bill) is the biggest item in the lifetime cost of large computing systems such as data centers. The terms “green IT” and “green computing” are often used; perhaps the adjective “green” is somewhat misleading, as it connotes “concerned with or supporting protection of the environment,” a strictly larger objective than mere energy efficiency, which is what most “green” computing is actually about.

The present anthology has its place in this context. It consists of nine chapters dealing with various topics in green computing. The first is an overview of power-reduction approaches used in the IBM Blue Gene supercomputer. Subsequent chapters, by various sets of authors, mostly focus on machine-level approaches to energy efficiency (for example, “Compiler-Driven Energy Efficiency,” “Cross-Layer Power Management”).

Though doubtless a very valuable addition to the literature on its topic, one downside of the book is that it seems to use as a baseline literature that is partly dated; most references given within the chapters are older than one would expect, with many being from 2000 or earlier, and the median being perhaps from 2006 or 2007. Only a handful of the references are from 2010 or later. In a field that is seeing rapid progress, this is a significant matter. As a concrete example of how the dated discussion may be significant, we can see that power usage effectiveness (PUE), an industry standard recommended by the US Environmental Protection Agency under its Energy Star program, is casually mentioned in a couple of places in the book, but not discussed in detail, and well-known criticisms of PUE [2] are not mentioned at all. Likewise, the book does not seem to directly discuss cloud computing, mobile computing, and so on, though these are doubtless relevant to any contemporary discussion of “green” computing.

There are other recent anthologies [3,4,5] dealing with energy efficiency in large computing systems, which perhaps are more up to date in their analyses, and certainly do at least as good a job of considering various other energy efficiency issues, including full-system (data center-level) problems and energy efficiency in mobile devices. These, and a judicious selection of recent research papers (for example, [6,7,8]), would nicely supplement any study of this book.