By David J. Lilja
Measuring computing device functionality units out the basic innovations utilized in studying and figuring out the functionality of computers. through the e-book, the emphasis is on sensible equipment of dimension, simulation, and analytical modeling. the writer discusses functionality metrics and offers specific assurance of the thoughts utilized in benchmark programmes. He offers intuitive motives of the foremost statistical instruments had to interpret measured functionality facts. He additionally describes the final 'design of experiments' strategy, and exhibits how the utmost volume of data could be bought for the minimal attempt. The e-book closes with a bankruptcy at the means of queueing research. Appendices directory universal chance distributions and statistical tables are incorporated, besides a word list of vital technical phrases. This practically-oriented ebook can be of significant curiosity to an individual who wishes an in depth, but intuitive, realizing of computers functionality research.
Read Online or Download Measuring computer performance : a practitioner's guide PDF
Similar systems analysis & design books
This ebook collects the examine paintings of modern researchers and practitioners within the components of research, synthesis, layout and implementation of real-time structures with functions in quite a few business fields. Their works are grouped into six elements, jointly encompassing twenty chapters. every one half is dedicated to a mainstream topic, the chapters therein constructing one of many significant elements of real-time approach concept, modeling, layout, and useful purposes.
The net has created new company parts for the picture arts undefined: details is formatted to be released either on paper and online and new kinds of cooperation with consumers and co-producers advance. the writer provides a large overview of latest companies and methods, the web has made attainable.
Measuring machine functionality units out the elemental concepts utilized in examining and realizing the functionality of desktops. during the ebook, the emphasis is on functional equipment of dimension, simulation, and analytical modeling. the writer discusses functionality metrics and offers particular assurance of the thoughts utilized in benchmark programmes.
Getting to know software program venture standards is a concise step by step consultant to constructing frameworks for the powerful administration and improvement of serious software program necessities. It re-defines the method and illustrates the necessities lifestyles cycle from elicitation to research, and specification to validation.
Extra info for Measuring computer performance : a practitioner's guide
6 Means versus ends metrics One of the most important characteristics of a performance metric is that it be reliable (characteristic 2). One of the problems with many of the metrics discussed above that makes them unreliable is that they measure what was done whether or not it was useful. What makes a performance metric reliable, however, is that it accurately and consistently measures progress towards a goal. Metrics that measure what was done, useful or not, have been called means-based metrics whereas ends-based metrics measure what is actually accomplished.
However, the sum of the MFLOPS ratings that could be calculated using these execution times is not a meaningful value. 2). Finally, if the sample data contain a few values that are not clustered together with the others, the median may give a more meaningful or intuitive indication of the central tendency of the data than does the mean. As an example, assume that we wish to determine how much memory is installed in the workstations in our laboratory. We investigate and ®nd that 25 machines contain 16 MBytes of memory, 38 machines contain 32 Mbytes, four machines contain 64 Mbytes, and one machine contains 1024 Mbytes.
The second method, however, would increase the value of f in the above calculation of the MFLOPS rating, thereby arti®cially in¯ating its value. This ¯exibility in counting the total number of ¯oating-point operations causes MFLOPS to violate characteristic 6 of a good performance metric. It is also unreliable (characteristic 2) and inconsistent (characteristic 5). 4 SPEC To standardize the de®nition of the actual result produced by a computer system in `typical' usage, several computer manufacturers banded together to form the System Performance Evaluation Cooperative (SPEC).
Measuring computer performance : a practitioner's guide by David J. Lilja