Previous | Table of Contents | Next |
Gibson Mix is described in Gibson (1970). Lucas, Jr. (1971) presents one of the earliest survey of test workloads. The articles by Hansen et al. (1982), Patterson (1982), and Levy and Clark (1982) argue the proper (or improper) use of kernels in comparing processors. Gilbreath (1981) describes the Sieve kernel and provides its listings in FORTRAN, C, PL/I, BASIC, FOURTH, and COBOL. Wright, Kohler, and Zahavi (1989) describe one interpreation of the debit-credit benchmark.
Surveys of workload characterization techniques can be found in Ferrari (1972), Agrawala et al. (1976), and Oed and Mertens (1981). The data presented in Tables 6.1 and 6.2 are from Jain and Turner (1982). Sreenivasan and Kleinman (1974) describe a case study of the the multiparameter histogram technique. Application of principal-component analysis techniques can be found in Serazzi (1981, 1985). Clustering analysis has been applied in numerous case studies. See, for example, Agrawala and Mohr (1977), Mamrak and Amer (1977), Artis (1978), Wight (1981), and Calzarossa. and Ferrari (1986). For description of various clustering techniques see Everitt (1974), which also has a good discussion of problems with clustering.
Nutt (1975, 1979) and Svobodova (1976) present some of the earliest discussions on H/W and S/W monitors. Pinkerton (1977) describes common forms and mechanisms of software monitors. Ramage (1977) discusses various issues of software monitor design. Hughes (1980) presents the design of a hybrid monitor.
For a sample description of a commercially available hardware monitor, see COMTEN (1977). Howard (1983) presents a description of several commercially available performance analysis tools. Leavitt (1986b) has a more recent listing of commercially available monitors. EDP Performance Review publishes an annual issue listing all commercially available performance analysis tools. Plattner and Nievergelt (1981) describe performance measurements of programs.
Bouhana (1979) describes the issues related to use of accounting logs. The case study on accounting log usage is from Jain and Turner (1982).
Houghton, Jr. (1982) describes and compares four different ways to instrument a program. Power (1983) presents a survey of program execution monitoring and has an extensive bibiliography. Bentley (1982) discusses several techniques to improve the efficiency of programs.
The discussion on distributed monitoring is based on the features available in several commercially available network monitors. Joyce et al. (1987) present a case study of a distributed-monitor design.
For a survey of capacity planning issues, see the report prepared by Data-metrics (1983) for the National Bureau of Standards. The report has an extensive bibliography. The CMG conference proceedings and EDP performance review regularly have articles on this topic. See also Strauss (1981). Leavitt (1986a) provides a list of commercial capacity planning products and vendor addresses.
For surveys of RTEs see Watkins and Abrams (1977a, 1977b) and Spiegel (1980). Figure 9.5 is based on Proppe and Wallack (1982). McGalliard and Theraja (1988) and Trehan (1978) discuss the problems of remote workstation emulation and describe the keyboard monitor emulation approach.
Many of the benchmarking mistakes discussed in this part are from Goldstein (1988), which has an interesting discussion of other issues related specifically to storage benchmarking.
Advice on preparing good graphic charts can be found in the books by Schmid and Schmid (1979), Lefferts (1981), Schmid (1983), Holmes (1984), and White (1984). Articles by Powell (1988), Finehirsh (1985), and MacKinnon (1987) also present additional information on graphic presentations.
The idea of Kiviat graphs was first proposed by Kolence (1973a). It immediately became a popular topic of discussion in Performance Evaluation Review for the next three years. Kolence and Kiviat (1973b) describe several common shapes. The paper by Morris (1974) gives an excellent history and classification of graphs. Merrill (1974) proposed a numerical figure of merit to compare Kiviat graphs, and a number of papers including those by Bahr (1975), Merrill (1975), Stevens (1975), and Calcagni (1976) subsequently proposed other alternative figures of merits. Schumacher charts are described in more detail in Morris and Roth (1982).
The data for Case Study 11.2 was adopted from Patterson and Sequin (1982). We have deleted one row and one column containing missing values.
The book by Ferrari, Serazzi, and Zeigner (1983) has an excellent treatment of many topics related to measurement and modeling of computer systems. Other books discussing performance measurements are Ferrari (1978), Barnes (1979), and McKerrow (1987).
Previous | Table of Contents | Next |