**Measuring Computer Performance**

A Practitioner's Guide

A Practitioner's Guide

**David J. Lilja
Cambridge University Press, New York, NY, 2000**

**ISBN 0-521-64105-5**

**Errata**

The following corrections should be made to the book.

- p. 31 -- The example at the bottom of this page should be replaced with this file: p31-correction.pdf.
- p. 35 -- In the first sentence immediately following Equation (3.12), "w_{i} = 0.5" should be "w_{1} = 0.5."
- p. 53 -- In Equation (4.22), the term "z_(1-alpha/2)" should be replaced with the term "t_(1-alpha/2 ; n-1)." Then add the following sentence as the last sentence in this example, "Note that in Equation (4.22) we have replaced the z value in Equation (4.21) with the corresponding value from the t table since only n=8 measurements were used to compute the sample mean and the sample standard deviation."
- p. 54 -- The first paragraph should read, "A binomial distribution with parameters p and n has a mean of pn and a variance of p(1-p)n. If np >= 10, we can approximate the binomial distribution with a Gaussian distribution with mean p~n and variance p~(1-p~)n. Applying ..." (The notation p~ is supposed to be the letter p with a bar over the top.)
- p. 64 -- There is a small inconsistency between the explanation on p. 64 and Table 5.1 on p. 65. In the second sentence on p. 64, the difference values are computed as (di = ai - bi). However, in Table 5.1, the difference values are computed as (di = bi - ai). Either way is fine -- the order of subtraction changes only the sign of the difference (di) values. The final conclusion about the comparison will remain the same.
- p. 69 -- The last sentence of the first paragraph in Section 5.1.3 should read, "From the characteristics ... mean for system i is simply p_i n_i with a variance of p_i(1-p_i)n_i." (The notation p_i means the letter p with the subscript i.)
- p. 69 -- The first sentence of the second paragraph in Section 5.1.3 should read, "If both ... with mean p_i n_i and variance p_i(1-p_i)n_i."
- p. 70 -- Above Equation (5.21) the calculation for p should be, "p = p1 - p2 = 0.0250 ..."
- p. 70 -- Equations (5.22) and (5.23) show the calculation for a 95% confidence, not a 90% confidence interval as stated in the text above and below these equations.
- p. 72 -- The first summation in Equation (5.25) should be the sum from j=1 to k.
- p. 72 -- The entry y_kn in Table 5.3 should be y_nk.
- p. 77 -- In Table 5.6, the subscript for the F value from the table should have 0.95 in place of 0.05.
- p. 79 -- In Equation (5.52), the summation from j=1 to 3 should be omitted from the second radical.
- p. 91 -- In the next-to-last sentence of the second paragraph, "Thus, we have ... one tick while are measuring the event." should be "Thus, we have ... one tick while measuring the event."
- p. 101 -- In the first sentence of the second paragraph of Section 6.4.2.1, "interdeterminate" should be "indeterminate".
- p. 131 -- In the last paragraph of Section 7.3.3.1, "...for both TPC and UDP network..." should be "...for both TCP and UDP network..."
- p. 144 -- The denominator in Equation (8.32) should be "square root [7(86,611,800)]". The remainder of the expression is correct.
- p. 148 -- In Equation (8.41), "y" should be "y_i".
- p. 154 -- Near the end of the first paragraph, "t_1-0.95;4 = 2.132" should be "t_0.95;4 = 2.132".
- p. 154 -- In the last paragraph, y'_10 should be y_10 (that is, remove the prime mark).
- p. 163 -- In Equation (9.13), the last term, y_bar, should be added instead of subtracted.
- p. 164-165 -- In Equations (9.29), (9.30), and (9.31), the subscript for F on the right-hand side should have (1 - alpha) instead of alpha.
- p. 165 -- In Table 9.3, the subscripts for all of the F values from the table should have (1 - alpha) instead of alpha.
- p. 168 -- -- In Table 9.5, the subscript for the F values from the table should have 0.95 in place of 0.05.
- p. 169 -- In step #5, the subscript for F should have (1 - alpha) instead of alpha.
- p. 174 -- In Table 9.7, the subscripts for all of the F values from the table should have (1 - alpha) instead of alpha.
- p. 175 -- The sentence above Equation (9.62) should be, "For example, from Table 9.11, we find that the contrast for the interaction between factors A and C is ..."
- p. 211 -- In the third paragraph, the sentence that ends with, "... where n is the total of points plotted." should be "... where n is the total number of random values generated."
- p. 219 -- In Figure 11.2, the arrow pointing into the server of Disk 1 from the right-hand side should be removed. The arrow should be pointing out of this server.
- p. 230 -- Equation (11.16) should be P_0 = 1 / (rho^0 + rho^1 + rho^2 + ...) = 1 / sum(rho^n) where the summation still runs from n=0 to infinity.
- p. 243 -- In the last line on this page, the term "ln(1-p)" should be "ln(p)". This term should be corrected in both of the places it appears in this line.