### Table 2 Performance of the fuzzy implication operator, according to different authors

1992

Cited by 3

### Table 1: Performance implications of runtime cross referencechecks, as a percentage of the original speed.

### Table 2: Performance implications of node selection using Remos in the presence of external traffic. Measurements use a synthetic program that generates significant traffic between nodes m-6 and m-8 on our IP-based testbed

1998

Cited by 53

### Table 3.3: System performance without critical clauses or non-local implications

1989

Cited by 32

### Table 3. Performance and size implications of computing 3x3 convolutions on 1024x1024 pixel images by multiple vertical bands. A single-cycle processing rate and 4 pixels overlap are assumed.

1999

"... In PAGE 20: ... One possible way of reducing this cost is to slice an image in vertical bands, and to process these images one band at time. Table3 and figure 10 show the effects of... ..."

Cited by 3

### Table 10: AMAD and AD_SE with the best defuzzification method (D1 and D6) for R- implications.

"... In PAGE 14: ... Thus, the appropriate defuzzification method is D6. Table10 shows the good performance of the combination of the former implication operators with good defuzzification methods with respect to the average of the others. Again, we found some cases where usual fuzzy implication operators perform better in combination with defuzzification methods in mode B.... ..."

### Table 4.1 Execution trace of backtracing In this example, there are two failures in the entire process. For instance after assigning [a=0, b=1] by backtracing, we perform implication and nd that the original backtrace path is no longer available, i.e., stable time of i is less than its time bound as shown in Figure 4.9 and Table 4.1. Since the initial objective (j; 1; T) is not achieved, an alternative input, h, of gate 5 is backtraced with objective (h; 1; T). Objective value 1 for line h is a control value for gate 3 and the implication value of h is unknown. According to case 2 of the backtrace algorithm, an on-path x-input is required. Unfortunately, there is no on-path x-input and backtracing fails. This triggers backtracking on the last primary input assignment (b=1), and b is re-assigned 0 to recover PODEM from the previous incorrect decision. Implication is then performed and h = (1; ?) is derived. The initial objective is not accomplished, and backtracing from j is resumed. Another failure occurs, when line h is backtraced by case 4b: no on-path control input with unknown stable time. The rest of the backtracing and implication can be discussed similarly as shown in Table 4.1. 20

1995

Cited by 1

### Table 3 Example of an individual implication matrix (Respondent R)

in Marketing

2002

"... In PAGE 18: ... As an example, the implication matrix found for one particular respondent, say respondent R, is given in Table 3. lt; Table3 about here gt; ... In PAGE 19: ... Table 3 shows which ends j are implied by which means i, and therefore, Reynolds and Gutman (1988) called such matrices implication matrices . It is immediately clear, that Table3 contains quite a few mutuals . For example, this respondent has stated that feel fine (1) is a means to personal development (2), but at the same time that personal development is a means to feel fine .... In PAGE 19: ... Aggregate analysis First we look at the aggregate implication matrix. The aggregate implication matrix is obtained by simply taking the average ratings of the implication matrices of the individual respondents, such as the one shown in Table3 . The aggregate implication matrix is presented in Table 4.... In PAGE 20: ... Aggregation could then produce entries in the implication matrix, both in cell (i,j) and cell (j,i). To refrain from the effects of mixing different respondents, an analysis at the level of the individual implication matrices, such as the one of Table3 is needed. Let g be the number of concepts making up the network.... In PAGE 21: ...Table3 ). This implies that it serves four times as a means and nine times as an end.... In PAGE 21: ... (Xij= Xji =0). For instance, Table3 shows a null dyad for quot; quot;prove yourself quot; and quot;satisfaction . The next possibility is an asymmetric dyad from i to j, i.... In PAGE 21: ...yad from i to j, i.e. if Xij is 1 and Xji is 0. An example in Table3 is the relation between quot;personal development quot; and quot;perform properly quot;. It can also be an asymmetric dyad from j to i, i.... In PAGE 21: ... Finally, between two concepts there can be a mutual dyad, where both Xij and Xji are 1. For example, in Table3 we have a mutual between quot;feel fine quot; and quot;personal development quot;. Let M be the total number of mutual dyads.... In PAGE 21: ...ine quot; and quot;personal development quot;. Let M be the total number of mutual dyads. M is the critical number for testing hypotheses regarding the symmetry of means-end relations. For the respondent presented in Table3 (respondent R) M is 10. If means- end relations are neutral toward symmetry, the expected value of M is equal to the number of mutual relations under mere chance (given the values of the marginals).... In PAGE 21: ...elations. For the respondent presented in Table 3 (respondent R) M is 10. If means- end relations are neutral toward symmetry, the expected value of M is equal to the number of mutual relations under mere chance (given the values of the marginals). As we will see later, the expected number by mere chance in the case of respondent R of Table3 , is 8.... In PAGE 23: ... The parameters g114 were estimated for all 136 respondents using the UCINET V program (Borgatti, Everett and Freeman, 1999). For respondent R, with the implication matrix of Table3 , the estimated g114 is 1.15, which confirms its tendency towards symmetry (g114 gt;0).... In PAGE 25: ...000 matrices that have the same number of indegrees and outdegrees as had been observed with these respondents. For example, let us look at the results for respondent R, of which the implication matrix is given in Table3 . For this respondent, the observed number of mutual relations, Mobserved, is 10.... ..."

### Table 1: SOR Scalability Constants (Intel iPSC/860) Notice that, as we described earlier, we do not consider possible event overlaps in the shift. We will return to the performance implications of this assumption shortly. Finally, we estimate the cost for the pipelined loop with block size B as Tpipe = (P + NB ? 1)Titeration where the single iteration time Titeration is given by Titeration = Xrecv(B) + Kunbuf0 + NP Kupdate + Kbuf0 + Ksend0: Combining all these terms, the scalability prediction for the SOR code becomes TSOR = Kj loopN2 P + (Kbuf + Ksend + Kunbuf )(N ? 2) + Xrecv(N ? 2) + P + NB ? 1 Xrecv(B) + Kunbuf0 + NP Kupdate + Kbuf0 + Ksend0

1995

Cited by 8

### Table 2 TFA content and process scoping issues Scoping issue Some implications

"... In PAGE 7: ...The scope issues of a TFA are twofold: (1) issues related to the content of the activity and (2) issues relevant to the performance (processing) and organization of the TFA activity (process). Table2 lists scope issues. Note how issues and implications interact quite heavily with each other.... ..."