### Table 9: Comparison of speedups for different environments. The speedups of the non-numerical applications corre- spond to the average of the runs for the two input sets that are not used for profiling.

2001

"... In PAGE 11: ... The best strategy for non-overlapped execution in both types of applications is AdvCoarseR, which is applied with retraction when applicable. To summarize our results, Table9 compares the speedups for different environments. Relative to P.... In PAGE 11: ...ave P.host affinity. In this case, our strategies cannot help. To put our results in perspective, Table9 shows the speedups un- der two related scenarios. The first one is an ideal system with 2 P.... In PAGE 11: ... We ran the numerical and non-numerical applications on such a machine after the Polaris and the Silicon Graphics paralleliz- ing compilers, respectively, had marked the parallel sections. The speedups on such a machine are shown in Table9 . Clearly, the re- sults are not directly comparable to our simulation results because of differences in architecture.... ..."

Cited by 15

### Table 9: Comparison of speedups for different environments. The speedups of the non-numerical applications corre- spond to the average of the runs for the two input sets that are not used for profiling.

2001

"... In PAGE 11: ... The best non-overlapped execution strategy in both types of applications is AdvCoarseR, which is applied with retraction when applicable. To summarize our results, Table9 compares the speedups for different environments. Relative to P.... In PAGE 11: ...ave P.host affinity. In this case, our strategies cannot help. To put our results in perspective, Table9 shows the speedups un- der two related scenarios. The first one is an ideal system with 2 P.... In PAGE 11: ... We ran the numerical and non-numerical applications on such a machine after the Polaris and the Silicon Graphics paralleliz- ing compilers, respectively, had marked the parallel sections. The speedups on such a machine are shown in Table9 . Clearly, the re- sults are not directly comparable to our simulation results because of differences in architecture.... ..."

Cited by 15

### Table 9: Comparison of speedups for different environments. The speedups of the non-numerical applications corre- spond to the average of the runs for the two input sets that are not used for profiling.

"... In PAGE 11: ... The best non-overlapped execution strategy in both types of applications is AdvCoarseR, which is applied with retraction when applicable. To summarize our results, Table9 compares the speedups for different environments. Relative to P.... In PAGE 11: ...ave P.host affinity. In this case, our strategies cannot help. To put our results in perspective, Table9 shows the speedups un- der two related scenarios. The first one is an ideal system with 2 P.... In PAGE 11: ... We ran the numerical and non-numerical applications on such a machine after the Polaris and the Silicon Graphics paralleliz- ing compilers, respectively, had marked the parallel sections. The speedups on such a machine are shown in Table9 . Clearly, the re- sults are not directly comparable to our simulation results because of differences in architecture.... ..."

"... In PAGE 11: ... The best non-overlapped execution strategy in both types of applications is AdvCoarseR, which is applied with retraction when applicable. To summarize our results, Table9 compares the speedups for different environments. Relative to P.... In PAGE 11: ...ave P.host affinity. In this case, our strategies cannot help. To put our results in perspective, Table9 shows the speedups un- der two related scenarios. The first one is an ideal system with 2 P.... In PAGE 11: ... We ran the numerical and non-numerical applications on such a machine after the Polaris and the Silicon Graphics paralleliz- ing compilers, respectively, had marked the parallel sections. The speedups on such a machine are shown in Table9 . Clearly, the re- sults are not directly comparable to our simulation results because of differences in architecture.... ..."

### Table 7: Percentages of typed digits and non-numerical answers.

2000

Cited by 2

### Table 1. Examples of non-numeric belief, plausibility and probability functions induced by compatibility relations. A

"... In PAGE 16: ... Let W = fw1; w2; w3; w4g and = f 1; 2; 3g. Consider the following compatibility relations: 1(w1) = f 1; 2g; 1(w2) = f 2; 3g; 1(w3) = f 3g; 1(w4) = f 1; 3g; 2(w1) = f 1g; 2(w2) = f 2g; 2(w3) = f 3g; 2(w4) = f 3g; 3(w1) = f 1g; 3(w2) = f 1g; 3(w3) = f 1; 2g; 3(w4) = f 1; 2; 3g: The induced non-numeric functions for 1 and 2 are summarized in Table1 . Note that i1 and i1 de ne a pair of non-numeric belief and plausibility functions, while i2 = i2 = i2 de nes a non-numeric probability function.... ..."