### Table 2: E ect of DFA

1994

"... In PAGE 17: ... DFA also identi es operations that do not change the value of a relation (such as clearing an empty relation). The results of running the DFA optimizations on the benchmark suite (see Section 7) is shown in Table2 . The absolute error in the code speed improvements is one percentage point.... ..."

Cited by 12

### Table 1: DFA exact results

2004

"... In PAGE 6: ...5%. These results can be verified against the exact results obtained using brute force algorithms in Table1 and Table 2. 3 Sampling and prediction The formulae established for n = 2, 3 offer the exact values of PD (2, p), PD (3, p), PN (2, p), PN (3, p) , for any p = 2, 3, .... In PAGE 10: ... Using the first 13 values from Table 10, we obtain the results for the NTPS predictor g in Table 11. Using the first 13 values and the supplementary 11 values from Table1 0, we get the results in Table 12. As we can see, the difference between the percentage obtained by generating samples and the one computed using the NTPS predictor is again less than 1.... In PAGE 11: ... The results are presented in Table 7 and Table 10. Note that the statistics is very close to those in Table1 and Table 2, respectively. For most of them, the difference is less than 1%.... In PAGE 18: ...Table1 0: The number of NFAs accepting the empty language using randomly generated samples No. n p Total number of NFAs NFAs accepting non-empty lan- guages NFAs accepting non-empty lan- guages Non-empty per- cent 1 3 2 22500 18906 3594 84.... In PAGE 19: ...Table1 1: Comparative results for NFA NTPS predictor using 13 knots Number of NFAs Percent of NFAs NTPS No n p Total number of NFAs tested recognizing at least one word recognizing at least one word g(n, p) estimated empty percent Precision 1 6 8 22500 22161 98.490% 99.... In PAGE 19: ...000 0.000 0 Table1 2: Comparative results for NFA NTPS predictor using 24 knots Number of NFAs Percent of NFAs NTPS No n p Total number of NFAs tested recognizing at least one word recognizing at least one word g(n, p) estimated empty percent Precision 1 4 7 22500 21121 93.870% 92.... ..."

### Table 2. DFA exact results

"... In PAGE 10: ... Obviously, the number of automata grows exponentially with a3 and a7 (a3 a7 a1 a33 a40a67 a7 for DFAs and a67 a7 a26 a1 a5a14a7 for NFAs). The method was used for the values presented in Table2 and Table 4. One can see... In PAGE 11: ... The results are presented in Table 7 and Table 10. Note that the statistics is very close to those in Table2 and Table 4, respectively. For most of them, the difference is less than 1%.... In PAGE 11: ... The programs were run on a PC Pentium 4 1.6A with 64 MB memory, for more than 1 week to obtain the results in Table2 and Table 4. The size of memory was not important since every time we store only one automaton and no swapping of data is used.... ..."

### Table 4: Size of clients of the DFA framework.

"... In PAGE 5: ... for a wide range of optimizations. Table4 lists the sizes of the parts of the optimizer that use the DFA framework.... ..."

### Table 6. Test Coverage Using DFA Last

2003

"... In PAGE 5: ... To investigate the utility of applying DFA only once, when coverage is already very high, we applied more than twice as many random patterns, followed by a sin- gle round of single clock cycle DFA tests. Table6 shows the coverage results. There is no increase in the cover- age since the DFA rule-based method is used too late.... ..."

Cited by 1

### Table B.2: Methods of DFA Object

2003

Cited by 1

### Table 2. Test Coverage Using DFA

2003

Cited by 1

### Table 1: DFA applied to coloured noises.

"... In PAGE 5: ...qual 0.5, 1 and 1.5 respectively. We checked this assertion both numerically and algebraically [1]. Table1 summarizes these results. 1The random process X[t] is said to be wide sense stationary, or covariance stationary, when E(X[t]) is nite and independant of t, and R[s] = E(X[t]X[t+s]) is nite and does not depend on t.... ..."

### Table 3: NJ-NNJ comparisons for DFA, FAIN and FAIR

in non-native

"... In PAGE 3: ...Table 3: NJ-NNJ comparisons for DFA, FAIN and FAIR Table3 results indicated that DFA mean scores exhibited by the two groups of judges were very similar. Unsurprisingly, the T-test yielded no statistical significance for DFA.... ..."