### Table 2: Experimentally Determined Values of Hu n, for n apos; 300

1995

"... In PAGE 11: ... For this reason, we introduced Hu n, the average number of strategy changes per interaction, which we use as the second measure of strategy update function efficiency. Table2 shows, for each of the sixteen strategy update functions investigated, the experimentally determined values of Hu n, for n apos; 300. Table 3 shows, for each of the sixteen strategy update... ..."

Cited by 47

### Table 1: Experimentally Determined Values of Cu n, for n apos; 300

1995

"... In PAGE 5: ... Here, increasing the majority that agents use does have a significant effect on Cu n. While Table1 appears to be encouraging, we found that after testing each result against experiment (1a) for signi- ficance (using a significance level of 5%), no experiments appeared to represent a meaningful improvement (equally, no experiments appeared to represent a meaningful degrad- ation). When testing using a significance level of 10%, only experiment (4a) showed a meaningful improvement (no ex- periments showed a meaningful degradation at this level).... ..."

Cited by 47

### Table 3 lists experimentally determined values of k as a function of v and R/C, including additional data not shown in Figs. 14-16.

1970

### Table 2. Experimentally determined machine parameters for seth.

"... In PAGE 21: ...eigenvalues per on-diagonal processor are kept flxed going from a problem instance (n; k; p) to (2n; 2k; 4p), we assume ~ t(p) a (n; k) = ~ t(4p) a (2n; 2k): (15) From this assumption and the available data for ~ t(1) a derived from Table 1, we compute the fii-parameters from a least squares flt, see Table2 . With these parameters and flxed values on p and/or n, the model predicts that ~ t(p) a (n; k) decreases for an increasing value of k.... ..."

### Table 1 The peak VDR of 1 ms bipolar pulses of various shapes required for VDR of 470 V/cm to be exceeded without interruption for 200 and for 250 As, compared to the experimentally determined values of P50% 470 V/cm exceeded

"... In PAGE 5: ... To an extent, this argu- ment can also be supported quantitatively. In Table1 , the measured values of P50% for triangular, sine, and rectan- gular bipolar pulses are compared to the theoretical values of the peak VDR required for the critical VDR of 470 V/ cm to be exceeded without interruption for 200 As and for 250 As (depicted in Fig. 5).... ..."

### Table 2 Comparison between calculated and experimentally determined cycle statistics for the cycle

"... In PAGE 10: ... The standard deviations were roughly estimated from available test data for various kinds of steel. Table2 summarizes the calculated and experimentally obtained statistics for the number of cycles until the onset of instable crack growth. The results are in good agreement, especially the trend in the mean value and the standard deviations for the... ..."

### Table 13: Experimentally determined rectangular cuto parameters, when = 1 and = 0.

1996

"... In PAGE 55: ...6 percent reduction in computing time. Table13 summarizes, for each machine, the values obtained for the parameters m, k, and n, which describe the rectangular condition (18). These were computed using the method described in Section 5.... In PAGE 56: ... (16) = Simple Criterion from Square Cuto , (17) = Theoretically Scaled Criterion, (20) = New Rectangular Criterion. The data in Table13 shows that the performance of DGEMM is not symmetric in the matrix dimensions. Moreover, the asymmetry varies from machine to machine.... ..."

Cited by 15

### Table 2: Experimentally determined square cuto s on various machines.

1996

"... In PAGE 17: ...hose any of the points in the range. We opted to choose the cuto to be = 199, i.e., use DGEMM if m 199, since Strassen is almost always better than DGEMM for larger values, and when it is slower it is so by a very small amount. Table2 shows the results of performing this experiment for our test machines. The cuto s in Table 2 show the size at which our Strassen implementation becomes more e cient for square matrices.... In PAGE 17: ... Table 2 shows the results of performing this experiment for our test machines. The cuto s in Table2 show the size at which our Strassen implementation becomes more e cient for square matrices. The modest size of the cuto s show that using our Strassen code instead of standard matrix multiplication is advantageous in many applications.... In PAGE 20: ...3 Table 5: Times for DGEMM and DGEFMM for di erent numbers of recursions on all three machines with = 1=3 and = 1=4. is the square cuto value given in Table2 for each machine. 1800 for the RS/6000 and CRAY C90 and 1350 for the CRAY T3D.... ..."

Cited by 15

### TABLE 3 Structures of synthesized compounds All values are mean H11006 S.E.M. from three experimental determinations. Compound Structure

2001