### Table 9: Large-step optimization: small-steps: simulated annealing; large-steps: 2mach/BB.

1995

"... In PAGE 23: ... Overall the proposed methods to perform the large-step, the best results were obtained for the large-step method 2mach/BB and 2mach/EL. In Table9 we present the results when we applied the large-step optimization methods using the simulated annealing method and the 2mach/BB, where in this case we run just 10 iterations of the method but we let run the simulated annealing method for 1000 iterations. Comparing these results with the respective ones in Tables 6, 7 and 8, we can observe that by running fewer iterations of the large-step optimizations methods and more of the simulated annealing method we did not get better results.... ..."

Cited by 5

### TABLE V NUMBER OF STEPS WITH THE LARGE DISTURBANCES

2002

Cited by 20

### Table 4. Cost of the matrix step for the large matrix

2002

"... In PAGE 22: ...ake high bandwidth inter-wafer connections (cf. 5.3). Using the formulas given in the previous section, we obtain the costs in Table4 for the custom and FPGA implementations, for various parameter choices. The third line shows that here too, significant parallelism can be attained at very little cost (88% of the wafer area is occupied by DRAM storing the input).... ..."

Cited by 7

### Table 4. Cost of the matrix step for the large matrix

2002

Cited by 7

### TABLEII The number of candidate and large itemsets at each step

### Table 3: Total Data Exchange for Single Iteration Step of Large Simulation

1994

"... In PAGE 7: ... The last column of Table 2 shows the memory requirements of the executable code; these numbers show that the code is far from reaching the local memory limit of 128 Mb. A performance summary for the communication part of the algorithm is given in Table3 . The presented values for the message volume and the message size are calculated with the formula presented in section 2.... ..."

Cited by 2

### Table 2: Total Data Exchange for Single Iteration Step of Large Simulation

1993

"... In PAGE 5: ... The last column of Table 1 shows the memory requirements of the executable code; these numbers show that the code is far from reaching the local memory limit of 128 Mb. A performance sum- mary for the communication part of the algorithm is given in Table2 . Here, the reported bandwidth rep- resents a measured value which includes the message startup costs.... ..."

Cited by 2

### TABLE VI ONE STEP CONTROL COST WITH THE LARGE DISTURBANCES (AVERAGED OVER 100 STEPS)

2002

Cited by 20