### Table 6 End-of-Period Ranking of Percentage Change of RCA# EANIEs ASEAN LA

2002

"... In PAGE 17: ...15 In addition, we work out the changes in the export pattern obtained from the RCA averages between the two periods. Table6 shows the end-of-period ranking of sectors by the percentage change in the absolute value of RCA. We observe both the gain and loss relationships in the RCAs of the three country groupings.... ..."

### Table 4.3: The functional unit and memory access la- tencies (in cycles) used in our simulator.

### Table A2. Business cycle correlation with Germany in each period.

### Table 8: E ect of Number of Periods in the Life Cycle.

### Table 3.3 Clock cycle period for all designs (nsec)

2002

### Table 5.1: Results of the test scenario in ms per period of 100 period cycles, period 1500ms, wcet 100ms

in Contents

2005

### Table 4. Task Set Characteristics: Benchmark IDs, Phases[cycles] and Periods [cycles]

"... In PAGE 6: ....6, 0.7 and 0.8 are shown in Table4 . The first column shows the tasks used in each task set.... ..."

### Table 8: Different combinations of LA, LB and N initial with lim. scan

"... In PAGE 6: ... This number can be reduced by using larger values of LA, LB and/or N. To demonstrate this point, we show in Table8 the results obtained for some of the circuits of Table 6 by using several different combinations of LA, LB and N. The first row for every circuit in Table 8 is always the same as that in Table 6.... In PAGE 6: ... To demonstrate this point, we show in Table 8 the results obtained for some of the circuits of Table 6 by using several different combinations of LA, LB and N. The first row for every circuit in Table8 is always the same as that in Table 6. Table 8: Different combinations of LA, LB and N initial with lim.... In PAGE 6: ...a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 a3 det - number of detected faults cycles - number of clock cycles required for test application app - number of test sets applied with limited scan operations ls - average number of time units where limited scan operations are performed From Table8 it can be seen that it is possible to reduce the number of applications of the test set (and the number of (I ,D1) pairs) by using larger values of LA, LB and/or N; how- ever, in several cases, the number of clock cycles required for test application increases as well. 5.... ..."

### T able 8. C ompar ing Supe rs c alar SD F Per for manc e (FFT ) Super sca la r SDF Super sca la r SDF Super sca la r SDF Super sca la r SDF

"... In PAGE 14: ... LU s + #FP ALUs for Supers ca la r). The figure s hows the e xe cution times for 150*150 ma trix multiplica tion. Scalability of SDF (Matrix) 0 10000000 20000000 30000000 40000000 50000000 60000000 2+1 2+2 3+2 3+3 4+3 4+4 5+4 5+5 Number of Functional Units Execution Cycles In Order Out of Order SDF Figur e 6. Scalability of SDF over Supe rs c alar (M atr ix M ultiply) In the next Table ( Table8 ) we show the res ults for FFT. In this c as e, SD F outpe rforms O ut-of-O rde r Supers ca la r for data s iz es grea ter that 256 for a ll ma chine c onfigura tions .... ..."