### Table 1: The Arithmetic Cost of One Loop Iteration

1994

"... In PAGE 11: ... This will certainly limit the problem size we will be able to solve. Table1 tabulates the arithmetic cost of one loop of the inverse free iteration versus the Newton iteration #283.8#29 and #283.... In PAGE 11: ...teration #283.8#29 and #283.9#29 for the real ordinary and real generalized spectral divide and conquer problems, respectively. From Table1 , we see that for the standard spectral divide and conquer problem, the one loop of the inverse free iteration does about 6.7 times more arithmetic than the one... ..."

Cited by 51

### Table 7. Multiplication in first iteration (a) , other loop iterations (b)

2001

### TABLE I AVERAGE LOOP ITERATIONS NOT CAPTURED BY THE HOTSPOT FOR VARIOUS APPLICATION TYPES

### Table I summarizes the application area specific terminol- ogy used to describe data array elements, loop iterations, array distributions and loop iteration partition.

1995

Cited by 61

### Table 4: Max and min iterations in loops

"... In PAGE 5: ... Loops with small max and/or min number of iterations are not suitable to apply software pipelining technique. Table4 shows that the majority of the innermost loops can benefit from the software pipelining. However, 45% of the innermost loops are with small min iterations.... ..."

### TABLE I ITERATIONS OF THE LOOP DETECTION ALGORITHM

### Table 7.1: CEM Iteration Loop

### Table 3. Average Number of Main Loop Iterations Input Size in Decimal Digits Algorithm 100 250 500 1000

1998

"... In PAGE 11: ...Finally, we present the average number of main loop iterations performed by each algorithm in Table3 . Note that this data is independent of the particular implementation.... ..."

Cited by 1

### Table 3: Maximum speedups for loops with no cross-iteration dependences.

1994

"... In PAGE 35: ...2. Loops with No Cross-Iteration Dependences For loops with no cross-iteration dependences, Table3 shows that the parallelism that can be exploited usually is limited by the resource constraints since the number of independent iterations typically is much greater than the number of processors or functional units. Both trace scheduling and software pipelining can be used to schedule these loops to execute on a fine-grained processor.... ..."

Cited by 31