### Table 1. Comparison between conventional and improved reduce bit computation

"... In PAGE 8: ... The critical path of the improved reduce bit computation is significantly shorter since we can sum up the remaining terms in logarithmic time by using a binary tree structure of XOR gates. For d = 16, the critical path in the reduction circuit consists of 8 gates (2 ANDs and 7 XORs), whereas the critical path of the conventional reduce bit computation consists of 30 gates (15 ANDs and 15 XORs, see Table1 ). This results in a substantial performance gain, even when we take into account that the delay of an XOR gate is typically higher than that of an AND gate.... In PAGE 8: ...g., 774 gates instead of 240 when d = 16, see Table1 ). However, it must be con- sidered that the area of a digit-serial multiplier is primarily determined by the logic circuits for partial-product addition and reduction.... ..."

### Table 3 CPU time for 64 bit real computation

"... In PAGE 10: ... For multiplying N #02 N matrices the memory requirements for Strassen implementations are Implementation #0C 6 =0orA,B #0C= 0 and A, B overlaps with C do not overlap with C gemmw 1:67N 2 0:67N 2 Cray gemms 2:34N 2 2:34N 2 IBM ESSL gemms#7Breal not possible 1:40N 2 IBM ESSL gemms#7Bcomplex not possible 1:70N 2 The IBM ESSL routine gemms assumes #0B =1,#0C= 0 in #282#29, and no overlapping of A, B,orC. Table3 contains the results of the example problem for 64 bit real data. The highlights of the table as follows: Cray Competing against the hand tuned classical parallel matrix#7Bmatrix multiplica- tion supplied as part of the Cray Math and Scienti#0Cc Library turned out to be surprisingly di#0Ecult.... ..."

### Table 1. Comparison of computational complexity (bit operations)

1988

"... In PAGE 8: ... Therefore, the complexity of decryption is O(n3). Table1 shows the comparison of computational complexity. Table 1.... ..."

Cited by 12

### Table 4 CPU time for 64 bit complex computation

### Table 2: De nition of indexing and tagging schemes. The variable i denotes the number of index bits to compute and the variable t denotes the number of tag bits.

2005

### Table 2. Computational load of handshakes in bit multiplications.

"... In PAGE 10: ... The total numbers in bit multiplication to compute point multiplication are 36M (M denotes 106 ) . Table2 evaluates and compares the computational loads for the protocol in comparison with the original hand- shake. When compared with the original handshake, the proposed handshake included an one additional scalar multiplication and one off-line computation in each key generation part.... ..."

### Table 3: Computing the total number of bits used

2004

Cited by 27

### TABLE X OVERALL STATISTICS OF THE DIFFERENCE BETWEEN THE ADSL BIT RATE COMPUTED WITH NEXT SUMMATION METHODS AND THE ADSL BIT RATE COMPUTED BY THE MONTE CARLO METHOD ACROSS ALL SIMULATIONS

2002

Cited by 6

### TABLE X OVERALL STATISTICS OF THE DIFFERENCE BETWEEN THE ADSL BIT RATE COMPUTED WITH NEXT SUMMATION METHODS AND THE ADSL BIT RATE COMPUTED BY THE MONTE CARLO METHOD ACROSS ALL SIMULATIONS

2002

Cited by 6