A new class of irreducible pentanomials for polynomial based multipliers in binary fields

Transcription

1 Noname manuscript No. (will be inserted by the editor) A new class of irreducible pentanomials for polynomial based multipliers in binary fields Gustavo Banegas Ricardo Custódio Daniel Panario the date of receipt and acceptance should be inserted later Abstract We introduce a new class of irreducible pentanomials over F 2 of the form f(x) = x 2b+c + x b+c + x b + x c + 1. Let m = 2b + c and use f to define the finite field extension of degree m. We give the exact number of operations required for computing the reduction modulo f. We also provide a multiplier based on Karatsuba algorithm in F 2 [x] combined with our reduction process. We give the total cost of the multiplier and found that the bit-parallel multiplier defined by this new class of polynomials has improved XOR and AND complexity. Our multiplier has comparable time delay when compared to other multipliers based on Karatsuba algorithm. Keywords irreducible pentanomials polynomial multiplication modular reduction finite fields 1 Introduction Finite field extensions F 2 m of the binary field F 2 play a central role in many engineering applications and areas such as cryptography. Elements in these extensions are commonly represented using polynomial or normal bases. We center in this paper on polynomial bases for bit-parallel multipliers. When using polynomial bases, since F 2 m = F2 [x]/(f) for an irreducible polynomial f over F 2 of degree m, we write elements in F 2 m as polynomials over F 2 of degree smaller than m. When multiplying with elements in F 2 m, a polynomial of degree up to 2m 2 may arise. In this case, a division by f follows and the remainder of this division is the resulting element in F 2 m. Gustavo Banegas ( gustavo@cryptme.in) at Technische Universiteit Eindhoven, Eindhoven, The Netherlands Ricardo Custódio ( ricardo.custodio@ufsc.br) at Universidade Federal de Santa Catarina, Florianópolis, Brazil Daniel Panario ( daniel@math.carleton.ca) at Carleton University, Ottawa, Canada.

2 2 Gustavo Banegas et al. Mathematically, any irreducible polynomial can be used to define the extension. In practice, however, the choice of the irreducible f is crucial for fast and efficient field multiplication. The complexity of hardware circuits for finite field arithmetic in F 2 m is related to the amount of space and the time delay needed to perform the operations. Normally, the number of XOR and AND gates is a good estimation of the space complexity. The time complexity is the delay due to the use of these gates. Several special types of polynomials have been considered before, including polynomials with few nonzero terms like trinomials and pentanomials (three and five nonzero terms, respectively), equally spaced polynomials, all-one polynomials [, 9, 15], and other special families of irreducible polynomials [20]. In general, trinomials are preferred, but for degrees where there are no irreducible trinomials, pentanomials are considered. The analysis of the complexity using trinomials is known [19]. However, there is no general complexity analysis of a generic pentanomial in the literature. Previous results (see [3] for details) have focus on special classes of pentanomials, including: x m + x b+1 + x b + x b 1 + 1, where 2 b m/2 1 [6, 8, 1, 16, 21]; x m + x b+1 + x b + x + 1, where 1 < b < m 1 [6, 7, 1 16, 21]; x m + x m c + x b + x c + 1, where 1 c < b < m c [2]; x m + x a + x b + x c + 1, where 1 c < b < a m/2 [15]; x m + x m s + x m 2s + x m 3s + 1, where (m 1)/8 s (m 1)/3 [15]; x c + x 3c + x 2c + x c + 1, where c = 5 i and i 0 [, 5]. Like our family, these previous families focus on bit operations, i.e., operations that uses only AND and exclusive-or (XOR) gates. In the literature is possible to find studies that use computer words to perform the operations [13, 17] but this is not the focus of our work. 1.1 Contributions of this paper In this paper, we introduce a new class of irreducible pentanomials with the following format: f(x) = x 2b+c + x b+c + x b + x c + 1. (1) We compare our pentanomial with the first two families from the list above. The reason to choose these two family is that [1] presents a multiplier considering these families with complexity 25% smaller than the other existing works in the literature using quadratic algorithms. Since our multiplier is based on Karatsuba s algorithm, we also compare our method with Karatsuba type algorithms. An important reference for previously used polynomials and their complexities is the recent survey on bit-parallel multipliers by Fan and Hasan [3]. Moreover, we observe that all finite fields results used in this paper can be

3 A new class of irreducible pentanomials 3 found in the classical textbook by Lidl and Niederreiter [10]; see [11] for recent research in finite fields. We prove that the complexity of the reduction deps on the exponents b and c of the pentanomial. A consequence of our result is that for a given degree m = 2b + c, for any positive integers b > c, all irreducible polynomials in our family have the same space and time complexity. We provide the exact number of XORs and gate delay required for the reduction of a polynomial of degree 2m 2 by our pentanomials. The number of XORs needed is 3m 2 = 6b+3c 2 when b 2c; for b = 2c this number is 12 5 m 1 = 12c 1. We also show that AND gates are not required in the reduction process. For comparison purposes with other pentanomials proposed in the literature, since the operation considered in those papers is the product of elements in F 2 m, we also consider the number of ANDs and XORs used in the multiplication prior to the reduction. In the literature, one can find works that use the standard product or use some more efficient method of multiplication, such as Karatsuba, and then add the complexity of the reduction. However, most of the works perform an analysis of the multiplier as a single block, that is, consider multiplication and reduction as a single process. In this paper, we use a Karatsuba multiplier combined with our fast reduction method. The total cost is then Cm log m 2 or Cm log m 1, deping on b 2c or b = 2c, respectively. The constant C of the Karatsuba multiplier deps on the implementation. In our experiments, C is strictly less than 6 for all practical degrees, up to degrees For the reduction, we give algorithms that achieve the above number of operations using any irreducible pentanomial in our family. We compare the complexity of the Karatsuba multiplier with our reduction with the method proposed by Park et. al [1], as well as, with Karatsuba variants given in [3]. 1.2 Structure of the paper The structure of this paper is as follows. In Section 2 we give the number of required reduction steps when using a pentanomial f from our family. We show that for our pentanomials this number is 2 or 3. This fact is crucial since such a low number of required reduction steps of our family allows for not only an exact count of the XOR operations but also for a reduced time delay. Our strategy for that consists in describing the reduction process throughout equations, cleaning the redundant operations and presenting the final optimized algorithm. Section 3 provides the first component of our strategy. In this section, we simply reduce a polynomial of degree at most or exactly 2m 2 to a polynomial of degree smaller than m. The second component of our strategy is more delicate and it allows us to derive the exact number of operations involved when our pentanomial f is used to define F 2 m. Sections and 5 provide those analyses for the cases when two and three steps of reduction are needed, that is, when c = 1 and c > 1, respectively. We give algorithms and exact estimates for the space and time complexities in those cases. In Section 6, we

4 Gustavo Banegas et al. describe a Karatsuba multiplier implementation combined with our reduction. Based on our implementation, we show that the number of XOR and AND gates is better than the known space complexity in the literature. On the other hand, the time complexity (delay) in our implementation is worse than quadratic methods but comparable with Karatsuba implementations. Hence, our multiplier would be preferable in situations where space complexity and saving energy are more relevant than time complexity. We demonstrate that our family contains many polynomials, including degrees where pentanomials are suggested by NIST. Conclusions are given in Section 7. 2 The number of required reductions When operating with two elements in F 2 m, represented by polynomials, we obtain a polynomial of degree at most 2m 2. In order to obtain the corresponding element in F 2 m, a further division with remainder by an irreducible polynomial f of degree m is required. We can see this reduction as a process to bring the coefficient in interval [2m 2, m] to a position less than m. This is done in steps. In each step, the coefficients in interval [2m 2, m] of the polynomial is substituted by the equivalent bits following the congruence x m x a + x b + x c + 1. Once the coefficient in position 2m 2 is brought to a position less than m, the reduction is completed. In this section, we carefully look into the number of steps needed to reduce the polynomial by our polynomial f given in Equation (1). The most important result of this section is that we need at most 3 steps of this reduction process using our polynomials. This information is used in the next sections to give the exact number of operations when the irreducible pentanomial given in Equation (1) is employed. This computation was possible because our family has a small number of required reduction steps. Let D 0 (x) = 2m 2 d i x i be a polynomial over F 2. We want to compute D red, the remainder of the division of D 0 by f, where f has the form f(x) = x 2b+c + x b+c + x b + x c + 1 with 2b + c = m and b > c > 0. The maximum number k a of reduction steps for a pentanomial x m + x a + x b + x c + 1 in terms of the exponent a is given by Sunar and Koç [18] m 2 k a = + 1. m a In our case m = 2b + c and a = b + c, thus { 2b + c 2 c 2 2 if c = 1, k b+c = + 1 = + 3 = 2b + c b c b 3 if c > 1. Using the same method as in [18], we can derive the number of steps required associated to the exponents b and c. These numbers are needed in Section 3. We get k b = 2b + c = 2b + c b (2) b = 2, (3) b + c

5 A new class of irreducible pentanomials 5 and k c = 2b + c = 2b + c c c 2 2b { 1 if c = 1, + 2 = 2 if c > 1. () Thus, the reduction process for our family of pentanomials involves at most three steps. This is a special property that our family enjoys. The general process for the reduction proposed in this paper is given in the next section. The special case c = 1, that is when our polynomials have the form f(x) = x 2b+1 + x b+1 + x b + x + 1, requires two steps. This subfamily is treated in detail in Section. The general case of our family f(x) = x 2b+c + x b+c + x b + x c + 1 for c > 1 involves three steps and is treated in Section 5. 3 The general reduction process The general process that we follow to get the original polynomial D 0 reduced to a polynomial of degree smaller than m is depicted in Figure 1. Without loss of generality, we consider the polynomial to be reduced as always having degree 2m 2. Indeed, the cost to determine the degree of the polynomial to be reduced is equivalent to checking if the leading coefficent is zero. The polynomial D 0 to be reduced is split into two parts: A 0 is the piece of the original polynomial with degree at least m and hence that requires extra work, while B 0 is formed by the terms of D 0 with exponents smaller than m and so that it does not require to be reduced. Dividing the leading term of A 0 by f with remainder we obtain D 1. In the same way as before, we split D 1 in two parts A 1 and B 1 and repeat the process obtaining the tree of Figure 1. D 0 = A 0 + B 0 D 1 = A 1 + B 1 D 2 = A 2 + B 2 D 3 = A 3 + B 3 Fig. 1 Tree representing the general reduction strategy.

6 6 Gustavo Banegas et al. 3.1 Determining A 0 and B 0 We trivially have D 0 (x) = A 0 (x) + B 0 (x) = 2m 2 i=m m 1 d i x i + d i x i, and hence A 0 = 2m 2 i=m d i x i and B 0 = m 1 d i x i. (5) 3.2 Determining A 1 and B 1 Using for clarity the generic form of a pentanomial over F 2, f(x) = x m + x a + x b + x c + 1, dividing the leading term of A 0 by f and taking the remainder, we get D 1 = m 2 m 2 d i+m x i+a + m 2 d i+m x i+b + m 2 d i+m x i+c + d i+m x i. Separating the already reduced part of D 1 from the piece of D 1 that still requires more work, we obtain and B 1 = A 1 = m 1 i=a m+a 2 i=m m+b 2 d i+(m a) x i + m 1 d i+(m a) x i + i=m m+c 2 d i+(m b) x i + m 1 d i+(m b) x i + Since m = 2b + c and a = b + c, we have A 1 = 3b+2c 2 i=2b+c 3 d i+b x i + i=2b+c i=m m 2 d i+(m c) x i + 2b+2c 2 d i+b+c x i + i=2b+c d i+(m c) x i, (6) d i+2b x i, d i+m x i. B 1 = 2b+c 1 +c 2b+c 1 d i+b x i + 2b+c 1 d i+b+c x i + 2 d i+2b x i + d i+2b+c x i. (7)

7 A new class of irreducible pentanomials Determining A 2 and B 2 As before, we divide the leading term of A 1 by f and we obtain the remainder D 2. We get D 2 = D 2a + D 2b + D 2c, where D 2a, D 2b and D 2c refer to the reductions of the sums in Equation (6). We start with D 2a : a 2 D 2a = d i+2m a x i (x a + x b + x c + 1). Separating D 2a in the pieces A 2a and B 2a, we get A 2a = 2a 2 i=m d i+2m 2ax i since b + a 2 < m, and B 2a = m 1 i=a a+b 2 d i+2m 2a x i + a+c 2 d i+2m a b x i + Substituting m = 2b + c and a = b + c, we get A 2a B 2a = 2b+c 1 +c 2 d i+2b x i + b+2c 2 d i+2b+c x i + a 2 d i+2m a c x i + d i+2m a x i. = 2b+2c 2 i=2b+c d i+2b x i, and d i+3b x i + d i+3b+c x i. Proceeding with the reduction now of the second sum in Equation (6), we obtain D 2b = a+b 2 i=a 2b 2 d i+2m a b x i + d i+2m 2b x i + Clearly, D 2b is already reduced, and thus A 2b = 0, and B 2b = D 2c = 2 +c 2b 2 d i+2b+c x i + d i+2b+2c x i + b 2 d i+2m b c x i + We finally reduce the third and last sum in Equation (6): a+c 2 i=a d i+2m a c x i + 2c 2 d i+2m b c x i + d i+2m b x i. b 2 d i+3b+c x i + d i+3b+2c x i. c 2 d i+2m 2c x i + Again, we easily check that D 2c is reduced and so A 2c = 0, and B 2c = b+2c 2 +c d i+3b x i + Concluding, A 2 is given by 2c 2 d i+3b+c x i + A 2 = A 2a + A 2b + A 2c = 2a 2 i=m d i+2m c x i. c 2 d i+b x i + d i+b+c x i. d i+2m 2a x i, (8)

8 8 Gustavo Banegas et al. and B 2 = B 2a + B 2b + B 2c is 2b+c 1 B 2 = +c 2 b+2c 2 d i+2b x i + 2 d i+2b+c x i + +c d i+3b+c x i + b+2c 2 d i+3b x i + +c 2b 2 d i+2b+c x i + 2c 2 d i+3b x i + d i+b x i + d i+2b+2c x i + d i+3b+c x i + b 2 c 2 d i+3b+c x i + d i+3b+2c x i + d i+b+c x i. (9) 3. Determining A 3 and B 3 Dividing the leading term of A 2 in Equation (8) by f, we have D 3 = b+2c 2 +c d i+3b x i + 2c 2 d i+3b+c x i + We have that D 3 is reduced so A 3 = 0 and B 3 = b+2c 2 +c d i+3b x i + 2c 2 d i+3b+c x i + c 2 d i+b x i + d i+b+c x i. c 2 d i+b x i + d i+b+c x i. (10) 3.5 The number of terms in A r and B r Let G(i) = 1 if i > 0 and G(i) = 0 if i 0. Let r be a reduction step. It is clear now that the precise number of terms for A r and B r, for r 0, can be obtained using k b+c, k b and k c given in Equations (2), (3) and (). We have: i) The number of terms of A 0 and B 0 is 1. ii) For r > 0, the number of terms of A r is G(k b+c r)+g(k b r)+g(k c r), while the number of terms of B r is times the number of terms of A r 1. The family of polynomials f(x) = x 2b+1 + x b+1 + x b + x + 1 In this section, we consider the case when c = 1, that is, when k b+c = 2, as given in Equation (2). The polynomials in this subfamily have the form f(x) = x 2b+1 + x b+1 + x b + x + 1. For the subfamily treated in this section, since k b+c = 2, we immediately get A 2 = 0 and the expressions in the previous section simplify. As a consequence, the desired reduction is given by D red = B 0 + B 1 + B 2.

9 A new class of irreducible pentanomials 9 Using Equations (5), (7) and (9), we obtain 2b D red = d i x i + 2b +1 b 1 d i+2b+1 x i + 2b 1 +1 d i+b x i + b d i+2b x i + i=1 b 2b d i+3b x i + d i+b+1 x i + i=1 2b 2 d i+2b+1 x i + d i+2b+2 x i + b 2 d i+3b+2 x i + d 3b+1. (11) A crucial issue that allows us to give improved results for our family of pentanomials is the fact that redundancies occur for D red in Equation (11). Let b 2 T 1 (j) = (d i+2b+1 + d i+3b+2 )x i+j, T 2 (j) = d 3b x j, b 1 T 3 (j) = d 3b+1 x j, T (j) = (d i+2b+1 + d i+3b+1 )x i+j. A careful analysis of Equation (11) reveals that T 1, T 2 and T 3 are used more than once, and hence, savings can occur. We rewrite Equation (11) as D red =B 0 + T 1 (0) + T 1 (b) + T 1 (b + 1) + T 2 (b 1)+ T 2 (2b 1) + T 2 (2b) + T 3 (0) + T 3 (2b) + T (1). (12) One can check that by plugging T 1, T 2, T 3 and T in Equation (12) we recover Equation (11). Figure 2 shows these operations. We remark that even though the first row in this figure is B 0, the following two rows are not B 1 and B 2. Indeed, those rows are obtained from B 1 and B 2 together with the optimizations provided by T 1, T 2, T 3 and T. 2b 2b-1 2b-2 b+1 b-1 b B 0 m-1 m-2 m-3... a b... c T 3 T 2 T 1 T 2 T 1 T 2 T 1 = T T 3 D red B 0 + B 1 + B 2 Fig. 2 Representation of the reduction by f(x) = x 2b+1 + x b+1 + x b + x + 1. Using Equation (12), the number N of XOR operations is N = 6b + 1 = 3m 2.

10 10 Gustavo Banegas et al. It is also easy to see from Figure 2 that the time delay is 3T X, where T X is the delay of one 2-input XOR gate. We are now ready to provide Algorithm 1 for computing D red given in Equation (12), and as explained in Figure 2, for the pentanomials f(x) = x 2b+1 + x b+1 + x b + x + 1. Algorithm 1 Computing D red when f(x) = x 2b+1 + x b+1 + x b + x + 1. input : D 0 = d[b... 0] bits vector of length b + 1 output: D red for i 0 to b 2 do T 1 [i] d[i + 2b + 1] d[i + 3b + 2]; Definition of T 1 for i 0 to b 1 do T [i] d[i + 2b + 1] d[i + 3b + 1]; Definition of T D red [0] d[0] T 1 [0] d[3b + 1]; Column 0 of Fig. 2 for i 1 to b 2 do D red [i] d[i] T 1 [i] T [i 1]; Columns 1 to b 2 of Fig. 2 D red [b 1] d[b 1] d[3b] T [b 2] D red [b] d[b] T 1 [0] T [b 1] for i b + 1 to 2b 2 do D red [i] d[i] T 1 [i b] T 1 [i b 1]; Columns b + 1 to 2b 2 of Fig. 2 D red [2b 1] d[2b 1] d[3b] T 1 [b 2] D red [2b] d[2b] d[3b + 1] d[3b] return D red Putting all pieces together, we give next the main result of this section. Theorem 1 Algorithm 1 correctly gives the reduction of a polynomial of degree at most 2m 2 over F 2 by f(x) = x 2b+1 + x b+1 + x b + x + 1 involving N = 3m 2 = 6b + 1 number of XORs operations and a time delay of 3T X. 5 Family f(x) = x 2b+c + x b+c + x b + x c + 1, c > 1 For polynomials of the form f(x) = x 2b+c + x b+c + x b + x c + 1, c > 1, we have that k b+c = 3, implying that A 3 = 0. The reduction is given by D red = B 0 + B 1 + B 2 + B 3.

11 A new class of irreducible pentanomials 11 Using Equations (5), (7), (9) and (10), we have that D red satisfies 2b+c 1 D red = Let 2b+c 1 2b 2 2b+c 1 d i x i + +c b 1 d i+b+c x i + b+c 1 d i+b x i + 2 d i+2b+c x i + b+2c 2 d i+2b x i + +c d i+2b+c x i + c 1 b 2 d i+2b+2c x i + d i+3b+c x i + d i+3b+2c x i. b 2 T 1 (j) = (d i+2b+c + d i+3b+2c )x i+j, T 2 (j) = d 3b+c 1 x j, d i+3b x i + (13) c 1 b 2 b 2 T 3 (j) = d i+3b+c x i+j, T (j) = d i+2b+c x i+j, T 5 (j) = d i+3b+2c x i+j. Again, a careful analysis of Equation (13) shows that T 1, T 2 and T 3 are used more than once. Thus, we can rewrite Equation (13) for D red as D red =B 0 + T 1 (0) + T 1 (b) + T 1 (b + c)+ T 2 (b 1) + T 2 (b + c 1) + T 2 (2b 1) + T 2 (2b + c 1)+ T 3 (0) + T 3 (c) + T 3 (2b) + T (c) + T 5 (2c). (1) Figure 3 depicts these operations. Using Equation (1) and Figure 3, we have Algorithm 2. For code efficiency reasons, in contrast to Algorithm 1, in Algorithm 2 we separate the last line before the equality in Figure 3. The additions of this last line are done in lines 17 to 20 of Algorithm 2. As a consequence, lines 3 to 16 of Algorithm 2 include only the additions per column from 0 to 2b + c 1 of the first three lines in Figure 3. 2b+c-1 2b+c-2 2b 2b-1 b+c b+c-1 2c b-1 c-1 0 B 0 D red m-1 m a b... c... 2b+c-1 T 2b T 2b-2 2 T b 3 1 T b-2 2 T 0 1 T 2b+c-2 2 T b+c 1 T b+c-2 2 T c b+2c-2 T 2c 2c-1 5 T c c-1 3 T 0 3 = B 0 + B 1 + B 2 + B 3 Fig. 3 Representation of the reduction by f(x) = x 2b+c + x b+c + x b + x c + 1, c > 1. The time delay is 3T X ; after removal of redundancies and not counting repeated terms, we obtain that the number N of XORs is N = 6b + 3c 2 = 3m 2.

12 12 Gustavo Banegas et al. Algorithm 2 Computing D red for f(x) = x 2b+c + x b+c + x b + x c + 1. input : D 0 = d[2b + c ] bits vector of length 2b + c output: D red for i 0 to b 2 do T 1 [i] d[i + 2b + 1] d[i + 3b + 2c]; Definition of T 1 for i 0 to c 1 do D red [i] d[i] T 1 [i]; Columns 0 to c 1 of the first three lines of Fig. 3 for i c to b 2 do D red [i] d[i] T 1 [i] d[i + 2b] D red [b 1] d[b 1] d[3b + c 1] d[3b 1] for i b to b + c 2 do D red [i] d[i] T 1 [i b] d[i + 2b] D red [b + c 1] d[b + c 1] d[3b + c 1] T 1 [c 1] for i b + c to 2b 2 do D red [i] d[i] T 1 [i b] T 1 [i b c] D red [2b 1] d[2b 1] d[3b + c 1] T 1 [b c 1] for i 2b to 2b + c 2 do D red [i] d[i] T 1 [i b c] d[i + b + c] D red [2b + c 1] d[2b + c 1] d[3b + c 1] d[3b 1] for i 0 to c 1 do D red [i] D red [i] d[i + 3b + c]; Columns 0 to c 1 of the th line of Fig. 3 for i c to b + 2c 2 do D red [i] D red [i] d[i + 3b]; Cols c to b + 2c 2 of the th line of Fig. 3 return D red Theorem 2 Algorithm 2 correctly gives the reduction of a polynomial of degree at most 2m 2 over F 2 by f(x) = x 2b+c + x b+c + x b + x c + 1 involving N = 3m 2 = 6b + 3c 2 number of XORs operations and a time delay of 3T X. 5.1 Almost equally spaced pentanomials: the special case b = 2c Consider the special case b = 2c. In this case we obtain the almost equally spaced polynomials f(x) = x 5c + x 3c + x 2c + x c + 1. Our previous analysis when applied to these polynomials entails 5c 1 D red = 2c 1 5c 1 d i x i + i=3c 5c 2 d i+5c x i + 3c 1 d i+2c x i + i=3c c 2 d i+5c x i + c 2 d i+c x i + i=2c c 1 d i+6c x i + 5c 1 d i+6c x i + i=2c 2c 2 d i+7c x i + d i+3c x i + d i+8c x i. (15)

13 A new class of irreducible pentanomials 13 Let 2c 2 2c 2 T 1 (j) = (d i+5c + d i+c )x i+j, T 2 (j) = (d i+8c + d i+6c )x i+j, c 1 c 1 T 3 (j) = d 8c 1 x j, T (j) = d i+8c x i+j, T 5 (j) = d i+5c x i+j, c 2 T 6 (j) = d i+7c x i+j, T 7 (j) = 5c 1 i=c d i+2c x i+j. In the computation of D red, T 1, T 2, T 3 and T are used more than once. Figure shows, graphically, these operations. After removal of redundancies, the number N of XORs is N = 12c 1 = 12 m 1. This number of XORs 5 is close to 2.m providing a saving of about 20% with respect to the other pentanomials in our family. Irreducible pentanomials of this form are rare but they do exist, for example, for degrees 5, 155 and 805. We observe that the extension 155 is used in [1]. 5c-1 5c-2 c c-1 3c 3c-1 2c 2c-1 c-1 0 B 0 m-1 m-2... a... b... c... T 7 T T 3 T 2 T T 3 T 2 T 1 T 1 T 1 T 5 = T 6 D red B 0 + B 1 + B 2 + B 3 Fig. Representation of the reduction by the almost equally spaced pentanomials (the special case b = 2c). Using Equation (15) and Figure, we naturally have Algorithm 3. 6 Multiplier in F 2 [x], complexity analysis and comparison Our multiplier consists of two steps. The first is a multiplication itself using schoolbook or Karatsuba arithmetic, and the second is the reduction described in the previous sections. The choice of the first step method will basically dep on whether the application requirement is to minimize area, i.e., the number of ANDs and XORs gates, or to minimize the arithmetic delay; see [3] for several variants of both the schoolbook and Karatsuba algorithms. Minimizing the area is interesting in applications that need to save power at the expense of a longer runtime.

14 1 Gustavo Banegas et al. Algorithm 3 Computing D red for f(x) = x 5c + x 3c + x 2c + x c + 1. input : D 0 = d[5c ] bits vector of length 5c output: D red for i 0 to c 2 do T 1 [i] d[i + 6c] d[i + 5c] for i 0 to c 2 do T 2 [i] d[i + 9c] d[i + 7c] for i 0 to c 2 do D red [i] d[i] d[i + 8c] d[i + 5c] d[i + 7c] D red [c 1] d[c 1] d[9c 1] d[6c 1] for i c to 2c 2 do D red [i] d[i] T 1 [i c] T 2 [i c] D red [2c 1] d[2c 1] d[8c 1] T 1 [c 1] for i 2c to 3c 1 do D red [i] d[i] T 1 [i 2c] for i 3c to c 1 do D red [i] d[i] T 1 [i 3c] d[i + 5c] for i c to 5c 2 do D red [i] d[i] T 2 [i c] d[i + 2c] D red [5c 1] d[5c 1] d[8c 1] d[7c 1] return D red A summary of our results compared to that of Park et al. [1] is given in Table 1, since their algorithm is very efficient in terms of running time, even though its energy consumption is high due to a greater number of gates. With respect to Karatsuba variants, Table 3 of survey [3] shows several algorithms and their costs for the multiplication part, without reduction, for special irreducible pentanomials including Type 1, Type 2 and equally spaced pentanomials. We note that the given spatial complexity is given only for irreducible polynomials of degree powers of 2 and powers of 3 only. In contrast, our analysis holds for any degree for the multiplication step and irreducible polynomials as given by our family. In Table 1, we provide the number of XORs and ANDs gates, and the delay for Type 1 and Type 2 families in [1] and our family of pentanomials. In the case of our family, in addition to the number of XORs for the reduction considered in the previous sections, we include the cost for the multiplication using our recursive Karatsuba implementation multiplier. In our implementation, the constant of Karatsuba is strictly less than 6; see Figure 5 for degrees up to 1, 02. As can be seen, for degrees powers of 2, the constant achieves a local minimum. For the number of AND gates, we only provide an interval. The actual number of AND gates deps on the value of m; it only reaches a maximum when m = 2 k 1, for k 1.

15 A new class of irreducible pentanomials C ,000 Deg Fig. 5 Karatsuba constant for degrees up to 1, 02. The costs for using polynomials of degrees proposed by NIST can be found in Table 2. The AND count is the exact value of the number of products, where we note that this value lies in the interval given in Table 1. The delay is due to only 1 AND gate at the lowest level of the Karatsuba recursion. The XOR count is formed by the XORs of Karatsuba multiplier and the ones of the reduction part. For the Karatsuba multiplier we have 3 log 2 (m 1) since there are 3 delay XORs per level of the Karatsuba recursion. The reduction part accounts for only 3 delay XORs. Hence, the total number for the XOR delays 3 log 2 (m 1) + 3. Table 1 Space complexity of known pentanomials using classical arithmetic. Type # XOR # AND Type 1 x m + x b+1 + x b + x + 1, 1 < b m 2 1 3m 2 + 2m + 8b m 2 + 2m 1 [1] b is odd 3m 2 + 2m + 8b m 2 + 2m 1 [1] b is even Type 2 x m + x c+2 + x c+1 + x c + 1, 1 < b m 2 1 [1] c is odd, c 3 8 (m 7) 3m 2 + 2m + 1c m 2 + 2m 1 3m 2 + 2m + 1c + 5 3m 2 + 2m 1 [1] c is even Type 3 x 2b+c + x b+c + x b + x c + 1, c 1, b 2c Ours < 6m log m 2 (3 log 2 m, 3 log 2 m +1 ] Type x 5c + x 3c + x 2c + x c + 1, c 1 Ours < 6m log m 1 (3 log 2 m, 3 log 2 m +1 ] We point out that in [1] the authors use classical arithmetic and the multiplication and reduction is done as a unique block. In contrast, we separate these two parts and use Karatsuba for the multiplier followed by our reduction. In the following we comment on the density of irreducible pentanomials in our family. Table 3 lists (b, c, N ) for all irreducible Type 3 pentanomials of

16 16 Gustavo Banegas et al. Table 2 Costs for fixed degrees proposed by NIST. Degree XORs Karatsuba Reduction Total ANDs Delay , , 31, 19 T A + 27T X , , 71 6, 339 T A + 27T X 283 3, , , 305 T A + 30T X 09 75, 21 1, , 39 17, 133 T A + 30T X , 280 1, , , 203 T A + 33T X our family for degrees up to 1, 02; N is the number of XORs required for the reduction. We leave as an open problem to mathematically characterize under which conditions our pentanomials are irreducible. Table 3: Irreducible Type 3 pentanomials and their number of XORs (b, c, N ). 2, 1, 11 3, 2, 22, 1, 25 5, 1, 31 5, 2, 3 6, 1, 37 5, 3, 37 7, 2, 6 9, 5, 67 8, 7, 67 9, 6, 70 12, 1, 73 11, 3, 73 10, 7, 79 13, 3, 85 10, 9, 85 13,, 88 15, 6, 106 1, 9, , 2, , 6, , 10, , 11, , 12, , 5, , 7, , 15, , 6, 12 23, 5, , 7, , 2, 15 21, 11, , 13, , 5, , 13, , 2, , 10, , 1, , 12, 18 28, 7, , 1, , 9, ,, , 20, , 7, , 15, , 18, 21 35, 3, , 11, , 22, , 20, , 10, , 19, 21 38, 7, 27 31, 21, 27 1, 3, , 9, , 12, , 19, , 12, 268 3, 25, 277 5,, , 29, 283 7, 2, 286 0, 17, , 23, 295 8, 7, 307 0, 23, 307 6, 15, 319 2, 23, , 2, 322 5, 18, 322 1, 26, 322 5, 19, , 33, 325 1, 28, , 7, 331 1, 29, 331 7, 20, 30 5, 26, 36 3, 30, 36 9, 19, 39 1, 35, 39 5, 28, , 6, , 18, 358 5, 30, 358 6, 31, , 1, , 25, ,, , 7, 391 5,, 00 51, 3, 06 59, 19, 09 50, 1, 21 63, 18, 30 68, 9, 33 63, 19, 33 59, 27, 33 56, 33, 33 67, 12, 36 69, 11, 5 60, 31, 51 75, 2, 5 56, 1, 57 63, 29, 63 62, 31, , 37, 63 75, 6, 66 71, 1, 66 65, 26, 66 61, 36, 72 77, 5, 75 7, 15, 87 63, 37, 87 67, 30, 90 65, 3, 90 73, 19, 93 71, 30, 51 87, 2, , 6, , 30, , 2, , 17, 51 71, 6, , 9, , 28, , 36, , 21, , 61, , 28, , 10, , 30, , 50, , 12, 60 98, 9, , 33, , 3, , 9, , 51, , 3, , 28, , 37, , 55, , 11, 637 7, 65, , 7, 63 81, 5, 66 79, 60, , 61, , 18, , 53, , 1, ,, , 7, , 69, , 66, , 19, , 9, , 3, , 75, , 31, , 37, , 53, , 2, 706 9, 9, , 31, , 2, , 30, , 23, , 37, , 3, , 6, , 73, 75 99, 52, , 15, , 5, , 61, , 23, , 3, , 67, ,, , 68, , 2, , 83, , 1, , 1, , 67, , 7, , 7, , 59, , 81, , 30, , 70, , 36, , 31, , 61, , 3, , 63, , 1, , 13, , 1, 81 95, 91, , 37, , 61, , 5, , 9, , 59, , 6, , 76, , 31, , 2, , 3, , 15, , 22, , 50, , 70, , 3, , 1, , 25, , 3, , 102, 96 15, 9, 99 11, 89, , 99, 99 15, 3, , 50, , 5, , 78, , 33, , 36, , 60, , 85, , 6, ,, , 8, , 7, , 29, , 25, , 107, , 105, ,, , 52, , 68, , 92, , 67, , 78, , 90, , 91, , 8, , 7, , 103, , 2, , 35, , 9, , 93, , 18, , 5, , 90, , 5, , 103, , 27, , 1, , 81, , 99, , 121, , 121, , 113, , 29, , 3, , 89, ,, , 28, , 60, 11

17 A new class of irreducible pentanomials , 62, , 123, , 111, , 101, , 10, , 27, , 20, , 6, , 50, , 81, , 132, , 63, , 71, , 109, , 102, , 28, , 37, , 93, , 100, , 61, , 109, , 1, , 98, , 25, , 81, , 121, , 132, , 21, , 5, , 109, , 13, , 93, , 13, , 22, , 66, , 9, , 63, , 77, , 57, , 105, ,, , 108, , 6, , 9, , 19, , 137, , 10, , 55, , 77, , 69, , 18, , 2, , 78, , 76, , 108, , 7, , 69, , 59, , 155, , 10, , 12, , 12, , 9, , 17, , 90, ,, , 18, , 63, , 9, , 12, , 1, , 90, , 83, , 115, , 71, , 7, , 6, , 163, , 76, , 95, , 175, ,, , 53, , 13, , 2, , 2, , 10, , 18, , 58, , 90, , 98, , 126, , 130, , 9, , 1, , 153, , 161, , 10, , 39, , 161, , 37, , 127, , 77, , 6, , 102, , 67, , 89, , 163, , 10, , 71, , 68, , 172, , 173, , 138, , 17, , 180, , 9, , 139, , 8, , 135, , 9, , 13, , 91, , 15, , 130, , 13, , 150, , 8, , 71, , 167, , 163, , 12, , 77, , 189, , 129, , 163, , 92, , 2, , 66, , 86, , 25, , 30, , 90, , 150, , 67, , 75, , 181, , 18, , 36, , 116, , 9, , 81, , 107, , 171, ,, , 17, , 113, , 61, , 91, , 12, 186 2, 137, , 172, , 8, , 127, , 13, , 93, , 103, , 117, , 109, , 131, , 137, , 212, , 7, , 181, , 36, , 157, , 2, , 162, , 131, , 203, , 92, , 175, , 105, , 6, , 73, , 15, , 22, , 5, , 15, , 7, , 1, , 7, , 30, , 162, , 9, , 3, , 102, , 170, , 68, , 198, , 27, , 179, , 211, , 175, , 11, , 73, , 83, , 203, , 116, , 205, , 161, , 91, , 63, , 101, , 19, , 220, , 122, , 11, , 139, , 203, , 183, , 207, , 102, , 69, , 6, , 90, , 26, , 67, , 9, , 37, , 50, , 138, , 16, , 173, , 199, , 2, , 139, , 11, , 221, , 18, , 156, , 21, , 217, , 13, , 1, , 182, , 9, , 36, , 103, , 10, , 21, , 20, , 139, , 251, , 173, , 239, , 6, , 161, , 29, , 62, , 130, , 31, , 93, , 189, , 9, , 203, , 76, , 26, , 2, , 130, , 182, , 25, , 10, , 91, , 121, , 66, , 182, , 233, , 187, , 15, , 180, , 31, , 261, , 1, , 190, , 210, , 169, , 108, , 35, , 93, , 102, 2.6 0, 9, , 133, , 25, , 1, , 150, , 170, , 23, , 27, , 255, , 2, , 101, ,, , 1, , 270, , 76, , 5, , 269, , 203, , 38, , 25, , 211, , 185, , 217, , 38, , 5, , 150, , 182, , 217, , 18, , 59, , 91, , 153, , 228, , 231, , 115, , 27, , 99, , 173, , 10, , 81, , 89, , 181, , 231, , 29, , 1, , 121, , 16, , 252, , 28, , 270, , 91, , 23, , 18, , 1, , 66, , 7, , 150, , 210, , 298, , 177, , 10, , 175, , 19, , 173, , 98, , 17, , 163, , 102, , 106, , 162, , 67, , 289, , 5, , 25, , 287, ,, , 10, , 13, , 37, , 83, , 12, , 262, , 107, , 151, , 29, , 166, , 230, , 16, , 15, , 223, , 306, , 23, , 31, , 133, , 169, , 76, , 89, , 287, , 78, 2890

18 18 Gustavo Banegas et al. 373, 218, , 123, , 7, , 207, , 102, , 150, , 162, , 23, , 223, , 137, , 108, , 180, , 63, , 89, , 271, , 107, , 179, , 133, , 18, , 287, , 132, , 252, , 197, , 207, , 265, , 7, , 222, , 9, , 139, , 50, , 150, , 250, , 10, , 71, , 165, , 332, , 156, , 196, , 316, , 37, , 98, , 302, , 3, , 28, , 187, , 183, Table shows the number of irreducible pentanomials of degrees 163, 283 and 571 for the families considered since those are NIST degrees where pentanomials have been recommed [12]. Analyzing the table, we have that Family Type 1 has the most irreducible pentanomials, but few of them have degrees recommed by NIST [12]. Family Type 2 has less irreducible pentanomials, but has the highest number of representatives with NIST degrees of interest. Our family (Type 3 and ) has 728 polynomials of degrees up to 1, 02. Family Type 3 has 5 pentanomials of NIST degrees while Type has no NIST degrees pentanomials. Table Number of irreducible pentanomials for NIST degrees. Type #Irred Type 1 2, Type 2 1, Type Type Conclusions In this paper, we present a new class of pentanomials over F 2, defined by x 2b+c +x b+c +x b +x c +1. We give the exact number of XORs in the reduction process; we note that in the reduction process no ANDs are required. It is interesting to point out that even though the cases c = 1 and c > 1, as shown in Figures 2 and 3, are quite different, the final result in terms of number of XORs is the same. We also consider a special case when b = 2c where further reductions are possible. There are irreducible pentanomials of this shape for several degree extensions of practical interest. We provide a detailed analysis of the space and time complexity involved in the reduction using the pentanomials in our family. For the multiplication process, we simply use the standard Karatsuba algorithm. The proved complexity analysis of the multiplier and reduction considering the family proposed in this paper, as well as our analysis suggests that these pentanomials are as good as or possibly better to the ones already proposed.

19 A new class of irreducible pentanomials 19 References 1. G.B. Agnew, R.C. Mullin, and S.A. Vanstone. An implementation of elliptic curve cryptosystems over F IEEE Journal on Selected Areas in Communications, 11(5):80 813, A. Cilardo. Fast parallel GF (2 m ) polynomial multiplication for all degrees. IEEE Transactions on Computers, 62(5):929 93, H. Fan and M.A. Hasan. A survey of some recent bit-parallel GF (2 n ) multipliers. Finite Fields and Their Applications, 32:5 3, A. Halbutogullari and C.K. Koç. Mastrovito multiplier for general irreducible polynomials. IEEE Transactions on Computers, 9: , M.A. Hasan, M. Wang, and V.K. Bhargava. Modular construction of low complexity parallel multipliers for a class of finite fields GF (2 m ). IEEE Transactions on Computers, 1: , J.L. Imaña. High-speed polynomial basis multipliers over GF (2 m ) for special pentanomials. IEEE Transactions on Circuits and Systems I: Regular Papers, 63(1):58 69, J.L. Imaña, R. Hermida, and F. Tirado. Low complexity bit-parallel multipliers based on a class of irreducible pentanomials. IEEE Transactions on Very Large Scale Integration (VLSI) Systems, 1(12): , J.L. Imaña, R. Hermida, and F. Tirado. Low complexity bit-parallel polynomial basis multipliers over binary fields for special irreducible pentanomials. Integration, the VLSI Journal, 6(2): , T. Itoh and S. Tsujii. Structure of parallel multipliers for a class of fields GF (2 m ). Information and Computation, 83(1):21 0, R. Lidl and H. Niederreiter. Finite Fields. Cambridge University Press, Cambridge, G.L. Mullen and D. Panario. Handbook of Finite Fields. CRC Press, NIST. Digital signature standard. Report P-1, National Institute of Standards and Technology (NIST), T. Oliveira, J. López, and F. Rodríguez-Henríquez. Software implementation of Koblitz curves over quadratic fields. In International Conference on Cryptographic Hardware and Embedded Systems, pages Springer, S.-M. Park, K.-Y. Chang, D. Hong, and C. Seo. New efficient bit-parallel polynomial basis multiplier for special pentanomials. Integration, the VLSI Journal, 7(1): , A. Reyhani-Masoleh and M.A. Hasan. Low complexity bit parallel architectures for polynomial basis multiplication over GF (2 m ). IEEE Transactions on Computers, 53:95 959, F. Rodríguez-Henríquez and C.K. Koç. Parallel multipliers based on special irreducible pentanomials. IEEE Transactions on Computers, 52: , M. Scott. Optimal irreducible polynomials for GF (2 m ) arithmetic. IACR Cryptology eprint Archive, 2007:192, B. Sunar and C.K. Koç. Mastrovito multiplier for all trinomials. IEEE Transactions on Computers, 8: , H. Wu. Bit-parallel finite field multiplier and squarer using polynomial basis. IEEE Transactions on Computers, 51(7): , H. Wu. Bit-parallel polynomial basis multiplier for new classes of finite fields. IEEE Transactions on Computers, 57: , T. Zhang and K.K. Parhi. Systematic design of original and modified mastrovito multipliers for general irreducible polynomials. IEEE Transactions on Computers, 50(7):73 79, 2001.