Generating Primes Using Partitions

Ganesh Reddy Pittu

Abstract: This paper presents a new technique of generating large prime numbers using a
smaller one by employing Goldbach partitions. Experiments are presented showing how this
method produces candidate prime numbers that are subsequently tested using either Miller Rabin
or AKS primality tests.

Introduction
Generation of large prime numbers is fundamental to modern cryptography protocols [1],[2],
generation of pseudorandom sequences [3]-[5], and in new application of these protocols to
multi-party computing and cryptocurrencies [6]-[8]. Public-key cryptography requires random
generation of prime numbers to derive public key. This paper presents a new technique of
generating large prime numbers using a smaller one by employing Goldbach partitions [10]. The
algorithm is described and experiments are presented showing how this method gives large
primes in an effective manner. A candidate prime will be tested using either Miller-Rabin (MR)
or AKS primality tests [11],[12].

Generation of random primes

Large prime numbers are generated randomly by considering a random number and testing it
with MR or AKS primality test or one might use different sieves [13]-[18], with applications to a
variety of cryptography areas (e.g. [19],[20]). The table below presents the average of 10
executions for random generation of prime numbers in a typical experiment.
Table 1. Average attempts to generate a random prime

Length of the Average

Random Prime attempts to
number generate a
prime number

45 98

50 159

55 172

60 211

70 224

1
Figure 1. Average of 10 executions to generate a random prime number vs. Length of the prime
number. X-axis: Length of the prime number, Y-Axis: Average Attempts.
The average number for 10 executions increases with the size of the number, and it ranges from
about 100 for a 45-digit number to 224 for a 70-digit number.
Generating primes using Goldbach partitions
The method of prime number generation using Goldbach partitions is now presented. Let g(n) be
the number of unique ways n can be expressed as p + q, where p and q are primes. The partition
function g(n) has a local maximum for multiples of primes. We obtain local maximum at
multiples of primes such as 2x3 = 6; 2x3x5 = 30; 2x3x5x7 = 210; 2x3x5x11 = 330; 2x3x5x13 =
390; 2x3x5x17= 510; 2x3x5x19 = 570; 2x3x5x7x11 = 2310; 2x3x5x7x11x13 = 30030;
2x3x5x7x11x13x17 = 510510, etc. For n=, prime numbers in the range [n/2, n-2] are among the
partitions. The number n=210 is the largest number where all numbers in Goldbach partitions are
prime numbers.

Example for n=30: the Goldbach partitions are (23, 7), (19, 11) and (17, 13).
Example for n=210: the Goldbach partitions are (199,11), (197,13), (193,17), (191,19), (181,29),
(179,31), (173, 37), (167, 43), (163, 47), (157, 53), (151, 59), (149, 61), (139, 71), (137, 73),
(131, 79), (127, 83), (113, 97), (109, 101), (107, 103). If one counted from the smaller prime up
then (187,23) is one pair that is left out since 187 is not prime.

Since we are interested in generating candidate primes, we define h(n) as the count of all primes
in the range [n/2, n-2].

h(n)  g (n).

Equality is obtained only for n=6, 30, and 210.

Table 2 presents the number of candidate partitions in the range [n/2, n-2] obtained by the
product of prime numbers for some values of n.

2
 Table 2. Number of candidate partitions, h(n)
Prime n No. of Prime Numbers n No. of
Numbers Candidate candidate
Partitions Partitions
h(n) h(n)
2,3,5 30 3 2,3,5,23 690 56

2,3,7 42 4 2,3,7,17 714 56

2,3,11 66 7 2,3,7,19 798 60

2,3,13 78 9 2,3,11,13 858 65

2,3,17 102 10 2,3,5,29 870 66

2,3,19 114 13 2,3,5,31 930 67

2,3,23 138 13 2,3,7,23 966 70

2,3,29 174 16 2,3,5,37 1110 84

2,3,31 186 18 2,3,11,17 1122 85

2,3,5,7 210 19 2,3,7,29 1218 87

2,3,5,11 330 28 2,3,5,41 1230 87

2,3,5,13 390 32 2,3,11,19 1254 90

2,3,7,11 462 38 2,3,5,7,11 2310 151

2,3,5,17 510 42 2,3,5,7,13 2730 179

2,3,7,13 546 42 2,3,5,7,17 3570 223

2,3,5,19 570 42 2,3,5,7,19 3990 248

2,3,7,11,13 6006 352 2,3,7,11,17 7854 447

2,3,7,11,19 8778 496 2,3,11,13,17,19 277134 11323

2,3,7,13,17 9282 522 2,3,11,13,17,23 335478 13517

2,3,7,13,19 10374 582 2,3,11,17,19,23 490314 19148

3
 The graphs below compares the value of n and the number of candiate partitions computed.

Figure 2. No. of partition, h(n), for n<1000

Figure 3 No. of partitions, h(n), for 1000<n<50000

Using this technique, we investigate percentage of prime numbers obtained from the partitions
by MR and AKS primality testing algorithms.

4
 Table 3 Percentage of Prime Partitions

Prime Product Percentage of Prime Product Percentage of

Prime Partitions Prime Partitions

30 100 714 66.01

42 80 798 63.30

66 85.71 858 65.60

78 77.77 870 69.69

102 80 930 64.17

114 76.92 966 64.28

138 61.53 1110 64.28

174 68.75 1122 55.29

186 72.22 1218 56.32

210 100 1230 62.50

330 85.71 1254 56.66

390 84.37 2310 75.49

462 73.68 2730 71.50

510 76.19 3570 69.05

546 71.42 3990 65.72

570 73.80 6006 55.39

690 69.64 7854 52.12

8778 49.19 277134 33.52

9282 48.65 335478 32.72

10374 50 490314 30.64

The graphs below presents the percentage of prime numbers generated from the Goldbach
partitions.

5
 Figure 4. No. of candiate partitions vs percentage of primes for n<1000

Figure 5. No. of candiate partitions vs percentage of primes for 50000 <n<1000000

Now we present the new technique of generating a prime number using a smaller one and we call
our method the Goldbach Prime Generating Algorithm(GPGA) is presented. This method can be
used as an alternate method in public key cryptography and in authentication techniques.

The GPGA Algorithm:

1. Randomly pick a large even number n.

2. Use a known prime number of certain size, p a ,( p a << n) and p a + p b = n, where p b is a

candidate prime.

3. Test p b with AKS or MR primality testing algorithms.

4. Repeat 2,3 till p b is a prime number.

6
The above algorithm is implemented below.

GPGA pseudo code:

n:  p , p =2,3,5,7,…..i
i 1 i

Whlie prime == false

p a : Known Prime Number (p a << n)

pb: n - pa

Test p b for prime using MR or AKS

If p b prime

Then
Prime = true
EndWhile

Return p b .

Table 4. GPGA results for the average of 10 Executions

Length of n Length of p b (known Average attempts to

(even number) prime) generate prime number

45 4 21

45 5 18

45 6 20

45 7 15

45 8 18

45 9 15

60 4 10

60 5 16

7
 60 6 12

60 7 18

60 8 12

60 9 14

Figure 6. Attempts To Generate Candidate Prime when length of n = 45

Figure 7. Attempts To Generate Candidate Prime when length of n = 60

The above average case is taken for ten executions of the algorithm and are considered for 10
best cases. The while loop iteration might take even more attempts in few cases and will be the
worst case. So the best case to obtain a prime number would be O(1).

Conclusions
This paper has proposed a new method of generating prime candidates using partitions, which
are then tested using primality testing algorithms (MR and AKS). We provide results on the
8
efficiency of this algorithm. The use of the GPGA method reduced the number of trials in the
search for primes by a factor of nearly 10 for primes ranging from 45 to 70 digits long.

References
1. W. Diffie and M. Hellman, New directions in cryptography. IEEE Trans. Inf. Theory, 1976.
2. S. Singh, The Code Book: the Secret History of Codes and Code-breaking. Fourth Estate,
London, 1999.
3. S. Kak, Classification of random binary sequences using Walsh-Fourier analysis. IEEE
Trans. on Electromagnetic Compatibility, EMC-13: 74-77, 1971.
4. S. Kak and A. Chatterjee, On decimal sequences. IEEE Trans. on Information Theory IT-27:
647 – 652, 1981.
5. S. Kak, Encryption and error-correction coding using D sequences. IEEE Trans. on
Computers C-34: 803-809, 1985.
6. S. Kak, The piggy bank cryptographic trope. Infocommunications Journal 6: 22-25, 2014.
7. B. Prashanth Reddy, A review of the bitcoin network security framework. 2015.
https://subhask.okstate.edu/sites/default/files/prashanthcrypto.pdf
8. I. Miers. and C. Garman., M. Green., A. D. Rubin, Zerocoin: Anonymous distributed e-cash
from bitcoin, IEEE Symposium on Security and Privacy, 2013.
9. L. Washbourne, A survey of P2P Network security. arXiv:1504.01358, 2015
10. S. Kak, Goldbach partitions and sequences. Resonance 19: 1028-1037, 2014.
11. M. Agrawal, N. Kayal and N. Saxena, PRIMES is in P, Ann. of Math. (2) 160, 781-793,
2004.
12. R. Crandall and J. Papadopoulos, On the implementation of AKS-cClass Primality tests, 18
Mar 2003.
13. G.H Hardy, E.M. Wright , An Introduction to the Theory of Numbers, 2008.
14. A. Kak, Computer and Network Security, Purdue University, 2015.
15. S. Kak, A primes conjecture, 2015.
http://subhask.okstate.edu/sites/default/files/primes_conjecture.pdf
16. A. Atkin, D.J. Bernstein, Prime sieves using binary quadratic forms, Mathematics of
Computation 73: 1023–1030, 2004.
17. P. Pritchard, Improved incremental prime number sieves. Algorithmic Number Theory
Symposium, pp. 280–288, 1994.
18. J.P. Sorenson, Trading time for space in prime number sieves. Lecture Notes in Computer
Science 1423: 179–195, 1998.
19. P.S. Potluri, A review of anonymous file sharing on the cloud.
https://subhask.okstate.edu/sites/default/files/pushpa5.pdf
20. D. Eastlake 3rd, S. Crocker, J. Schiller, Randomness Recommendations for Security.
Network Working Group, MIT, 1994.