NEW MERSENNE PRIMES

By Alexander Hurwitz

    If p is prime, Mp = 2^p - 1 is called a Mersenne number. The primes M4253 and M4423 were discovered by coding the Lucas-Lehmer test for the IBM 7090. These two new primes are the largest prime numbers known; for other large primes see Robinson [4]. The computing was done at the UCLA Computing Facility. This test is described by the following theorem (see Lehmer [1, p. 443-4]).
    THEOREM. If S[1] = 4 and S[n+1] = S[n]^2 - 2 then Mp is prime if and only if S[p-1] == 0 (mod Mp).
    The test takes about 50 minutes of machine time for p = 4423. These results bring the number of known Mersenne primes to 20. The values of p for these twenty primes are listed in Table 1.
    If Mp is prime it is of interest to know the sign of the least absolute penultimate residue, that is, whether S[p-2] == +2^r (mod Mp) or S[p-2] == -2^r (mod Mp) where 2r = p+1. The Lucas-Lehmer test can also be used with S[1] = 10. The various penultimate residues of the known Mersenne primes were computed and the results appear in Table 1 (see Robinson [3]).
    In addition to testing the above Mersenne primes each Mersenne number with p < 5000 was tested unless a factor of Mp was known. The residues of S[p-1] (mod Mp) are available for checking purposes. The results for 3300 < p < 5000 are summarized in Table 2. The computer program also found (see [3, p.844]) that M8191 is not prime.
    The residue S[p-1] (mod Mp) for p > 3300 is output from the computer in a modified octal notation. That is, the residue is stored in the computer in 35 bit binary words and the output is a word by word conversion of the 35 bit words into octal (base 8) numbers. Thus the leading digit of each is quaternary (base 4). For p < 3300 the residue was printed in hexadecimal notation (see Robinson [3] and Reisel [2]).

TABLE 1
p S[1] = 4 S[1] = 10 p S[1] = 4 S[1] = 10
2     107 - +
3 + - 127 + +
5 + - 521 - +
7 - - 607 - -
13 + + 1279 - -
17 - + 2203 + -
19 - + 2281 - +
31 + + 3217 - +
61 + + 4253 + +
89 - + 4423 - -

TABLE 2
       R        R
           
3301     72013 4241     11012
3307     62061 4253     00000
3313     10050 4259     46007
3331     51270 4261     55632
3343     76415 4283     74774
3371     57040 4339     41356
3373     36120 4349     74465
3389     64705 4357     74271
3413     50261 4363     61114
3461     03241 4397     40174
           
3463     57665 4409     51070
3467     23046 4421     25131
3469     21765 4423     00000
3547     75574 4481     70216
3559     45350 4493     36053
3583     42507 4519     01571
3607     45062 4523     22235
3617     35431 4567     74267
3631     14530 4583     46556
3637     67413 4591     47243
           
3643     04606 4621     74601
3671     04031 4643     51444
3673     01626 4651     00707
3691     54715 4663     52442
3697     53743 4673     40333
3709     06427 4679     14305
3739     22413 4703     54013
3769     00747 4721     04420
3821     52075 4729     40137
3833     45453 4733     12774
           
3847     57652 4783     77350
3877     46507 4789     02364
3881     34503 4799     04305
3889     30737 4817     70020
3919     16520 4831     33213
3943     33442 4877     75412
4007     17770 4889     24410
4027     60265 4909     61113
4049     31260 4937     26525
4051     63236 4951     22271
           
4091     55650 4973     03354
4093     26670 4987     72275
4111     20437        
4133     66046 8191     03624
4157     43640      
4159     62544      
4177     16076      
4201     53211      
4219     51756      
5231     51457      

    The five least significant octal digits of the residue appear in Table 2 for each p > 3300 tested. If p (3300 < p < 5000) is omitted from Table 2 a factor of 2^p-1 is known. Some of these factors are not yet published but were communicated to the author by John Brillhart.
    My thanks to the Computing Facility for their help in this work, especially J. L. Selfridge and F. H. Hollander.

University of California at Los Angeles
Los Angeles, California

    1. D. H. LEHMER, "An extended theory of Lucas' functions," Ann. of Math. v. 31, 1930, p. 419-448.
    2. H. REISEL, "Mersenne numbers," MTAC, v. 12, 1958, p. 207-213.
    3. R. M. ROBINSON, "Mersenne and Fermat numbers," Proc. Amer. Math. Soc. v. 5, 1954, p. 842-846.
    4. R. M. ROBINSON, "A report on primes of the form k*2^n+1 and on factors of Fermat numbers," Proc. Amer. Math. Soc. v. 9, 1958, p. 673-681

Received November 3, 1964. The preparation of this paper was sponsored by the U. S. [photocopy illegible] Research.


Return to this paper's entry in Luke's Mersenne Library and Bibliography