Mersenne primes and perfect numbers are two deeply interlinked types of natural numbers in number theory. Mersenne primes, named after the friar Marin Mersenne, are prime numbers that can be expressed as 2p − 1 for some positive integer p. For example, 3 is a Mersenne prime as it is a prime number and is expressible as 22 − 1.[1][2] The numbers p corresponding to Mersenne primes must themselves be prime, although not all primes p lead to Mersenne primes—for example, 211 − 1 = 2047 = 23 × 89.[3] Meanwhile, perfect numbers are natural numbers that equal the sum of their positive proper divisors, which are divisors excluding the number itself. So, 6 is a perfect number because the proper divisors of 6 are 1, 2, and 3, and 1 + 2 + 3 = 6.[2][4]

Cuisenaire rods showing the proper divisors of 6 (1, 2, and 3) adding up to 6
Visualization of 6 as a perfect number
A graph plotting years on the x-axis with the number of digits of the largest known prime logarithmically on the y-axis, with two trendlines
Logarithmic graph of the number of digits of the largest known prime by year, nearly all of which have been Mersenne primes

There is a one-to-one correspondence between the Mersenne primes and the even perfect numbers, but it is unknown whether there exist odd perfect numbers or not. This is due to the Euclid–Euler theorem, partially proved by Euclid and completed by Leonhard Euler: even numbers are perfect if and only if they can be expressed in the form 2p − 1 × (2p − 1), where 2p − 1 is a Mersenne prime. In other words, all numbers that fit that expression are perfect, while all even perfect numbers fit that form. For instance, in the case of p = 2, 22 − 1 = 3 is prime, and 22 − 1 × (22 − 1) = 2 × 3 = 6 is perfect.[1][5][6]

It is currently an open problem as to whether there are an infinite number of Mersenne primes and even perfect numbers.[2][6] The frequency of Mersenne primes is the subject of the Lenstra–Pomerance–Wagstaff conjecture, which states that the expected number of Mersenne primes less than some given x is (eγ / log 2) × log log x, where e is Euler's number, γ is Euler's constant, and log is the natural logarithm.[7][8][9] It is also not known if any odd perfect numbers exist; various conditions on possible odd perfect numbers have been proven, including a lower bound of 101500.[10]

The following is a list of all currently known Mersenne primes and perfect numbers, along with their corresponding exponents p. As of 2023, there are 51 known Mersenne primes (and therefore perfect numbers), the largest 17 of which have been discovered by the distributed computing project Great Internet Mersenne Prime Search, or GIMPS.[2] New Mersenne primes are found using the Lucas–Lehmer test (LLT), a primality test for Mersenne primes that is efficient for binary computers.[2]

The displayed ranks are among indices currently known as of 2022; while unlikely, ranks may change if smaller ones are discovered. According to GIMPS, all possibilities less than the 48th working exponent p = 57,885,161 have been checked and verified as of January 2024.[11] The discovery year and discoverer are of the Mersenne prime, since the perfect number immediately follows by the Euclid–Euler theorem. Discoverers denoted as "GIMPS / name" refer to GIMPS discoveries with hardware used by that person. Later entries are extremely long, so only the first and last six digits of each number are shown.

More information Rank, p ...
Table of all 51 currently-known Mersenne primes and corresponding perfect numbers
Rank p Mersenne prime Mersenne prime digits Perfect number Perfect number digits Discovery Discoverer Method Ref.[12]
1 2 3 1 6 1 Ancient times[lower-alpha 1] Known to Ancient Greek mathematicians Unrecorded [13][14][15]
2 3 7 1 28 2 [13][14][15]
3 5 31 2 496 3 [13][14][15]
4 7 127 3 8128 4 [13][14][15]
5 13 8191 4 33550336 8 1200s/c. 1456[lower-alpha 2] Multiple[lower-alpha 3] Trial division [14][15]
6 17 131071 6 8589869056 10 1588[lower-alpha 2] Pietro Cataldi [2][18]
7 19 524287 6 137438691328 12 [2][18]
8 31 2147483647 10 230584...952128 19 1772 Leonhard Euler Trial division with modular restrictions [19][20]
9 61 230584...693951 19 265845...842176 37 November 1883 Ivan Pervushin Lucas sequences [21]
10 89 618970...562111 27 191561...169216 54 June 1911 Ralph Ernest Powers [22]
11 107 162259...288127 33 131640...728128 65 June 1, 1914 [23]
12 127 170141...105727 39 144740...152128 77 January 10, 1876 Édouard Lucas [24]
13 521 686479...057151 157 235627...646976 314 January 30, 1952 Raphael M. Robinson LLT on SWAC [25]
14 607 531137...728127 183 141053...328128 366 [25]
15 1,279 104079...729087 386 541625...291328 770 June 25, 1952 [26]
16 2,203 147597...771007 664 108925...782528 1,327 October 7, 1952 [27]
17 2,281 446087...836351 687 994970...915776 1,373 October 9, 1952 [27]
18 3,217 259117...315071 969 335708...525056 1,937 September 8, 1957 Hans Riesel LLT on BESK [28]
19 4,253 190797...484991 1,281 182017...377536 2,561 November 3, 1961 Alexander Hurwitz LLT on IBM 7090 [29]
20 4,423 285542...580607 1,332 407672...534528 2,663 [29]
21 9,689 478220...754111 2,917 114347...577216 5,834 May 11, 1963 Donald B. Gillies LLT on ILLIAC II [30]
22 9,941 346088...463551 2,993 598885...496576 5,985 May 16, 1963 [30]
23 11,213 281411...392191 3,376 395961...086336 6,751 June 2, 1963 [30]
24 19,937 431542...041471 6,002 931144...942656 12,003 March 4, 1971 Bryant Tuckerman LLT on IBM 360/91 [31]
25 21,701 448679...882751 6,533 100656...605376 13,066 October 30, 1978 Landon Curt Noll & Laura Nickel LLT on CDC Cyber 174 [32]
26 23,209 402874...264511 6,987 811537...666816 13,973 February 9, 1979 Landon Curt Noll [32]
27 44,497 854509...228671 13,395 365093...827456 26,790 April 8, 1979 Harry L. Nelson & David Slowinski LLT on Cray-1 [33][34]
28 86,243 536927...438207 25,962 144145...406528 51,924 September 25, 1982 David Slowinski [35]
29 110,503 521928...515007 33,265 136204...862528 66,530 January 29, 1988 Walter Colquitt & Luke Welsh LLT on NEC SX-2 [36][37]
30 132,049 512740...061311 39,751 131451...550016 79,502 September 19, 1983 David Slowinski et al. (Cray) LLT on Cray X-MP [38]
31 216,091 746093...528447 65,050 278327...880128 130,100 September 1, 1985 LLT on Cray X-MP/24 [39][40]
32 756,839 174135...677887 227,832 151616...731328 455,663 February 17, 1992 LLT on Harwell Lab's Cray-2 [41]
33 859,433 129498...142591 258,716 838488...167936 517,430 January 4, 1994 LLT on Cray C90 [42]
34 1,257,787 412245...366527 378,632 849732...704128 757,263 September 3, 1996 LLT on Cray T94 [43][44]
35 1,398,269 814717...315711 420,921 331882...375616 841,842 November 13, 1996 GIMPS / Joel Armengaud LLT / Prime95 on 90 MHz Pentium PC [45]
36 2,976,221 623340...201151 895,932 194276...462976 1,791,864 August 24, 1997 GIMPS / Gordon Spence LLT / Prime95 on 100 MHz Pentium PC [46]
37 3,021,377 127411...694271 909,526 811686...457856 1,819,050 January 27, 1998 GIMPS / Roland Clarkson LLT / Prime95 on 200 MHz Pentium PC [47]
38 6,972,593 437075...193791 2,098,960 955176...572736 4,197,919 June 1, 1999 GIMPS / Nayan Hajratwala LLT / Prime95 on IBM Aptiva with 350 MHz Pentium II processor [48]
39 13,466,917 924947...259071 4,053,946 427764...021056 8,107,892 November 14, 2001 GIMPS / Michael Cameron LLT / Prime95 on PC with 800 MHz Athlon T-Bird processor [49]
40 20,996,011 125976...682047 6,320,430 793508...896128 12,640,858 November 17, 2003 GIMPS / Michael Shafer LLT / Prime95 on Dell Dimension PC with 2 GHz Pentium 4 processor [50]
41 24,036,583 299410...969407 7,235,733 448233...950528 14,471,465 May 15, 2004 GIMPS / Josh Findley LLT / Prime95 on PC with 2.4 GHz Pentium 4 processor [51]
42 25,964,951 122164...077247 7,816,230 746209...088128 15,632,458 February 18, 2005 GIMPS / Martin Nowak [52]
43 30,402,457 315416...943871 9,152,052 497437...704256 18,304,103 December 15, 2005 GIMPS / Curtis Cooper & Steven Boone LLT / Prime95 on PC at University of Central Missouri [53]
44 32,582,657 124575...967871 9,808,358 775946...120256 19,616,714 September 4, 2006 [54]
45 37,156,667 202254...220927 11,185,272 204534...480128 22,370,543 September 6, 2008 GIMPS / Hans-Michael Elvenich LLT / Prime95 on PC [55]
46 42,643,801 169873...314751 12,837,064 144285...253376 25,674,127 June 4, 2009[lower-alpha 4] GIMPS / Odd Magnar Strindmo LLT / Prime95 on PC with 3 GHz Intel Core 2 processor [56]
47 43,112,609 316470...152511 12,978,189 500767...378816 25,956,377 August 23, 2008 GIMPS / Edson Smith LLT / Prime95 on Dell OptiPlex PC with Intel Core 2 Duo E6600 processor [55][57][58]
48 57,885,161 581887...285951 17,425,170 169296...130176 34,850,340 January 25, 2013 GIMPS / Curtis Cooper LLT / Prime95 on PC at University of Central Missouri [59][60]
* 69,129,889 Lowest unverified milestone[lower-alpha 5]
49[lower-alpha 6] 74,207,281 300376...436351 22,338,618 451129...315776 44,677,235 January 7, 2016[lower-alpha 7] GIMPS / Curtis Cooper LLT / Prime95 on PC with Intel Core i7-4790 processor [61][62]
50[lower-alpha 6] 77,232,917 467333...179071 23,249,425 109200...301056 46,498,850 December 26, 2017 GIMPS / Jonathan Pace LLT / Prime95 on PC with Intel Core i5-6600 processor [63][64]
51[lower-alpha 6] 82,589,933 148894...902591 24,862,048 110847...207936 49,724,095 December 7, 2018 GIMPS / Patrick Laroche LLT / Prime95 on PC with Intel Core i5-4590T processor [65][66]
* 120,271,181 Lowest untested milestone[lower-alpha 5]
Close

Historically, the largest known prime number has often been a Mersenne prime.

Notes

  1. The first four perfect numbers were documented by Nicomachus circa 100, and the concept was known (along with corresponding Mersenne primes) to Euclid at the time of his Elements. There is no record of discovery.
  2. Islamic mathematicians such as Ismail ibn Ibrahim ibn Fallus (1194–1239) may have known of the fifth through seventh perfect numbers prior to European records.[16]
  3. Found in an anonymous manuscript, Clm 14908, dated 1456 and 1461, and in Ibn Fallus' earlier work, which was not widely distributed[14][17]
  4. M42,643,801 was first reported to GIMPS on April 12, 2009 but was not noticed by a human until June 4, 2009 due to a server error.
  5. As of 18 September 2024[11]
  6. It has not been verified whether any undiscovered Mersenne primes exist between the 48th (M57,885,161) and the 51st (M82,589,933) on this table; the ranking is therefore provisional.
  7. M74,207,281 was first reported to GIMPS on September 17, 2015 but was not noticed by a human until January 7, 2016 due to a server error.

References

Wikiwand in your browser!

Seamless Wikipedia browsing. On steroids.

Every time you click a link to Wikipedia, Wiktionary or Wikiquote in your browser's search results, it will show the modern Wikiwand interface.

Wikiwand extension is a five stars, simple, with minimum permission required to keep your browsing private, safe and transparent.