### 5 April 2005

<ph f="cmbx">The Carmichael numbers up to </ph> <math xmlns="http://www.w3.org/1998/Math/MathML"> <mn>1017</mn> </math>

### Richard G.E. Pinch

2 Eldon Road, Cheltenham, Glos GL52 6TU, U.K. E-mail address : rgep@chalcedon.demon.co.uk

• Abstract. We extend our previous computations to show that there are 585355 Carmichael numbers up to ${10}^{17}$  . As before, the numbers were generated by a back-tracking search for possible prime factorisations together with a “large prime variation”. We present further statistics on the distribution of Carmichael numbers.

1 Introduction

A Carmichael number $N$  is a composite number $N$  with the property that for every $b$  prime to $N$  we have ${b}^{N-1}\equiv 1modN$  . It follows that a Carmichael number $N$  must be square-free, with at least three prime factors, and that $p-1|N-1$  for every prime $p$  dividing $N$  : conversely, any such $N$  must be a Carmichael number.
For background on Carmichael numbers and details of previous computations we refer to our previous paper [4: in that paper we described the computation of the Carmichael numbers up to ${10}^{15}$  and presented some statistics. These computations have since been extended to ${10}^{16}$  [5and now to ${10}^{17}$  , using similar techniques, and we present further statistics.
The complete list of Carmichael numbers up to ${10}^{17}$  is available from the author.

2 Organisation of the search

We used improved versions of strategies described in [4.
The principal search was a depth-first back-tracking search over possible sequences of primes factors ${p}_{1},\dots ,{p}_{d}$  . Put ${P}_{r}={\prod }_{i=1}^{r}{p}_{i}$  , ${Q}_{r}={\prod }_{i=r+1}^{d}{p}_{i}$  and ${L}_{r}=lcm\left\{{p}_{i}-1:i=1,\dots ,r\right\}$  . We find that ${Q}_{r}$  must satisfy the congruence $N={P}_{r}{Q}_{r}\equiv 1mod{L}_{r}$  and so in particular ${Q}_{d}={p}_{d}$  must satisfy a congruence modulo ${L}_{d-1}$  : further ${p}_{d}-1$  must be a factor of ${P}_{d-1}-1$  . We modified this to terminate the search early at some level $r$  if the modulus ${L}_{r}$  is large enough to limit the possible values of ${Q}_{r}$  , which may then be factorised directly.
We also employed the variant based on proposition 2 of [4which determines the finitely many possible pairs $\left({p}_{d-1},{p}_{d}\right)$  from ${P}_{d-2}$  . In practice this was useful only when $d=3$  allowing us to determine the complete list of Carmichael numbers with three prime factors up to ${10}^{18}$  . The count of these is 35586 (not 35585 as incorrectly reported in our preliminary work).
Finally we employed a different search over large values of ${p}_{d}$  , in the range ${10}^{7}<{p}_{d}<{10}^{9}$  , using the property that ${P}_{d-1}\equiv 1mod\left({p}_{d}-1\right)$  and factorising the numbers in this arithmetic progression less than ${10}^{18}/{p}_{d}$  .

3 Statistics

 $n$ $C\left({10}^{n}\right)$ 3 1 4 7 5 16 6 43 7 105 8 255 9 646 10 1547 11 3605 12 8241 13 19279 14 44706 15 105212 16 246683 17 585355
Table 1 . Distribution of Carmichael numbers up to ${10}^{17}$  .
 $X$ 3 4 5 6 7 8 9 10 11 total 3 1 0 0 0 0 0 0 0 0 1 4 7 0 0 0 0 0 0 0 0 7 5 12 4 0 0 0 0 0 0 0 16 6 23 19 1 0 0 0 0 0 0 43 7 47 55 3 0 0 0 0 0 0 105 8 84 144 27 0 0 0 0 0 0 255 9 172 314 146 14 0 0 0 0 0 646 10 335 619 492 99 2 0 0 0 0 1547 11 590 1179 1336 459 41 0 0 0 0 3605 12 1000 2102 3156 1714 262 7 0 0 0 8241 13 1858 3639 7082 5270 1340 89 1 0 0 19279 14 3284 6042 14938 14401 5359 655 27 0 0 44706 15 6083 9938 29282 36907 19210 3622 170 0 0 105212 16 10816 16202 55012 86696 60150 16348 1436 23 0 246683 17 19539 25758 100707 194306 172234 63635 8835 340 1 585355
Table 2 . Values of $C\left(X\right)$  and $C\left(d,X\right)$  for $d\le 10$  and $X$  in powers of 10 up to ${10}^{17}$  .
We have shown that there are 585355 Carmichael numbers up to ${10}^{17}$  , all with at most 11 prime factors. We let $C\left(X\right)$  denote the number of Carmichael numbers less than $X$  and $C\left(d,X\right)$  denote the number with exactly $d$  prime factors. Table  1 gives the values of $C\left(X\right)$  and Table  2 the values of $C\left(d,X\right)$  for $X$  in powers of 10 up to ${10}^{17}$  .
We have used the same methods to calculate the smallest Carmichael number ${S}_{d}$  with $d$  prime factors for $d$  up to 34. The results are given in Table  3 . Heuristics suggest that a good approximation for $log{S}_{d}$  might be $2log2\left(d-1\right)log\left(d-1\right)$  .
There is some support for this from Table  4 .
 $d$ $N$ 3 561 4 41041 5 825265 6 321197185 7 5394826801 8 232250619601 9 9746347772161 10 1436697831295441 11 60977817398996785 12 7156857700403137441 13 1791562810662585767521 14 87674969936234821377601 15 6553130926752006031481761 16 1590231231043178376951698401 17 35237869211718889547310642241 18 32809426840359564991177172754241 19 2810864562635368426005268142616001 20 349407515342287435050603204719587201 21 125861887849639969847638681038680787361 22 12758106140074522771498516740500829830401 23 2333379336546216408131111533710540349903201 24 294571791067375389885907239089503408618560001 25 130912961974316767723865201454187955056178415601 26 13513093081489380840188651246675032067011140079201 27 7482895937713262392883306949172917048928068129206401 28 1320340354477450170682291329830138947225695029536281601 29 379382381447399527322618466130154668512652910714224209601 30 70416887142533176417390411931483993124120785701395296424001 31 2884167509593581480205474627684686008624483147814647841436801 32 4754868377601046732119933839981363081972014948522510826417784001 33 1334733877147062382486934807105197899496002201113849920496510541601 34 260849323075371835669784094383812120359260783810157225730623388382401
Table 3 . The smallest Carmichael numbers with $d$  prime factors for $d$  up to 34.
 $d$ $\frac{log\left({S}_{d}\right)}{2log2\left(d-1\right)log\left(d-1\right)}$ 3 3.293621187 4 2.324869052 5 1.772215418 6 1.755823184 7 1.503593258 8 1.385943139 9 1.296862582 10 1.273113126 11 1.210794211 12 1.187291825 13 1.183842403 14 1.142841324 15 1.115637251 16 1.112255293 17 1.068846695 18 1.086833722 19 1.067861876 20 1.055266652 21 1.056211783 22 1.041906608 23 1.034833298 24 1.024202006 25 1.026042173 26 1.014073506 27 1.017122489 28 1.010169037 29 1.007224804 30 1.000945160 31 0.984182030 32 0.993535534 33 0.990337138 34 0.984854273
Table 4 . The smallest Carmichael number with $d$  prime factors ${S}_{d}$  compared to $2log2\left(d-1\right)log\left(d-1\right)$  .

Figure 1 . $k\left(X\right)$  versus $X$  (expressed in powers of 10).

 $n$ $k\left({10}^{n}\right)$ $C\left({10}^{n}\right)/C\left({10}^{n-1}\right)$ 3 2.93319 4 2.19547 7.000 5 2.07632 2.286 6 1.97946 2.688 7 1.93388 2.441 8 1.90495 2.429 9 1.87989 2.533 10 1.86870 2.396 11 1.86421 2.330 12 1.86377 2.286 13 1.86240 2.339 14 1.86293 2.319 15 1.86301 2.353 16 1.86406 2.335 17 1.86472 2.373
Table 5 . The function $k\left(X\right)$  and growth of $C\left(X\right)$  for $X={10}^{n}$  , $n\le 17$  .
In Table  5 and Figure  1 we tabulate the function $k\left(X\right)$  , defined by Pomerance, Selfridge and Wagstaff [6by $C\left(X\right)=Xexp\left(-k\left(X\right)\frac{logXlogloglogX}{loglogX}\right).$  They proved that $liminfk\ge 1$  and suggested that $limsupk$  might be $2$  , although they also observed that within the range of their tables $k\left(X\right)$  is decreasing: Pomerance [7,[8gave a heuristic argument suggesting that $limk=1$  . The decrease in $k$  is reversed between ${10}^{13}$  and ${10}^{14}$  : see Figure  1 . We find no clear support from our computations for any conjecture on a limiting value of $k$  .
In Table  5 we also give the ratios $C\left({10}^{n}\right)/C\left({10}^{n-1}\right)$  investigated by Swift [10.
Swift's ratio, again initially decreasing, also increases again before ${10}^{15}$  .

Figure 2 . $C\left(X\right)$  as a power of $X$  (expressed in powers of 10).

 $n$ $logC\left({10}^{n}\right)/\left(nlog10\right)$ 4 .21127 5 .24082 6 .27224 7 .28874 8 .30082 9 .31225 10 .31895 11 .32336 12 .32633 13 .32962 14 .33217 15 .33480 16 .33700 17 .33926
Table 6 . $C\left(X\right)$  as a power of $X$  .
In Table  6 and Figure  2 we see that within the range of our computations $C\left(X\right)$  is a slowly growing power of $X$  : about ${X}^{0.339}$  for $X$  around ${10}^{17}$  .
In Table  7 we give the number of Carmichael numbers in each class modulo $m$  for $m=5$  , 7, 11 and 12.
In Tables  8 and  9 we give the number of Carmichael numbers divisible by primes $p$  up to 97. In Table  8 we count all Carmichael numbers divisible by $p$  : in Table  9 we count only those for which $p$  is the smallest prime factor.
The largest prime factor of a Carmichael number up to ${10}^{17}$  is $p=223401361$  , dividing $n=99816335969903281=17\cdot 31\cdot 379\cdot 2237\cdot p$  . We have $p={n}^{0.4911}$  :
indeed $n=p\left(2p-1\right)$  and $n-1=\left(p-1\right)\left(2p+1\right)$  .
The largest prime to occur as the smallest prime factor of a Carmichael number in this range is 380251, dividing $90256390764228001=380251\cdot 410761\cdot 577981.$  We note that this number is of the form $\left(25k+1\right)\left(27k+1\right)\left(38k+1\right)$  with $k=15210$  .
We define the index of a Carmichael number $N$  to be the integer $i\left(N\right)=\left(N-1\right)/\lambda \left(N\right)$  where $\lambda$  is the Carmichael function, the exponent of the multiplicative group. A result of Somer [9implies that $i\left(N\right)\to \infty$  as $N$  runs through Carmichael numbers. In Table  10 we list the Carmichael numbers known to have index less than 100: the list is complete for numbers up to ${10}^{17}$  .
We define the Lehmer index of a number $N$  to be $\ell \left(N\right)=\left(N-1\right)/\phi \left(N\right)$  . Lehmer's totient problem asks whether there is a number of integral Lehmer index greater than 1 (that is, whether there is a composite $N$  such that $\phi \left(N\right)$  divides $N-1$  ).
Such a number would have to be a Carmichael number. See Guy [2 §B37 for further details.
The Carmichael numbers up to ${10}^{17}$  with Lehmer index $\ge 2$  are listed in Table  11 .
 $m$ $c$ ${25.10}^{9}$ ${10}^{11}$ ${10}^{12}$ ${10}^{13}$ ${10}^{14}$ ${10}^{15}$ ${10}^{16}$ ${10}^{17}$ 5 0 203 312 627 1330 2773 5814 12200 25523 1 1652 2785 6575 15755 37467 90167 215713 520990 2 82 154 327 702 1484 3048 6094 12912 3 102 172 344 725 1463 3059 6285 12845 4 124 182 368 767 1519 3124 6391 13085 7 0 401 634 1334 2774 5891 12691 27550 60602 1 1096 1885 4613 11447 28001 69131 168856 414999 2 105 186 432 967 2109 4599 10011 21961 3 152 232 496 1055 2178 4707 10039 21918 4 129 211 450 985 2122 4592 9944 21832 5 138 222 454 1033 2224 4777 10125 21911 6 142 235 462 1018 2181 4715 10158 22132 11 0 335 547 1324 3006 7032 16563 38576 90731 1 640 1131 2770 6786 16548 40891 100071 248840 2 139 217 473 1068 2361 5338 12054 27268 3 142 220 457 1045 2348 5319 12186 27483 4 104 187 442 1026 2317 5261 11917 27276 5 152 243 466 1066 2370 5316 12194 27670 6 116 198 440 1061 2400 5384 12155 27362 7 122 195 458 1023 2223 5165 11853 26960 8 129 222 475 1107 2450 5449 12012 27639 9 131 218 465 1042 2285 5179 11835 27158 10 153 227 471 1049 2372 5347 11830 26968 12 1 2071 3462 7969 18761 43760 103428 243382 579215 3 0 0 1 2 2 5 5 7 5 20 32 64 124 228 448 805 1470 7 47 75 147 289 547 1027 1906 3589 9 25 36 60 103 165 294 560 1018 11 0 0 0 0 4 10 25 56
Table 7 . The number of Carmichael numbers in each class modulo $m$  for $m=5$  , 7, 11 and 12.
 $p$ ${25.10}^{9}$ ${10}^{11}$ ${10}^{12}$ ${10}^{13}$ ${10}^{14}$ ${10}^{15}$ ${10}^{16}$ ${10}^{17}$ 3 25 36 61 105 167 299 565 1025 5 203 312 627 1330 2773 5814 12200 25523 7 401 634 1334 2774 5891 12691 27550 60602 11 335 547 1324 3006 7032 16563 38576 90731 13 483 807 1784 3998 9045 20758 47785 110320 17 293 489 1182 2817 6640 16019 38302 91783 19 372 608 1355 3345 7797 18638 44389 106273 23 113 207 507 1282 3135 7716 18867 46612 29 194 336 832 2094 5158 12721 31110 76647 31 335 571 1320 3086 7270 17382 41440 99845 37 320 535 1270 2926 6826 16220 38647 92744 41 227 390 1001 2418 5896 14344 34759 84977 43 184 296 772 1920 4663 11594 28650 70904 47 53 80 199 492 1223 2873 6810 17002 53 92 160 351 813 2041 5143 12256 30560 59 26 41 92 262 644 1611 3959 9613 61 269 453 1075 2542 6047 14429 34503 83268 67 110 178 407 1063 2540 6306 15295 37521 71 104 194 521 1320 3351 8546 21485 53857 73 198 348 849 2145 4925 11929 29072 71137 79 64 107 247 686 1728 4318 10693 26766 83 14 24 56 137 340 838 1929 4640 89 68 131 320 788 1951 4981 12178 30737 97 123 193 495 1277 3123 7594 18706 45542
Table 8 . Primes occurring in Carmichael numbers.
 $p$ ${25.10}^{9}$ ${10}^{11}$ ${10}^{12}$ ${10}^{13}$ ${10}^{14}$ ${10}^{15}$ ${10}^{16}$ ${10}^{17}$ 3 25 36 61 105 167 299 565 1025 5 202 309 624 1325 2765 5797 12175 25481 7 364 579 1218 2557 5461 11874 25915 57459 11 263 428 1071 2509 5979 14397 33893 80745 13 237 431 1058 2462 5699 13514 32025 76256 17 117 206 496 1318 3244 8114 20206 50170 19 152 244 532 1401 3358 8141 20020 49413 23 37 78 207 535 1360 3317 8195 20803 29 55 103 284 729 1822 4659 11577 29149 31 101 168 390 876 2116 5153 12575 30667 37 60 95 219 551 1401 3418 8594 21382 41 35 68 171 414 1092 2736 6788 17275 43 35 65 168 403 943 2308 5520 13636 47 14 16 36 81 195 459 1135 2854 53 19 30 55 147 363 973 2327 5842 59 2 4 11 43 100 272 618 1542 61 34 58 148 364 851 1978 4722 11278 67 8 18 50 123 317 815 1950 4843 71 15 25 66 161 389 979 2480 6178 73 14 28 68 175 406 1015 2508 6277 79 4 10 17 66 175 467 1163 2873 83 1 1 4 8 39 79 175 457 89 10 16 23 55 148 409 1003 2523 97 10 20 50 106 261 606 1413 3445
Table 9 . Primes occurring as least prime factor in Carmichael numbers.
 $i$ $N$ factors 5 6601 $7\cdot 23\cdot 41$ 7 561 $3\cdot 11\cdot 17$ 18 42018333841 $11\cdot 47\cdot 1049\cdot 77477$ 18 55462177 $17\cdot 23\cdot 83\cdot 1709$ 18 8885251441 $11\cdot 47\cdot 1109\cdot 15497$ 21 10585 $5\cdot 29\cdot 73$ 22 2465 $5\cdot 17\cdot 29$ 23 1105 $5\cdot 13\cdot 17$ 25 11921001 $3\cdot 29\cdot 263\cdot 521$ 31 62745 $3\cdot 5\cdot 47\cdot 89$ 37 11972017 $43\cdot 433\cdot 643$ 37 67902031 $43\cdot 271\cdot 5827$ 39 334153 $19\cdot 43\cdot 409$ 43 52633 $7\cdot 73\cdot 103$ 44 15841 $7\cdot 31\cdot 73$ 45 8911 $7\cdot 19\cdot 67$ 47 2821 $7\cdot 13\cdot 31$ 48 1729 $7\cdot 13\cdot 19$ 49 1208361237478669 $53\cdot 653\cdot 26479\cdot 1318579$ 50 4199932801 $29\cdot 499\cdot 503\cdot 577$ 52 206955841 $17\cdot 71\cdot 277\cdot 619$ 53 1271325841 $17\cdot 31\cdot 179\cdot 13477$ 54 4169867689 $13\cdot 29\cdot 383\cdot 28879$ 55 271794601 $13\cdot 19\cdot 743\cdot 1481$ 60 6840001 $7\cdot 17\cdot 229\cdot 251$ 61 1962804565 $5\cdot 103\cdot 149\cdot 25579$ 67 11985924995083901 $29\cdot 101\cdot 1427\cdot 16349\cdot 175403$ 67 410041 $41\cdot 73\cdot 137$ 70 162401 $17\cdot 41\cdot 233$ 76 4752717761 $11\cdot 17\cdot 107\cdot 173\cdot 1373$ 81 35575075809505 $5\cdot 197\cdot 223\cdot 353\cdot 458807$ 82 1496405933740345 $5\cdot 47\cdot 317\cdot 40253\cdot 499027$ 83 142159958924185 $5\cdot 37\cdot 107\cdot 58379\cdot 123017$ 83 158664761899885 $5\cdot 37\cdot 107\cdot 53987\cdot 148469$ 83 204370370140285 $5\cdot 37\cdot 107\cdot 48677\cdot 212099$ 83 24831908105124205 $5\cdot 29\cdot 719\cdot 3023\cdot 78790717$ 85 171189355538562901 $23\cdot 71\cdot 983\cdot 1031\cdot 103437589$ 90 3778118040573702001 $11\cdot 47\cdot 1051\cdot 67967\cdot 102302009$ 92 520178982961 $11\cdot 29\cdot 131\cdot 607\cdot 20507$ 94 12782849065 $5\cdot 7\cdot 269\cdot 317\cdot 4283$ 95 8956911601 $11\cdot 17\cdot 127\cdot 131\cdot 2879$ 97 5472940991761 $199\cdot 241\cdot 863\cdot 132233$ 97 721574219707441 $167\cdot 241\cdot 5039\cdot 3557977$ 97 83565865434172201 $103\cdot 1993\cdot 9551\cdot 42622169$ 97 9729822470631481 $127\cdot 409\cdot 110681\cdot 1692407$ 99 438253965870337 $43\cdot 139\cdot 49409\cdot 1484009$
Table 10 . Carmichael numbers with small index
 $\ell$ $N$ factors 2.00163 3852971941960065 $3\cdot 5\cdot 23\cdot 89\cdot 113\cdot 1409\cdot 788129$ 2.00388 655510549443465 $3\cdot 5\cdot 23\cdot 53\cdot 389\cdot 2663\cdot 34607$ 2.00837 13462627333098945 $3\cdot 5\cdot 23\cdot 53\cdot 197\cdot 8009\cdot 466649$ 2.01001 26708253318968145 $3\cdot 5\cdot 17\cdot 113\cdot 57839\cdot 16025297$ 2.03067 269040992399565 $3\cdot 5\cdot 23\cdot 29\cdot 4637\cdot 5799149$ 2.03207 158353658932305 $3\cdot 5\cdot 17\cdot 89\cdot 149\cdot 563\cdot 83177$ 2.03614 1817671359979245 $3\cdot 5\cdot 23\cdot 29\cdot 359\cdot 11027\cdot 45893$ 2.07138 16057190782234785 $3\cdot 5\cdot 17\cdot 29\cdot 269\cdot 6089\cdot 1325663$ 2.07294 75131642415974145 $3\cdot 5\cdot 23\cdot 29\cdot 53\cdot 617\cdot 9857\cdot 23297$ 2.08129 881715504450705 $3\cdot 5\cdot 17\cdot 47\cdot 89\cdot 113\cdot 503\cdot 14543$ 2.08381 31454143858820145 $3\cdot 5\cdot 17\cdot 23\cdot 2129\cdot 39293\cdot 64109$ 2.08894 6128613921672705 $3\cdot 5\cdot 17\cdot 23\cdot 353\cdot 7673\cdot 385793$ 2.09841 12301576752408945 $3\cdot 5\cdot 23\cdot 29\cdot 53\cdot 113\cdot 197\cdot 1042133$ 2.11432 1886616373665 $3\cdot 5\cdot 17\cdot 23\cdot 83\cdot 353\cdot 10979$ 2.11493 3193231538989185 $3\cdot 5\cdot 17\cdot 23\cdot 113\cdot 167\cdot 2927\cdot 9857$ 2.13936 11947816523586945 $3\cdot 5\cdot 17\cdot 23\cdot 89\cdot 113\cdot 233\cdot 617\cdot 1409$
Table 11 . Carmichael numbers with Lehmer index greater than 2
References

1. Jean-Marie De Koninck and Claude Levesque (eds.), Number theory: proceedings of the international number theory conference, Université de Laval, 1987, Berlin, Walter de Gruyter, 1989.
2. Richard K. Guy, Unsolved problems in number theory, second ed., Unsolved problems in intuitive mathematics, vol. 1, Springer–Verlag, New York, 1994.
3. R.A. Mollin (ed.), Number theory and its applications, Dordrecht, Kluwer Academic, 1989, Proceedings of the NATO Advanced Study Institute on Number Theory and Applications.
4. Richard G.E. Pinch, The Carmichael numbers up to ${10}^{15}$  , Math. Comp. 61 (1993), 381–391, Lehmer memorial issue.
5. , The Carmichael numbers up to ${10}^{16}$  , March 1998, arXiv:math.NT/9803082.
6. C. Pomerance, J.L. Selfridge, and S.S. Wagstaff jr, The pseudoprimes up to ${25.10}^{9}$  , Math. Comp. 35 (1980), no. 151, 1003–1026.
7. Carl Pomerance, On the distribution of pseudoprimes, Math. Comp. 37 (1981), 587–593.
8. , Two methods in elementary analytic number theory, in Mollin [3, Proceedings of the NATO Advanced Study Institute on Number Theory and Applications, pp. 136–161.
9. Lawrence Somer, On Fermat $d$  -pseudoprimes, in De Koninck and Levesque [1, pp. 841–860.
10. J.D. Swift, Review 13[9] — table of Carmichael numbers to ${10}^{9}$  , Math. Comp. 29 (1975), 338–339.

2 Eldon Road, Cheltenham, Glos GL52 6TU, U.K. E-mail address : rgep@chalcedon.demon.co.uk