Problems & Puzzles: Puzzles

Puzzle 74.- SOD(26972593-1)

The current largest known prime number is 26972593-1, a number supposedly composed of 2,098,960 digits.

Questions:a) Confirm the quantity of digits of 26972593-1 and b) calculate the sum of those 2,098,960 digits (explain the method/tools used). 

Hints:

1)  The expected sum of digits of 26972593-1 should be near to 45*209896 = 9445320 if we suppose that the decimal digits of this prime number appear at random, but - of course - it should not to be exactly 9445320 because then 3 would divide 26972593-1  … J

2)  Can you obtain the asked SOD adding by separate the digits in odd position and the digits in even position? In such a way you would demonstrate very easily 11 do not divide that 26972593-1 … J


Solution

Enoch Haga (7/9/99) has calculated that SOD(26972593-1) = 9440671 . How? This is his own explanation:


"I have confirmed the count by first downloading M38* and then saving to my Zip disk. From there I moved to a word processor. In Word 97 I removed commas and verified the count with the File / Properties / Statistics feature. Then I did the necessary digit counts using the Find / Replace feature, and finally multiplied to get the total SOD = 9440671...

...My counts were as follows:
1 - 210744*1 = 210744
2 - 209678*2 = 419356
3 - 209382*3 = 628146
4 - 209832*4 = 839
328
5 - 209863*5= 1049315
6 - 210356*6= 1262136
7 - 209314*7= 1465198
8 - 209961*8= 1679688
9 - 209640*9= 1886760
0 - 210190*0= 0

Totals -...... 2098960 & 9440671"

Note: * http://reality.sgi.com/chongo/tech/math/prime/merdigit/index.html (detected broken 1/9/01)

Can you confirm this results not using any published result of M38?

***
Landon Curt Noll and - independently - Jim Howell confirmed the Enoch's results. Howell also has calculated "SOD adding by separate the digits in odd position and the digits in even position as was asked in the hint 2. Here are their contributions in detail.

Landon Curt Noll wrote:

"The known Mersenne primes may be found at:

        http://reality.sgi.com/chongo/tech/math/prime/merdigit/index.html

The decimal values were computed using calc (I ran calc on my laptop PC which is runing Linux):

        http://reality.sgi.com/chongo/tech/comp/calc/index.html

and the English names were computed (with input from calc) using number:

        http://reality.sgi.com/chongo/tech/math/number/number.html
        http://reality.sgi.com/chongo/cgi-bin/number.cgi

The decimal digits of the M38 number may be obtained (without commas or dots)
from the URL:
 http://reality.sgi.com/chongo/tech/math/prime/merdigit/m6972593/prime-l.html

The answer to the 'sum of the decimal digits' depends on your exact meaning.
If you are asking. What is the decimal sum of the decimal digits of 2^6972593-1? Then the answer may be obtained thru one of two ways:

method 1

    Go to the above URL (http:// ... /m6972593/prime-l.html)

    Save the URL into a file called prime-l.html

    extract the prime from the file (strip off the HTML header):

        head -29 prime-l.html | tail -1 > /tmp/prime

    Verify that you have the right sized file and it starts/ends with
    the correct digits (to be sure your browser download the
    entire number):

        wc /tmp/prime
        1       1 2098961 /tmp/prime

        cut -c1-10 /tmp/prime
        4370757441

        cut -c2098951-2098960 /tmp/prime
        2924193791

    Use a perl script to sum the digits:

        perl -e '@dig = split(//,<>); for ($sum=0,$i=0; $i<$#dig; ++$i) {$sum += $dig[$i];} print "$sum\n"' < /tmp/prime
        9440671

The output of the above perl command, 9440671, is your answer.

method 2

Note that using calc and perl, one could both compute the decimal value
and print the answer:

        calc -p '2^6972593-1' | perl -e '@dig = split(//,<>); for ($sum=0,$i=0; $i<$#dig; ++$i) {$sum += $dig[$i];} print "$sum\n"'
        9440671

I like the method #2 because it uses a combination of calc and perl.
Each tool is being used to do what it does best.  Last, you can obtain
the answer on a single command line. But for those who do not have calc, or have a slow computer, method 1 (using the URL containing the value pre-calculated by calc) works just
as well
".

***

Jim Howell wrote:

" I wrote a program, using my own big-arithmetic routines, which computes 2^6972593-1 and writes the number to a disk file.  Then a second program to read the file, and add all the digits, as well as count the number of digits.  The output is:

Number of digits = 2098960
Sum of digits =    9440671

Sum of 1st, 3rd, 5th etc. digits = 4722301
Sum of 2nd, 4th, 6th etc. digits = 4718370
Difference =                       3931

... Since the difference of these two sums (3931) is not divisible by 11, the
original number is also not divisible by 11...

My big arithmetic routines are written in C, with a small amount of assembler code.  The program to read the file and add the digits was written in C also.
On my Pentium-III/450, it takes about 8 minutes to compute 2^6972593-1 and about 1.5 hours to convert it from binary to decimal.  Adding and counting the digits takes only a few seconds
".

***

Enoch also counted the now (17/12/2001) supposed M39 :2^13466917-1:

I get 4053946 digits and a sum of digits of 18243946

0 - 405083*0=0
1 - 405614*1=405614
2 - 405068*2=810136
3 - 405928*3=1217784
4 - 405491*4=1621964
5 - 404915*5=2024575
6 - 405154*6=2430924
7 - 405308*7=2837156
8 - 406672*8=3253376
9 - 404713*9=3642417
T - 4053946 - 18243946

***

Andrew Rupinski updated these calculations (March 12, 2005):

I notice on Puzzle 74 that Enoch Haga updated the original question to address M13466917. Since 3 additional Mersennes have since been discovered, including one just a few weeks ago, I decided to go ahead and compute their digit counts if you are interested. They are as follows (the expected sum is
4.5 x #digits, the sum if all digits were equidistributed):

M20996011
0 - 631705
1 - 632720
2 - 630989
3 - 631467
4 - 632004
5 - 633283
6 - 630929
7 - 633503
8 - 632964
9 - 630866
Total: 6320430
SOD(M20996011) = 28445131
Expected sum = 28441935

M24036583
0 - 722613
1 - 723188
2 - 722754
3 - 722181
4 - 723758
5 - 724196
6 - 723856
7 - 724543
8 - 723551
9 - 725093
Total: 7235733
SOD(M24036583) = 32580433
Expected sum = 32560798.5

M25964951
0 - 782138
1 - 782118
2 - 781551
3 - 781856
4 - 780817
5 - 781588
6 - 780774
7 - 781662
8 - 781424
9 - 782302
Total: 7816230
SOD(M25964951) = 35170384
Expected sum = 35173035

***

 


Records   |  Conjectures  |  Problems  |  Puzzles