|
||||
Title: A Beastly Number Post by ThudanBlunder on Nov 26th, 2008, 9:05am a) Find the first 6 digits of (10666)! b) How many trailing zeros does the above number have? |
||||
Title: Re: A Beastly Number Post by Barukh on Dec 2nd, 2008, 10:32am For an easier part b), I get: [hide]2664(5667 - 1)[/hide] |
||||
Title: Re: A Beastly Number Post by Barukh on Dec 3rd, 2008, 12:31am According to the following article (http://www.pims.math.ca/pi/issue7/page10-12.pdf), solving part a) may require calculation of a certain logarithm to a precision of more than 670 decimal digits! :o |
||||
Title: Re: A Beastly Number Post by ThudanBlunder on Dec 4th, 2008, 6:56pm on 12/02/08 at 10:32:03, Barukh wrote:
If we take n/4 as an estimate for the number of trailing zeros, we get 2.5*10665 Your number is 5 times this. |
||||
Title: Re: A Beastly Number Post by Barukh on Dec 5th, 2008, 12:15pm on 12/04/08 at 18:56:10, ThudanBlunder wrote:
Yes, I should've written 5666 instead. But now I realized the answer is incorrect anyway (doesn't take into account fractions). To write the answer in a "compact form" may be as difficult as part a) then... |
||||
Title: Re: A Beastly Number Post by SMQ on Dec 5th, 2008, 12:32pm http://www.ocf.berkeley.edu/~wwu/YaBBImages/symbols/subinfty.gif 2) http://www.ocf.berkeley.edu/~wwu/YaBBImages/symbols/sum.gif http://www.ocf.berkeley.edu/~wwu/YaBBImages/symbols/lfloor.gif10666/5nhttp://www.ocf.berkeley.edu/~wwu/YaBBImages/symbols/rfloor.gif n=1 Where in practice the upper bound can be reduced to http://www.ocf.berkeley.edu/~wwu/YaBBImages/symbols/lfloor.gif666 log 10 / log 5http://www.ocf.berkeley.edu/~wwu/YaBBImages/symbols/rfloor.gif = 952
--SMQ |
||||
Title: Re: A Beastly Number Post by Barukh on Dec 6th, 2008, 9:52am SMQ, you are right. My answer is wrong, since it doesn't take into account the second term in your formula - the sum which is challenging to evaluate. After working it out, I get the following answer to b): 26645666 - 143. ::) I will supply details later, after I find the answer to the first question. |
||||
Title: Re: A Beastly Number Post by Eigenray on Dec 6th, 2008, 12:25pm on 12/06/08 at 09:52:06, Barukh wrote:
So is there a clever way to compute [hide]the sum of the digits of 2666 = 34004...233245[/hide]? Actually, the approximation [hide]333*log5(2)[/hide] is pretty good here. |
||||
Title: Re: A Beastly Number Post by Barukh on Dec 7th, 2008, 6:27am on 12/06/08 at 12:25:31, Eigenray wrote:
I don't know. I used high-precision software (MPFR) to compute the number (I still want to get a confirmation it's correct). Quote:
Yes, but it may be quite inaccurate for other exponents. In the attached graph, I plotted the discrepancies between your approximation and actual number for all cases 10n with 10 < n < 2000. Again, if my calculations are correct. |
||||
Title: Re: A Beastly Number Post by Barukh on Dec 7th, 2008, 11:21pm I get the following answer to part a): 13407273847... |
||||
Title: Re: A Beastly Number Post by Eigenray on Dec 8th, 2008, 7:57am on 12/07/08 at 23:21:10, Barukh wrote:
Are you sure it's not 134072738469787? But the first 6 digits are correct :) And the 143 is correct, as the following short (but rather inefficient) program shows: Code:
|
||||
Title: Re: A Beastly Number Post by Barukh on Dec 8th, 2008, 10:19am on 12/08/08 at 07:57:32, Eigenray wrote:
Hmm... I did calculate the logarithm with very high precision, so at least 15 digits should be accurate. Then, I did exponentiation with a double precision. Could it be I lost 5 digits of accuracy there? ??? What's your method? |
||||
Title: Re: A Beastly Number Post by Eigenray on Dec 8th, 2008, 12:44pm Code:
What do you get? Edit: And here it is with [link=http://www.mpfr.org/]MPFR[/link]: Code:
Gives 1.3407273846978712508 (I already had GMP but apparently it doesn't do logs. So I downloaded the MPFR sources and started ./configure; make. Then I realized I could install it through Cygwin, and wrote the above before it finished compiling.) |
||||
Title: Re: A Beastly Number Post by ThudanBlunder on Dec 9th, 2008, 4:03pm My source (http://books.google.co.uk/books?id=52N0JJBspM0C&pg=PA350&dq=leviathan+1735+n!&ei=zAc_SdD1AqaGzgTPnbnADg)for this problem got yet a different answer, 1.340727397... |
||||
Title: Re: A Beastly Number Post by Eigenray on Dec 9th, 2008, 5:32pm The funny thing is that he did compute the fractional part of 10666/log(10) correctly to 14 digits. But he decided to round it to 8 digits before exponentiating, and then claim 10 digits of accuracy in the result: Quote:
which is silly. |
||||
Powered by YaBB 1 Gold - SP 1.4! Forum software copyright © 2000-2004 Yet another Bulletin Board |