1
$\begingroup$

I am trying to calculate the value of $b=\dfrac{2^{3^{4^5}}}{e^{10240}}$. Is there any method to solve this efficiently?

$\endgroup$
4
  • $\begingroup$ Do you prioritize the exponents or the base in the power-tower? $\endgroup$ Sep 1, 2014 at 18:04
  • $\begingroup$ yes, the exponents is prioritize top down, that is 2^(3^(4^5)). I want to find out a numerical method to calculate the value with enough accuracy. $\endgroup$
    – magicworks
    Sep 1, 2014 at 18:11
  • 1
    $\begingroup$ Is there more context for this problem? Is this homework for a class or a question you have made up? $\endgroup$
    – graydad
    Sep 1, 2014 at 18:28
  • $\begingroup$ This comes from a students in physics class and he asked me this. He said he want to estimate the number of electrons from an emission. Actually I don't know very clearly how this number is derived. So I post here to see if anyone can give me some help. Thanks! $\endgroup$
    – magicworks
    Sep 2, 2014 at 5:39

2 Answers 2

5
$\begingroup$

As you can see WolframAlpha can give you an approximative power of 10 representation: $$10^{10^{10^{2.688465006395752}}}.$$

It is hard to get a better solution, because this number is too large. To show it, we will calculate the number of digits in base-$10$ representation. If $n$ is a number in base-$b$, with $d$ digits, then we know, that the formula for $d$ is the following: $$d=\lfloor\log_{b}n\rfloor+1.$$ Now let $$n:=\dfrac{2^{3^{4^5}}}{e^{10240}}.$$ Then the number of digits of $n$ in base-$10$ is $$d = \lfloor \log_{10} \dfrac{2^{3^{4^5}}}{e^{10240}}\rfloor+1 \\ d = \lfloor \log_{10} 2^{3^{4^5}} - \log_{10} e^{10240}\rfloor+1 \\ d = \lfloor 3^{4^5} \cdot \log_{10} 2 - 10240 \cdot \log_{10} e\rfloor+1.$$

We get the following for $d$:

$1124021466074751860097567522104789648012545442387518261576295420518 \\5174447660807915950553426138321488548657928846792570107753324167422 \\0100211779337007726069891241143955562498313809792175538301726950271 \\2513614070749429457547548532111853109636637797579524872247171419406 \\3487219461564568491620652987627661309480232956516340085351404053765 \\2037205369420431855146383193275981445894731731211119067826441631620 \\7609542700946643046958255703325110043123352486373327969799306832787 \\29227794366058969345$

So $n$ has $1.2402146607475186 \cdot 10^{489}$ digits.

You can easily check the result with Maple:

floor(3^(4^5)*log10(2)-10240*log10(exp(1)))+1;

This $n$ number is larger then a googleplex, which number also doesn't have a base-$10$ representation, because that is also too large. On the other hand this is less then Skewes' number and much much less then Graham's number. Read more at Large numbers wikipedia article.

The problem is that $2^{3^{4^5}}$ is too large. It is much larger then the denominator. The numerator and the whole fraction have just $4$ order of magnitude difference in base-$10$, the difference of the lengths is $4447$. The above give approximation has $1.124020417 \cdot 10^{489}$ digits, so the real value and the approximation has $1.049 \cdot 10^{483}$ difference in the numbers of digits.

$\endgroup$
1
  • 1
    $\begingroup$ Thank you user153012. After your analysis I realize the number is too large to be estimated with desired accuracy. Thank you! $\endgroup$
    – magicworks
    Sep 2, 2014 at 5:50
1
$\begingroup$

You can use Wolfram Alpha for the question in the title, or here for the one in the body. Either one is enormous.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .