Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 13 May 1999 20:49:35 -0400 (EDT)
From:      Thomas David Rivers <rivers@dignus.com>
To:        advocacy@freebsd.org, jesus.monroy@usa.net, licia@o-o.org, rivers@dignus.com
Subject:   Re: [Re: BUDS Coming to you soon.]
Message-ID:  <199905140049.UAA45243@lakes.dignus.com>
In-Reply-To: <Pine.BSF.4.05.9905131916060.3370-100000@o-o.org>

next in thread | previous in thread | raw e-mail | index | archive | help
> 
> Uhm, not to be a nitpicker, but 2^313 isn't -that- huge, according to bc,
> it's 
> 
> 166873987181321100187111070794496258953336290809113497652112625611110\
> 91607661254297054391304192
> 
> Which is a lot of kernels, but I could print it out on a 3x5 note card pretty
> easily ;)
> 
> 

 Are you sure that's right?  Seems kinda small to me :-)  But, I'm
 wrong more than I care to admit :-)

 So, 2**64 is 18446744073709551615 (according to "%qu") - 20 digits, 
 seems like 2**313 is going to be a lot longer...  let's say we gain
 a 10's place digit for every 4 powers of two (which is, of course,
 way conservative, as you actually gain more than one 10s digit.)  
 313-64 is 249; 249/4 is 62...

 Hmm... seems like you're right - you should have about 20+62 with
 my quick conservative math - and you have about 100 digits up there
 (I didn't count them.. just guessed) 
 
 So - yep - you could print it on a 3x5 card :-)

 But - it still is an awful lot of kernels...

 How long does it take to build a single kernel - say, a minute,
 or better, say,  30 seconds... let's make that a power of two and say, uh,
 32 seconds (2**5). 

 So, we need 2**5 * 2**313 total seconds... that's 2**318 seconds.

 Again, fudging a little on the math; let's pretend that a minute
 is 64 seconds instead of 60 - which is 2**6.  So, we need
 2**312 minutes.  Again, fudging and saying an hour is 64 minutes,
 we'd need 2**306 hours.

 (bc could likely figure this out much more precisely - but I'm 
 "thinking with my fingers")

 Fudging again, and saying there are 32 hours in a day, that's
 2**301 days.

 If we make a week 8 days long, that's 2**298 weeks.

 Making a year 64 weeks long (we are *really* stretching things here, 
 but I think we're stretching in the "right" way), that makes 2**293 years.

 Make a millenium 1024 years, and we have 2**283 millenia.

 If I've kept everything straight - that's an *awfully* long time :-)

 Maybe I've messed up along the way... so, just lop off a whole
 chunk of years and say 2**250 millenia...

 Now, how old is the universe?


 Of course, there are other ways to tackle the problem:

    If you had 8 billion people (2**33) (the world has about
 6 billion people on it, right now) and each did, say 2*280
 kernel builds a night, you could get it in one night.

    Let's give each of these 8 billion people 4 computers.  That's
 2**35 computers - so, each computer only needs to do 2**278
 kernel builds per night... 

 
  :-) :-) :-)



	- Dave Rivers -


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-advocacy" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199905140049.UAA45243>