mpdecimal for Windows

[srvaldez]

I built the Windows libraries of mpdecimal for anyone interested

[deltarho[1859]]

I had a bit of a struggle to get the required files into a folder so that FB could find them. Anyway, I got there.

FB hasn't got a log base 10 but there is an example in the manual to achieve that and I got log base 10 of 8 as 0.9030899869919434.

Code: Select all

The 'ke!san Online Calculator' gave me 0.9030899869919435856412
Your build (50 sig figs) gave me       0.90308998699194358564121668417347908030456964438633

and we can have arbitrary precision?

I expected mpDecimal-test.exe to take a while but it came in faster than I can blink.

WOW! The mind boggles.

Well done, srvaldez.

[srvaldez]

@deltarho[1859]
thank you for testing :smile:

[jdebord]

Just downloaded. Not yet tested ...

The web site for mpdecimal is here:

http://www.bytereef.org/mpdecimal/index.html

EDIT. Tested now. The two demo programs work well. The DLL too, but the file name (libmpdec.so.2.5.0) cannot be modified.

[dodicat]

Hi srvaldez
I keep going to your download link (Dark page with the filename at the top left), but I always get that little win 10 waiting circle going round and round.
Maybe it is because of the slow broadband connection here.
I'll keep trying.

[srvaldez]

try again, I uploaded it to GitHub and updated the link

[dodicat]

Thanks srvaldez.
I'll test it later.
deltarho[]
fb has log10

Code: Select all

 
#include "crt.bi"

print log10f(8)
print log10(8)

sleep
 

I also read this:
Benford’s Law,states that the fractional part of the logarithm of random data is uniformly distributed between 0 and 1.
So, for instance, in a set of figures taken from real life, 1 is by far the most common first digit.
This has a neat application in fraud detection, in that most people ignore Benford’s Law when making up figures, and create their first digits in equal numbers.

[deltarho[1859]]

fb has log10

Thanks for that.

[hhr]

Code: Select all

''=============================================================================
#include "mpdecimal.bi"

'we need to initialize mpd_context before including the overloded operators
Dim Shared As mpd_context_t mpd_context
mpd_init(@mpd_context, 8)

#include "mpdoverload.bi" 'mpdecimal overloaded operators
''=============================================================================
mpd_context.prec = 32
Dim As mpd summand, sum, n, ten

sum = 0
n = 0
ten = 10

Do ' Bellard's formula
   sum += (-32/(4*n+1)-1/(4*n+3)+256/(10*n+1)-64/(10*n+3)-4/(10*n+5)-4/(10*n+7)+1/(10*n+9)) / (1024^n)
   n += 1
   summand = -(-32/(4*n+1)-1/(4*n+3)+256/(10*n+1)-64/(10*n+3)-4/(10*n+5)-4/(10*n+7)+1/(10*n+9)) / (1024^n)
   sum += summand
   n += 1
Loop While Abs(summand) > ten^(-mpd_context.prec)
sum /= 64

Print sum
Print "press any key to exit ";
Sleep

[srvaldez]

@hhr
nice example :-)

[Andrew Lindsay]

Does anyone have a copy of the Bi and DLL for this?

I'd love to have a copy if anyone can oblige.

Regards

Andrew

[srvaldez]

hi Andrew Lindsay
link above updated, try again

[Andrew Lindsay]

Thanks, that’s greatly appreciated.