Need > 15 digits precision without big performance hit

[Jasper]

Hello,

I am doing calculations where I need ~50 decimal places of
precision..I have been writing code and Perl and noticed that even if
I convert my floating point numbers to integers, the maximum # decimal
places is 15 !! And Adding in Math::BigInt or BigFloat slows
performance to a CRAWL...does anyone have any recommendations here to
get around this problem...Perhaps a way to have 15 digits on either
side of the decimal, which it does not allow for today..If there is no
recommendation, does adding more CPU's/clustered machines buy me speed
of computation even though BigInt works at a software level ?? Does
sprintf help this issue ?

Thank you for your time,

Jasper

 

[Joe Smith]

I convert my floating point numbers to integers, the maximum # decimal
places is 15 !! And Adding in Math::BigInt or BigFloat slows
performance to a CRAWL...does anyone have any recommendations here

Buy a computer with a 64-bit processor and use a version of perl
that was compiled to use 64-bit integers.
-Joe