Printable View
I am writing a program that can calculate something that never stops like 1/3, and i was wondering what is the max place i could take something to? If i have this calcuating will it freez the computer? Im not doing 1/3 but im doing a calulatino that goes on for ever. And am just wondering the max calcuation i can do.
depends what you want to calculate and where you want to store your number, the decimals of rational numbers repeats themselves, so its not particularly hard but calculating trancendent numbers needs some strategy, but theoretically i dont think theres but physical limits, like ram/hdd space.
im going to calculate it on a computer with 512mb ram, and a 100gig harddrive. im going to store it in just like a text file or something.
On my system, .dat files larger than 10MB tend to become difficult to read (and to send via internet. I have 56k :) ) However, depending on how you set your data structure / technique of storage, you can get pretty large.
If you don't me asking, what are you calculating? :confused:
well its not an actualy calculation i guess. but im going to have the program calculate every posible outcome for 20 digets with all caricters.
what operating system should i use?
if you are doing this in C++, it does not really matter, though I'm a bit confused of what you are attempting.
I think you should limit your concerns to what datatype to use to store your intermediate and final results.
For floating point use long double, integers use _int64
20 significant digits is no big deal, really.
From MSDN
Quote:
long double
The long double contains 80 bits: 1 for sign, 15 for exponent, and 64 for mantissa. Its range is +/–1.2E4932 with at least 19 digits of precision. Although long double and double are separate types, the representation of long double and double is identical.
_int64
64 bit signed or unsigned integer
VC5 and above doesn't support long double. long double defaults to double in VC, though they cannot be interchanged.
MS made this decision when it was designing WinNT and NT is supposed to run on other platforms as well and not all CPU's FPUs (other than Intel's) support 80bit long double.
If that is a problem, then you can always implement it in an array.