@programminginc - страница 966

Zorrito

int main(){ mpf_set_default_prec(256); mpf_t pi; mpf_set(pi, 3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628); cout << setprecision(256); cout << "GMP:" << pi << endl; return 0; }

S.

int main(){ mpf_set_default_prec(256); mpf_t pi; mpf_set(pi, 3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628); cout << setprecision(256); cout << "GMP:" << pi << endl; return 0; }

Unless mpf_set is a macro (which seems not), 3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628 will be treated as only double in c/c++

S.

You may need mpf_set_str

S.

Sorry I have no knowledge on GMP, but your code just doesn't make sense in c/c++

Zorrito

Unless mpf_set is a macro (which seems not), 3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628 will be treated as only double in c/c++

mpf_set() is a function that sets a variable to a constant with type mpf_t.

S.

Then, what would happed when passing a 3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628 to mpf_t?

S.

imho, it would be rounded as double, lossing its precision, and then converted to mpf_t

S.

So to keep the precision, you need to convert a c-style string to mpf_t

Zorrito

You may need mpf_set_str

I tried this, and it's giving me the same kind of error.

S.

show the code and the full complaint from the compiler

Zorrito

show the code and the full complaint from the compiler

#include <iostream> #include <gmp.h> #include <iomanip> const int prec = 256; int main(){ float pif = 3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628; double pidf = 3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628; long double pildf = 3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628L; mpf_set_default_prec(prec); mpf_t pimpf; mpf_set(pimpf, 3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628); cout << setprecision(prec); cout << "Float: " << pif << '\n'; cout << "Double: " << pidf << '\n'; cout << "Long: " << pildf << '\n'; cout << "GMP: " << pimpf << endl; return 0; }

Zorrito

#include <iostream> #include <gmp.h> #include <iomanip> const int prec = 256; int main(){ float pif = 3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628; double pidf = 3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628; long double pildf = 3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628L; mpf_set_default_prec(prec); mpf_t pimpf; mpf_set(pimpf, 3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628); cout << setprecision(prec); cout << "Float: " << pif << '\n'; cout << "Double: " << pidf << '\n'; cout << "Long: " << pildf << '\n'; cout << "GMP: " << pimpf << endl; return 0; }

I use to compile it g++ -std=c++11 filename.cpp -lgmpxx -lgmp

S.

float pif = 3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628; double pidf = 3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628; long double pildf = 3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628L; they're plainly wrong here

S.

None of float, double, and long double, can store such a high precision

Zorrito

None of float, double, and long double, can store such a high precision

I know they can't. They're just test variables for me to check their precision against my mpf_t variable.

Zorrito

I know they can't. They're just test variables for me to check their precision against my mpf_t variable.

When I use mpf_set_d, I get the same precision as the double variable (as expected), but the compiler doesn't know what to do when I do mpf_set.

S.

You may have tried, const char* p = "3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628"; mpf_t f; ... mpf_set_str(f, p, 10);

S.

Isn't there any example snippet of GMP from the official documentation?

Anonymous

Nice bot

Zorrito

Isn't there any example snippet of GMP from the official documentation?

https://gmplib.org/manual/Assigning-Floats.html#Assigning-Floats

Zorrito

You may have tried, const char* p = "3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628"; mpf_t f; ... mpf_set_str(f, p, 10);

That got rid of the error, but now, it's printing pi=4.

S.

That got rid of the error, but now, it's printing pi=4.

you may need gmp_printf and proper calling

S.

That got rid of the error, but now, it's printing pi=4.

By a simple search I found: mpf_t f; mpf_init2(f, 64); mpf_set_str(f, "0.0415373652931074065807663354", 10); gmp_printf ("fixed point mpf %Ff\n", f);

S.

Please test it

Zorrito

Using gmp_printf: For pi, I'm just getting 4.000000, and for the code you want me to test, I got 0.000000.

S.

what's your full code for test?

Zorrito

Oh, I had a typo. Your test code worked fine (printed 0.041537).

S.

mp_printf ("fixed point mpf %.64Ff\n", f);

S.

try again?

S.

by default it only outputs 6 digits

S.

( like printf

Zorrito

mp_printf ("fixed point mpf %.64Ff\n", f);

Ok, that worked :D

S.

Full precision?

Zorrito

Do you have a link to a reference?

S.

No. YOu may need to read the official manual

S.

Or some examples from github

Zorrito

Type long double gave me 3.141592653589793238(51280895940618620443274267017841339111328125) Type mpf_t gave me 3.1415926535897932384626433832795028841971693993751058209749445923 (all decimal places are correct)

Zorrito

No. YOu may need to read the official manual

Okay, thanks :)

S.

ur welcome

Zorrito

A big part of the problem was that I did not get to the part about the gmp_printf function, so I used what I knew lol. I still think it's weird that mpf_set doesn't work, but as long as I have this workaround, I guess it won't matter too much.

S.

A big part of the problem was that I did not get to the part about the gmp_printf function, so I used what I knew lol. I still think it's weird that mpf_set doesn't work, but as long as I have this workaround, I guess it won't matter too much.

Not weird. Just the compiler cannot decide which overloaded function to use.

S.

A big part of the problem was that I did not get to the part about the gmp_printf function, so I used what I knew lol. I still think it's weird that mpf_set doesn't work, but as long as I have this workaround, I guess it won't matter too much.

... and to assign a high precision number, using a c style string is not a "wrokaround", it's the only way to go

S.

because none of the default data types has such a high precision

Zorrito

because none of the default data types has such a high precision

Oh, I see :o When you put it that way, it makes a lot more sense.

Medina