It has to be said, that AFAIK there is actually no guarantee that every possible combination of digits can be found in PI. Feel free to correct me.
Well, you can now calculate the N'th digit of PI if you want to find out, so you can skip all the known digits
https://math.hmc.edu/funfacts/finding-the-n-th-digit-of-pi/
Wow, it sounds really fun to be retired, I'd never have time for this sort of cool things!
Thanks for the link, it looks interesting!
Yes, I see I was a bit unclear about finding the index or the number itself.Thanks for the link, it looks interesting!
Ah, you have to find the index of that Fibonacci number, not the number itself. I was starting to wonder what I was missing.
5 seconds seems pretty fast indeed. It's tempting to try in Rust but I'd probably cheat and use the bigdecimal crate. Did you store your big integer values as decimal digits, or as a large binary value?
That must be the fun part.
decNumber* decNumberSin (decNumber *_result, decNumber *y, decContext *set)
{
decNumber pi, pi2, zero, one, two, x, fctl, term, cmp, result;
int i;
int negate = 0;
DEC_STASH_CONTEXT (set, &zero);
decNumberFromString (&one, "1", set);
decNumberFromString (&two, "2", set);
decNumberFromString (&pi, PI , set);
// Copy the argument y, so we can modify it.
decNumberCopy (&x, y);
// sin -x = - sin x
/* if (decCompare (&x, &zero) < 0) { */
if (decNumberIsNegative (&x)) { // x < 0
decNumberMinus (&x, &x, set);
negate = 1;
}
// We now have x >= 0
decNumberMultiply (&pi2, &pi, &two, set); // pi2 = 2*pi
decNumberMod (&x, &x, &pi2, set);
// We now have 0 <= x < 2*pi
/*if (decCompare (&x, &pi) >= 0) {*/
decNumberCompare (&cmp, &x, &pi, set);
if (!decNumberIsNegative (&cmp)) {
// x >= pi
decNumberSubtract (&x, &x, &pi, set);
negate = 1-negate;
}
// We now have 0 <= x < pi
decNumberDivide (&pi2, &pi, &two, set); // pi2 = pi/2
/*if (decCompare (&x, &pi2) >= 0) {*/
decNumberCompare (&cmp, &x, &pi2, set);
if (!decNumberIsNegative (&cmp)) {
// x >= pi/2, so let x = pi-x
decNumberSubtract (&x, &pi, &x, set);
}
// We now have 0 <= x <= pi/2.
// x^3 x^5 x^7
// sin x = x - --- + --- - ---- + ...
// 6 120 5040
//
// term(0) = x
// term(i) = - term(i-1) * x^2 / ((2*i)*(2*i+1))
decNumberCopy (&fctl, &one);
decNumberCopy (&term, &x);
decNumberCopy (&result, &x);
// DECNUMDIGITS+3 terms are enough to achieve the required precision.
for (i=0; i<DECNUMDIGITS+3; i++) {
decNumber tmpd, termout;
// term = -term * x^2 / (cnt*(cnt+1))
// cnt = cnt+2
decNumberMinus (&term, &term, set);
decNumberMultiply (&term, &term, &x, set);
decNumberMultiply (&term, &term, &x, set);
// Compute denominator and generate term.
decNumberFromInt32 (&tmpd, ((i+1)*2 + 1) * ((i+1)*2));
decNumberMultiply (&fctl, &fctl, &tmpd, set);
decNumberDivide (&termout, &term, &fctl, set);
// sum = sum + term
decNumberAdd (&result, &result, &termout, set);
}
if (negate) {
decNumberMinus (&result, &result, set);
}
DEC_RESTORE_ROUND (set, _result, &result, &zero)
return _result;
} /* decNumberSin */Well, I'm still married ....
Not sure if I understand the question. I installed Octave using a Windows installer, so I'm not compiling Octave. I'm just trying to link C++ programs to it. Visual Studio/MSVC shouldn't have problems with that, at least I haven't found anything indicating that on the web. (And for what it's worth: using C# with Octave works very well, so Octave should be familiar with the VS environment). And as far as I know (I can be wrong of course) MSVC is among the better compilers when it comes to following standards. At the very least it implemented full support for C++20 before any other of the well known compilers.
Can you post the error? Typically an include file should include all include files it require but sometime they forget. Also they sometime use name spaces without declaring them. There are so many minor possibilities nothing can be said fer sure without the error message. Of course sometimes they actually have syntax errors
What I mean is that (I think) Octave is meant to be compiled with gcc, so the sources are for GNU tools and the libraries too. The same is probably true of the headers they provide to link and interface external code, but I may be wrong.
static inline are not recognized.
msys2_shell.cmd. There's no compiler, so you'll have to install it with pacman. Hum. That seems a little... that's what I did:pacman -S cmake (ou -Syu but that's not necessary if you do it right after the other, it's just to update at the same time).