Home | History | Annotate | Download | only in builtins

Lines Matching defs:reciprocal

78     // [1, 2.0) and get a Q32 approximate reciprocal using a small minimax
79     // polynomial approximation: reciprocal = 3/4 + 1/sqrt(2) - b/2.  This
84     // Now refine the reciprocal estimate using a Newton-Raphson iteration:
108     uint64_t correction, reciprocal;
112     reciprocal = (uint64_t)recip32*cHi + ((uint64_t)recip32*cLo >> 32);
115     // 64-bit reciprocal estimate downward to ensure that it is strictly smaller
116     // than the infinitely precise exact reciprocal.  Because the computation
119     reciprocal -= 2;
121     // The numerical reciprocal is accurate to within 2^-56, lies in the
122     // interval [0.5, 1.0), and is strictly smaller than the true reciprocal
123     // of b.  Multiplying a by this reciprocal thus gives a numerical q = a/b
134     wideMultiply(aSignificand << 2, reciprocal, &quotient, &quotientLo);