78     // [1, 2.0) and get a Q64 approximate reciprocal using a small minimax  in __divtf3()
79     // polynomial approximation: reciprocal = 3/4 + 1/sqrt(2) - b/2.  This  in __divtf3()
85     // Now refine the reciprocal estimate using a Newton-Raphson iteration:  in __divtf3()
112     rep_t correction, reciprocal;  in __divtf3()  local
128     reciprocal = r64cH + (r64cL >> 64);  in __divtf3()
131     // 128-bit reciprocal estimate downward to ensure that it is strictly smaller  in __divtf3()
132     // than the infinitely precise exact reciprocal.  Because the computation  in __divtf3()
135     reciprocal -= 2;  in __divtf3()
137     // The numerical reciprocal is accurate to within 2^-112, lies in the  in __divtf3()
138     // interval [0.5, 1.0), and is strictly smaller than the true reciprocal  in __divtf3()
139     // of b.  Multiplying a by this reciprocal thus gives a numerical q = a/b  in __divtf3()
150     wideMultiply(aSignificand << 2, reciprocal, &quotient, &quotientLo);  in __divtf3()