I use the _mm512_div_round_pd intrinsic, the input parameters are (0.0030214, 12458.125, 0.12, 54.4445785, 1.0, 31.125, 10.254, -57.147) and (-0.0030214, 15.11, 0.000001, 78.223245, 2, 31.25, 6.58, -4487.12). output = _mm512_div_round_pd(a, b, _MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC);
I have checked the assembly instruction, the instruction is vdivpd {rn - sae}. However, the output is not a rounding result, it's just the same as _mm512_div_pd. I have tryed _MM_FROUND_TO_NEG_INF and other parameters, but the answer is same. The question also appears in _mm512_add_round_pd function.
ps. OS is CentOs Linux release 7.6.1810, CPU model name is Intel(R) Xeon(R) Gold 6130 CPU @ 2.10GHz
