Output is the opposite of its input. Input can be either float or cfloat forming a 512-bit or a 1024-bit buffer (vector<float,32>, vector<cfloat,16>, vector<cfloat,16>, vector<cfloat,8>). The output is a 256-bit buffer as all the floating-point operators (vector<float,8>, vector<cfloat,4>).
vector<float,8> fpneg (vector<float,32> xbuf, int xstart, unsigned int xoffs)
for (i = 0 ; i < 8 ; i++)
ret[i] = - xbuf[xstart + xoffs[i]]