| #include <stdio.h> | |
| #include <altivec.h> | |
| #if 0 | |
| vector float func(vector float vi){ | |
| #else | |
| void func(vector float *arg){ | |
| vector float vi = *arg; | |
| #endif | |
| vector float vzero = {0.0f, 0.0f, 0.0f, 0.0f}; | |
| vector float vres; | |
| vector int vmask; | |
| #ifdef __LITTLE_ENDIAN__ | |
| vmask = vec_slo(vec_splat_s32(-1), | |
| vec_splat_u8(12) * vec_splat_u8(8)); | |
| #else | |
| vmask = vec_sld(vec_splat_s32(0), vec_splat_s32(-1), 4); | |
| #endif | |
| vi = vec_sel(vi, vzero, (vector unsigned int)vmask); /* (x, y, z, 0) */ | |
| vres = vec_madd(vi, vi, vzero); /* (x^2, y^2, z^2, 0) */ | |
| if(vec_all_eq(vres, vzero)) /* Is squared vec equal to the zero vec? */ | |
| { | |
| vres = vec_ctf(vec_sld(vmask, vmask, 4), 0); /* return (0 0 -1 0) */ | |
| } | |
| else | |
| { | |
| vector float vtmp; | |
| #ifdef __LITTLE_ENDIAN__ | |
| vtmp = vec_sld(vres, vres, 12); /* (y^2, z^2, 0, x^2) */ | |
| vtmp[3] = 0.0f; /* (y^2, z^2, 0, 0) */ | |
| #else | |
| vtmp = vec_sld(vres, vres, 4); | |
| #endif | |
| vtmp = vec_add(vtmp, vres); /* (x^2 + y^2, y^2 + z^2, z^2, x^2) */ | |
| vtmp = vec_sld(vtmp, vtmp, 8); /* (z^2, x^2, x^2 + y^2, y^2 + z^2) */ | |
| vtmp = vec_add(vtmp, vres); /* (x^2 + z^2, x^2 + y^2, x^2 + y^2 + z^2, y^2 + z^2) */ | |
| vtmp = vec_splat(vtmp, 2); | |
| vtmp = vec_rsqrte(vtmp); | |
| vres = vec_madd(vtmp, vi, vzero); | |
| } | |
| #if 0 | |
| return vres; | |
| #else | |
| *arg = vres; | |
| #endif | |
| } | |
| int main(int argc, char *argv[]) | |
| { vector float v = {0.0f, 3.0f, 1.0f, 4.0f}; | |
| float f[4]; | |
| #if 0 | |
| v = func(v); | |
| #else | |
| func(&v); | |
| #endif | |
| f[0] = ((float *) &v)[0]; | |
| f[1] = ((float *) &v)[1]; | |
| f[2] = ((float *) &v)[2]; | |
| f[3] = ((float *) &v)[3]; | |
| printf("[%f %f %f %f]\n", f[0], f[1], f[2], f[3]); | |
| return 0; | |
| } | |