// 
// sub_101EF4 from 0x101ef4 to 0x1022d4 (992 bytes)
// 
// 1 xrefs:
// > <???> @ 0x11a0f8
// 
// This function is hookable on this platform!
// 
// Using BromaIDA 8.0.0 @ https://github.com/Stazzical/BromaIDA
// Using bindings at commit ba9f177b at Thu Jul 16 22:52:06 2026 from https://github.com/geode-sdk/bindings
// 

float32x4_t *__fastcall sub_101EF4(
        float32x4_t *result,
        int a2,
        __int64 a3,
        __int64 a4,
        unsigned __int64 *a5,
        double a6,
        double a7,
        double a8,
        long double a9,
        double a10,
        long double a11)
{
  float v11; // w8
  long double v12; // q0
  float v13; // w9
  float i; // w10
  unsigned int v15; // w11
  _WORD *v16; // x12
  _WORD *v17; // x14
  float v18; // s4
  int v19; // w1
  int v20; // w8
  int64x2_t v21; // q4
  int64x2_t v22; // q2
  float32x4_t v23; // q0
  int64x2_t v24; // q5
  int64x2_t v25; // q2
  int64x2_t v26; // q4
  unsigned __int64 v27; // d7
  int16x8_t *v28; // x10
  __int64 v29; // x9
  int16x8_t v30; // q7
  int16x8_t v31; // q17
  __int128 v32; // q19
  unsigned __int64 v33; // d16
  int32x4_t v34; // q6
  int32x4_t v35; // q20
  int8x16_t v36; // q7
  int8x16_t v37; // q17
  int8x16_t v38; // q21
  int8x16_t v39; // q18
  float32x4_t v40; // q22
  float32x4_t v41; // q21
  float32x4_t v42; // q20
  int16x8_t *v43; // x9
  int16x8_t v44; // q7
  int16x8_t v45; // q16
  int16x4_t *v46; // x9
  int32x4_t v47; // q20
  int8x16_t v48; // q7
  int8x16_t v49; // q16
  int8x16_t v50; // q21
  int8x16_t v51; // q18
  float32x4_t v52; // q22
  float32x4_t v53; // q21
  float32x4_t v54; // q20
  float32x4_t v55; // q21
  float v56; // w9
  long double v57; // q0
  float v58; // w10
  float v59; // w11
  unsigned int v60; // w12
  _WORD *v61; // x13
  _WORD *v62; // x15
  float v63; // s4
  _QWORD *v64; // x4
  int v65; // w8

  v11 = 4.6566e-10;
  *(_QWORD *)&v12 = 0x3800000038000000LL;
  *((_QWORD *)&v12 + 1) = 0x3800000038000000LL;
  v13 = 0.000030518;
  for ( i = -0.000030518; a2; *(_QWORD *)a4 += *a5 )
  {
    if ( ((unsigned __int8)result & 0xF) == 0 )
      break;
    v15 = 6 * *(_DWORD *)(a4 + 4);
    v16 = (_WORD *)(a3 + 2LL * v15);
    LOWORD(a11) = *v16;
    v17 = (_WORD *)(a3 + 2LL * (v15 + 6));
    LOWORD(a9) = *v17;
    WORD1(a11) = v16[1];
    WORD2(a11) = *(_WORD *)(a3 + 2LL * (v15 + 2));
    WORD1(a9) = v17[1];
    WORD2(a9) = *(_WORD *)(a3 + 2LL * (v15 + 8));
    WORD3(a11) = *(_WORD *)(a3 + 2LL * (v15 + 3));
    v18 = (float)*(__int16 *)(a3 + 2LL * (v15 + 5)) * i;
    WORD3(a9) = *(_WORD *)(a3 + 2LL * (v15 + 9));
    result = (float32x4_t *)sub_102D58(
                              v12,
                              v18 + (float)((float)*(__int16 *)(a3 + 2LL * (v15 + 11)) * v13),
                              (float)(*(_DWORD *)a4 >> 1) * v11,
                              a9,
                              v18,
                              a11,
                              (float)((float)*(__int16 *)(a3 + 2LL * (v15 + 4)) * i)
                            + (float)((float)*(__int16 *)(a3 + 2LL * (v15 + 10)) * v13));
    a2 = v19 - 1;
  }
  v20 = a2 >> 2;
  if ( a2 >> 2 )
  {
    v21 = vld1q_dup_f64((const double *)a4);
    v22.n128_u64[0] = 0;
    v23.n128_u64[0] = 0x3000000030000000LL;
    v23.n128_u64[1] = 0x3000000030000000LL;
    v24.n128_u64[0] = 2 * *a5;
    v22.n128_u64[1] = *a5;
    v24.n128_u64[1] = 3 * *a5;
    a9 = COERCE_LONG_DOUBLE(vdupq_n_s64(4 * *a5));
    v25 = vaddq_s64(v21, v22);
    v26 = vaddq_s64(v24, v21);
    *(_QWORD *)&a11 = 0x3800000038000000LL;
    *((_QWORD *)&a11 + 1) = 0x3800000038000000LL;
    do
    {
      v27 = vmul_s32(
              vzip2_s32((int32x2_t)v25.n128_u64[0], (int32x2_t)vextq_s8(v25, v25, 8u).n128_u64[0]),
              (int32x2_t)0x600000006LL).n64_u64[0];
      v28 = (int16x8_t *)(a3 + 2LL * (unsigned int)v27);
      v29 = a3 + 2LL * HIDWORD(v27);
      v30 = *v28;
      v31 = v28[1];
      v32 = *(_OWORD *)(v29 + 16);
      v33 = vmul_s32(
              vzip2_s32((int32x2_t)v26.n128_u64[0], (int32x2_t)vextq_s8(v26, v26, 8u).n128_u64[0]),
              (int32x2_t)0x600000006LL).n64_u64[0];
      v34 = vmulq_f32(vcvtq_f32_u32(vshrq_n_u32(vuzp1q_s32(v25, v26), 1u)), v23);
      LODWORD(v28) = HIDWORD(v33);
      v35 = vmovl_s16((int16x4_t)v30.n128_u64[0]);
      v36 = vcvtq_f32_s32(vmovl_high_s16(v30));
      v37 = vcvtq_f32_s32(vmovl_s16((int16x4_t)v31.n128_u64[0]));
      v38 = vcvtq_f32_s32(vmovl_s16(*(int16x4_t *)v29));
      v39 = vcvtq_f32_s32(vmovl_high_s16(*(int16x8_t *)v29));
      v40 = vextq_s8(v36, v37, 8u);
      v36.n128_u64[1] = v38.n128_u64[0];
      v41 = vextq_s8(v38, v39, 8u);
      v42 = vcvtq_f32_s32(v35);
      v43 = (int16x8_t *)(a3 + 2LL * (unsigned int)v33);
      result[1] = vmulq_f32(
                    vmlaq_f32(
                      v36,
                      vuzp1q_s32(
                        vdupq_lane_s32((int32x2_t)v34.n128_u64[0], 0),
                        vdupq_lane_s32((int32x2_t)v34.n128_u64[0], 1)),
                      vsubq_f32(vextq_s8(v37, vextq_s8(v39, v37, 8u), 8u), v36)),
                    *(float32x4_t *)&a11);
      result[2] = vmulq_f32(
                    vmlaq_lane_f32(
                      v41,
                      vsubq_f32(vcvtq_f32_s32(vmovl_s16(*(int16x4_t *)&v32)), v41),
                      (float32x2_t)v34.n128_u64[0],
                      1),
                    *(float32x4_t *)&a11);
      *result = vmulq_f32(vmlaq_n_f32(v42, vsubq_f32(v40, v42), v34.n128_f32[0]), *(float32x4_t *)&a11);
      v44 = *v43;
      v45 = v43[1];
      v46 = (int16x4_t *)(a3 + 2LL * (unsigned int)v28);
      v47 = vmovl_s16((int16x4_t)v44.n128_u64[0]);
      v48 = vcvtq_f32_s32(vmovl_high_s16(v44));
      v49 = vcvtq_f32_s32(vmovl_s16((int16x4_t)v45.n128_u64[0]));
      v50 = vcvtq_f32_s32(vmovl_s16((int16x4_t)v46->n64_u64[0]));
      v51 = vcvtq_f32_s32(vmovl_high_s16(*(int16x8_t *)v46->n64_u64));
      v52 = vextq_s8(v48, v49, 8u);
      v48.n128_u64[1] = v50.n128_u64[0];
      v53 = vextq_s8(v50, v51, 8u);
      v54 = vcvtq_f32_s32(v47);
      v55 = vmlaq_laneq_f32(v53, vsubq_f32(vcvtq_f32_s32(vmovl_s16(v46[2])), v53), v34, 3);
      --v20;
      v25 = vaddq_s64(v25, *(int64x2_t *)&a9);
      v26 = vaddq_s64(v26, *(int64x2_t *)&a9);
      result[4] = vmulq_f32(
                    vmlaq_f32(
                      v48,
                      vuzp1q_s32(vdupq_laneq_s32(v34, 2), vdupq_laneq_s32(v34, 3)),
                      vsubq_f32(vextq_s8(v49, vextq_s8(v51, v49, 8u), 8u), v48)),
                    *(float32x4_t *)&a11);
      result[5] = vmulq_f32(v55, *(float32x4_t *)&a11);
      result[3] = vmulq_f32(vmlaq_laneq_f32(v54, vsubq_f32(v52, v54), v34, 2), *(float32x4_t *)&a11);
      result += 6;
    }
    while ( v20 );
    *(_QWORD *)a4 = v25.n128_u64[0];
  }
  v56 = 4.6566e-10;
  *(_QWORD *)&v57 = 0x3800000038000000LL;
  *((_QWORD *)&v57 + 1) = 0x3800000038000000LL;
  v58 = 0.000030518;
  v59 = -0.000030518;
  if ( (a2 & 3) != 0 )
  {
    do
    {
      v60 = 6 * *(_DWORD *)(a4 + 4);
      v61 = (_WORD *)(a3 + 2LL * v60);
      LOWORD(a11) = *v61;
      v62 = (_WORD *)(a3 + 2LL * (v60 + 6));
      LOWORD(a9) = *v62;
      WORD1(a11) = v61[1];
      WORD2(a11) = *(_WORD *)(a3 + 2LL * (v60 + 2));
      WORD1(a9) = v62[1];
      WORD2(a9) = *(_WORD *)(a3 + 2LL * (v60 + 8));
      WORD3(a11) = *(_WORD *)(a3 + 2LL * (v60 + 3));
      v63 = (float)*(__int16 *)(a3 + 2LL * (v60 + 5)) * v59;
      WORD3(a9) = *(_WORD *)(a3 + 2LL * (v60 + 9));
      result = (float32x4_t *)sub_102D58(
                                v57,
                                v63 + (float)((float)*(__int16 *)(a3 + 2LL * (v60 + 11)) * v58),
                                (float)(*(_DWORD *)a4 >> 1) * v56,
                                a9,
                                v63,
                                a11,
                                (float)((float)*(__int16 *)(a3 + 2LL * (v60 + 4)) * v59)
                              + (float)((float)*(__int16 *)(a3 + 2LL * (v60 + 10)) * v58));
      *(_QWORD *)a4 += *v64;
    }
    while ( v65 != 1 );
  }
  return result;
}