/external/XNNPACK/src/f32-vsqrt/ |
D | neonfma-nr2fma1adj.c.in | 74 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); variable 89 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); variable
|
D | neonfma-nr1rsqrts1fma1adj.c.in | 76 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); variable 91 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); variable
|
D | avx512f-nr1fma1adj.c.in | 71 const __m512 vadjustment = _mm512_fnmadd_ps(vsqrtx, vsqrtx, vx); variable 91 const __m512 vadjustment = _mm512_fnmadd_ps(vsqrtx, vsqrtx, vx); variable
|
D | fma3-nr1fma1adj.c.in | 70 const __m256 vadjustment = _mm256_fnmadd_ps(vsqrtx, vsqrtx, vx); variable 89 const __m256 vadjustment = _mm256_fnmadd_ps(vsqrtx, vsqrtx, vx); variable
|
/external/XNNPACK/src/f32-vsqrt/gen/ |
D | neonfma-nr2fma1adj-x4.c | 40 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr2fma1adj_x4() local 55 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr2fma1adj_x4() local
|
D | avx512f-nr1fma1adj-x16.c | 39 const __m512 vadjustment = _mm512_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__avx512f_nr1fma1adj_x16() local 59 const __m512 vadjustment = _mm512_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__avx512f_nr1fma1adj_x16() local
|
D | fma3-nr1fma1adj-x8.c | 38 const __m256 vadjustment = _mm256_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__fma3_nr1fma1adj_x8() local 57 const __m256 vadjustment = _mm256_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__fma3_nr1fma1adj_x8() local
|
D | neonfma-nr1rsqrts1fma1adj-x4.c | 40 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr1rsqrts1fma1adj_x4() local 55 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr1rsqrts1fma1adj_x4() local
|
D | neonfma-nr2fma1adj-x8.c | 77 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr2fma1adj_x8() local 92 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr2fma1adj_x8() local
|
D | avx512f-nr1fma1adj-x32.c | 70 const __m512 vadjustment = _mm512_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__avx512f_nr1fma1adj_x32() local 90 const __m512 vadjustment = _mm512_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__avx512f_nr1fma1adj_x32() local
|
D | fma3-nr1fma1adj-x16.c | 69 const __m256 vadjustment = _mm256_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__fma3_nr1fma1adj_x16() local 88 const __m256 vadjustment = _mm256_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__fma3_nr1fma1adj_x16() local
|
D | fma3-nr1fma1adj-x24.c | 79 const __m256 vadjustment = _mm256_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__fma3_nr1fma1adj_x24() local 98 const __m256 vadjustment = _mm256_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__fma3_nr1fma1adj_x24() local
|
D | neonfma-nr2fma1adj-x12.c | 90 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr2fma1adj_x12() local 105 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr2fma1adj_x12() local
|
D | neonfma-nr1rsqrts1fma1adj-x8.c | 78 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr1rsqrts1fma1adj_x8() local 93 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr1rsqrts1fma1adj_x8() local
|
D | avx512f-nr1fma1adj-x48.c | 80 const __m512 vadjustment = _mm512_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__avx512f_nr1fma1adj_x48() local 100 const __m512 vadjustment = _mm512_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__avx512f_nr1fma1adj_x48() local
|
D | neonfma-nr2fma1adj-x40.c | 146 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr2fma1adj_x40() local 181 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr2fma1adj_x40() local 196 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr2fma1adj_x40() local
|
D | neonfma-nr2fma1adj-x16.c | 103 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr2fma1adj_x16() local 118 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr2fma1adj_x16() local
|
D | avx512f-nr1fma1adj-x64.c | 90 const __m512 vadjustment = _mm512_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__avx512f_nr1fma1adj_x64() local 110 const __m512 vadjustment = _mm512_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__avx512f_nr1fma1adj_x64() local
|
D | fma3-nr1fma1adj-x32.c | 89 const __m256 vadjustment = _mm256_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__fma3_nr1fma1adj_x32() local 108 const __m256 vadjustment = _mm256_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__fma3_nr1fma1adj_x32() local
|
D | neonfma-nr1rsqrts1fma1adj-x12.c | 91 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr1rsqrts1fma1adj_x12() local 106 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr1rsqrts1fma1adj_x12() local
|
D | avx512f-nr1fma1adj-x80.c | 100 const __m512 vadjustment = _mm512_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__avx512f_nr1fma1adj_x80() local 120 const __m512 vadjustment = _mm512_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__avx512f_nr1fma1adj_x80() local
|
D | fma3-nr1fma1adj-x40.c | 99 const __m256 vadjustment = _mm256_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__fma3_nr1fma1adj_x40() local 118 const __m256 vadjustment = _mm256_fnmadd_ps(vsqrtx, vsqrtx, vx); in xnn_f32_vsqrt_ukernel__fma3_nr1fma1adj_x40() local
|
D | neonfma-nr2fma1adj-x20.c | 116 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr2fma1adj_x20() local 131 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_f32_vsqrt_ukernel__neonfma_nr2fma1adj_x20() local
|
/external/XNNPACK/src/math/ |
D | sqrt-neonfma-nr1rsqrts1fma1adj.c | 47 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_math_f32_sqrt__neonfma_nr1rsqrts1fma1adj() local
|
D | sqrt-neonfma-nr2fma1adj.c | 46 const float32x4_t vadjustment = vfmsq_f32(vx, vsqrtx, vsqrtx); in xnn_math_f32_sqrt__neonfma_nr2fma1adj() local
|