| /external/XNNPACK/src/f32-vscaleexpminusmax/gen/ |
| D | avx512f-p5-scalef-x80.c | 62 __m512 vn4 = _mm512_roundscale_ps(_mm512_mul_ps(vx4, vlog2e), 0); in xnn_f32_vscaleexpminusmax_ukernel__avx512f_p5_scalef_x80() local
|
| D | avx512f-p5-scalef-x96.c | 64 __m512 vn4 = _mm512_roundscale_ps(_mm512_mul_ps(vx4, vlog2e), 0); in xnn_f32_vscaleexpminusmax_ukernel__avx512f_p5_scalef_x96() local
|
| D | avx2-p5-x40.c | 66 __m256 vn4 = _mm256_fmadd_ps(vx4, vlog2e, vmagic_bias); in xnn_f32_vscaleexpminusmax_ukernel__avx2_p5_x40() local
|
| D | avx512f-p5-scalef-x112.c | 66 __m512 vn4 = _mm512_roundscale_ps(_mm512_mul_ps(vx4, vlog2e), 0); in xnn_f32_vscaleexpminusmax_ukernel__avx512f_p5_scalef_x112() local
|
| /external/XNNPACK/src/f32-vscaleextexp/gen/ |
| D | avx512f-p5-scalef-x80.c | 56 const __m512 vn4 = _mm512_roundscale_ps(_mm512_mul_ps(vx4, vlog2e), 0); in xnn_f32_vscaleextexp_ukernel__avx512f_p5_scalef_x80() local
|
| D | avx512f-p5-scalef-x96.c | 57 const __m512 vn4 = _mm512_roundscale_ps(_mm512_mul_ps(vx4, vlog2e), 0); in xnn_f32_vscaleextexp_ukernel__avx512f_p5_scalef_x96() local
|
| D | avx2-p5-x40.c | 62 …const __m256 vn4 = _mm256_round_ps(_mm256_mul_ps(vx4, vlog2e), _MM_FROUND_TO_NEAREST_INT | _MM_FRO… in xnn_f32_vscaleextexp_ukernel__avx2_p5_x40() local
|
| D | avx512f-p5-scalef-x112.c | 58 const __m512 vn4 = _mm512_roundscale_ps(_mm512_mul_ps(vx4, vlog2e), 0); in xnn_f32_vscaleextexp_ukernel__avx512f_p5_scalef_x112() local
|
| D | avx512f-p5-scalef-x128.c | 59 const __m512 vn4 = _mm512_roundscale_ps(_mm512_mul_ps(vx4, vlog2e), 0); in xnn_f32_vscaleextexp_ukernel__avx512f_p5_scalef_x128() local
|
| /external/XNNPACK/src/f32-velu/gen/ |
| D | velu-wasm-rr2-p6-x5.c | 59 float vn4 = vz4 * vlog2e + vmagic_bias; in xnn_f32_velu_ukernel__wasm_rr2_p6_x5() local
|
| D | velu-scalar-rr2-p6-x5.c | 59 float vn4 = vz4 * vlog2e + vmagic_bias; in xnn_f32_velu_ukernel__scalar_rr2_p6_x5() local
|
| D | velu-wasm-rr2-p6-x6.c | 61 float vn4 = vz4 * vlog2e + vmagic_bias; in xnn_f32_velu_ukernel__wasm_rr2_p6_x6() local
|
| D | velu-avx512f-rr1-p6-x80.c | 59 __m512 vn4 = _mm512_fmadd_ps(vz4, vlog2e, vmagic_bias); in xnn_f32_velu_ukernel__avx512f_rr1_p6_x80() local
|
| D | velu-avx2-rr1-p6-x40.c | 57 __m256 vn4 = _mm256_fmadd_ps(vz4, vlog2e, vmagic_bias); in xnn_f32_velu_ukernel__avx2_rr1_p6_x40() local
|
| D | velu-scalar-rr2-p6-x6.c | 61 float vn4 = vz4 * vlog2e + vmagic_bias; in xnn_f32_velu_ukernel__scalar_rr2_p6_x6() local
|
| D | velu-wasm-rr2-lut16-p3-x5.c | 59 float vn4 = vz4 * vlog2e + vmagic_bias; in xnn_f32_velu_ukernel__wasm_rr2_lut16_p3_x5() local
|
| /external/XNNPACK/src/f16-raddstoreexpminusmax/gen/ |
| D | avx2-rr1-p2-x40.c | 57 __m256 vn4 = _mm256_fmadd_ps(vx4, vlog2e, vmagic_bias); in xnn_f16_raddstoreexpminusmax_ukernel__avx2_rr1_p2_x40() local
|
| D | avx2-rr1-p2-x40-acc2.c | 58 __m256 vn4 = _mm256_fmadd_ps(vx4, vlog2e, vmagic_bias); in xnn_f16_raddstoreexpminusmax_ukernel__avx2_rr1_p2_x40_acc2() local
|
| D | avx2-rr1-p2-x40-acc5.c | 61 __m256 vn4 = _mm256_fmadd_ps(vx4, vlog2e, vmagic_bias); in xnn_f16_raddstoreexpminusmax_ukernel__avx2_rr1_p2_x40_acc5() local
|
| D | avx2-rr1-p2-x48-acc2.c | 60 __m256 vn4 = _mm256_fmadd_ps(vx4, vlog2e, vmagic_bias); in xnn_f16_raddstoreexpminusmax_ukernel__avx2_rr1_p2_x48_acc2() local
|
| D | neonfp16arith-rr2-p2-x40.c | 57 float16x8_t vn4 = vfmaq_f16(vmagic_bias, vx4, vlog2e); in xnn_f16_raddstoreexpminusmax_ukernel__neonfp16arith_rr2_p2_x40() local
|
| /external/XNNPACK/src/f32-vsigmoid/gen/ |
| D | vsigmoid-avx512f-rr1-p5-scalef-div-x80.c | 55 __m512 vn4 = _mm512_mul_ps(vz4, vlog2e); in xnn_f32_vsigmoid_ukernel__avx512f_rr1_p5_scalef_div_x80() local
|
| D | vsigmoid-avx512f-rr1-p5-scalef-div-x96.c | 57 __m512 vn4 = _mm512_mul_ps(vz4, vlog2e); in xnn_f32_vsigmoid_ukernel__avx512f_rr1_p5_scalef_div_x96() local
|
| D | vsigmoid-avx512f-rr1-p5-scalef-nr1fma-x80.c | 55 __m512 vn4 = _mm512_mul_ps(vz4, vlog2e); in xnn_f32_vsigmoid_ukernel__avx512f_rr1_p5_scalef_nr1fma_x80() local
|
| /external/XNNPACK/src/f32-raddexpminusmax/gen/ |
| D | avx512f-p5-scalef-x128-acc2.c | 68 const __m512 vn4 = _mm512_roundscale_ps(_mm512_mul_ps(vx4, vlog2e), 0); in xnn_f32_raddexpminusmax_ukernel__avx512f_p5_scalef_x128_acc2() local
|