| /external/XNNPACK/src/f32-raddstoreexpminusmax/gen/ |
| D | neon-lut64-p2-x8.c | 87 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidx67]); in xnn_f32_raddstoreexpminusmax_ukernel__neon_lut64_p2_x8() local
93 vl67 = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neon_lut64_p2_x8()
94 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_raddstoreexpminusmax_ukernel__neon_lut64_p2_x8()
|
| D | neonfma-lut64-p2-x8-acc2.c | 87 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidx67]); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_lut64_p2_x8_acc2() local
93 vl67 = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_lut64_p2_x8_acc2()
94 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_lut64_p2_x8_acc2()
|
| D | neon-lut64-p2-x8-acc2.c | 88 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidx67]); in xnn_f32_raddstoreexpminusmax_ukernel__neon_lut64_p2_x8_acc2() local
94 vl67 = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neon_lut64_p2_x8_acc2()
95 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_raddstoreexpminusmax_ukernel__neon_lut64_p2_x8_acc2()
|
| D | neonfma-lut64-p2-x8.c | 86 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidx67]); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_lut64_p2_x8() local
92 vl67 = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_lut64_p2_x8()
93 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_lut64_p2_x8()
|
| D | neonfma-lut64-p2-x12-acc2.c | 94 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidx67]); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_lut64_p2_x12_acc2() local
102 vl67 = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_lut64_p2_x12_acc2()
103 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_lut64_p2_x12_acc2()
|
| D | neon-lut64-p2-x12.c | 94 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidx67]); in xnn_f32_raddstoreexpminusmax_ukernel__neon_lut64_p2_x12() local
102 vl67 = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neon_lut64_p2_x12()
103 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_raddstoreexpminusmax_ukernel__neon_lut64_p2_x12()
|
| D | neonfma-lut64-p2-x12.c | 93 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidx67]); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_lut64_p2_x12() local
101 vl67 = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_lut64_p2_x12()
102 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_lut64_p2_x12()
|
| D | neon-lut64-p2-x12-acc3.c | 96 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidx67]); in xnn_f32_raddstoreexpminusmax_ukernel__neon_lut64_p2_x12_acc3() local
104 vl67 = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neon_lut64_p2_x12_acc3()
105 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_raddstoreexpminusmax_ukernel__neon_lut64_p2_x12_acc3()
|
| D | neonfma-lut64-p2-x12-acc3.c | 95 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidx67]); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_lut64_p2_x12_acc3() local
103 vl67 = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_lut64_p2_x12_acc3()
104 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_lut64_p2_x12_acc3()
|
| D | neon-lut64-p2-x12-acc2.c | 95 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidx67]); in xnn_f32_raddstoreexpminusmax_ukernel__neon_lut64_p2_x12_acc2() local
103 vl67 = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neon_lut64_p2_x12_acc2()
104 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_raddstoreexpminusmax_ukernel__neon_lut64_p2_x12_acc2()
|
| /external/XNNPACK/src/f32-sigmoid/gen/ |
| D | neonfma-rr1-lut64-p2-div-x8.c | 60 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x8() local
66 vl67 = vld1_lane_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x8()
67 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x8()
|
| D | neonfma-rr1-lut2048-p1-div-x8.c | 59 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x8() local
65 vl67 = vld1_lane_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x8()
66 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x8()
|
| D | neonfma-rr1-lut64-p2-nr2recps-x8.c | 60 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x8() local
66 vl67 = vld1_lane_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x8()
67 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x8()
|
| D | neon-rr2-lut64-p2-nr2recps-x8.c | 61 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neon_rr2_lut64_p2_nr2recps_x8() local
67 vl67 = vld1_lane_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_sigmoid_ukernel__neon_rr2_lut64_p2_nr2recps_x8()
68 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_sigmoid_ukernel__neon_rr2_lut64_p2_nr2recps_x8()
|
| D | neon-rr2-lut2048-p1-nr2recps-x8.c | 60 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neon_rr2_lut2048_p1_nr2recps_x8() local
66 vl67 = vld1_lane_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_sigmoid_ukernel__neon_rr2_lut2048_p1_nr2recps_x8()
67 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_sigmoid_ukernel__neon_rr2_lut2048_p1_nr2recps_x8()
|
| D | neonfma-rr1-lut64-p2-nr1recps1fma-x8.c | 60 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x8() local
66 vl67 = vld1_lane_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x8()
67 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x8()
|
| D | neonfma-rr1-lut2048-p1-nr2recps-x8.c | 59 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_nr2recps_x8() local
65 vl67 = vld1_lane_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_nr2recps_x8()
66 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_nr2recps_x8()
|
| D | neonfma-rr1-lut2048-p1-nr2fma-x8.c | 59 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_nr2fma_x8() local
65 vl67 = vld1_lane_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_nr2fma_x8()
66 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_nr2fma_x8()
|
| D | neonfma-rr1-lut2048-p1-nr1recps1fma-x8.c | 59 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_nr1recps1fma_x8() local
65 vl67 = vld1_lane_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_nr1recps1fma_x8()
66 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_nr1recps1fma_x8()
|
| D | neonfma-rr1-lut64-p2-nr2fma-x8.c | 60 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x8() local
66 vl67 = vld1_lane_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x8()
67 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x8()
|
| D | neonfma-rr1-lut2048-p1-div-x12.c | 64 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x12() local
74 vl67 = vld1_lane_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x12()
75 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x12()
|
| D | neonfma-rr1-lut64-p2-div-x12.c | 65 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x12() local
75 vl67 = vld1_lane_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x12()
76 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x12()
|
| D | neonfma-rr1-lut2048-p1-nr2recps-x12.c | 64 float32x2_t vl67 = vld1_dup_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) vidx67]); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_nr2recps_x12() local
74 vl67 = vld1_lane_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) (vidx67 >> 32)], vl67, 1); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_nr2recps_x12()
75 const float32x4_t vl4567 = vcombine_f32(vl45, vl67); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_nr2recps_x12()
|
| /external/XNNPACK/src/f32-velu/gen/ |
| D | velu-neonfma-rr1-lut16-p3-x8.c | 69 …int32x2_t vl67 = vld1_dup_s32((const int32_t*) ((uintptr_t) xnn_table_exp2minus_k_over_16 + (uint3… in xnn_f32_velu_ukernel__neonfma_rr1_lut16_p3_x8() local
71 …vl67 = vld1_lane_s32((const int32_t*) ((uintptr_t) xnn_table_exp2minus_k_over_16 + (uint32_t) (vid… in xnn_f32_velu_ukernel__neonfma_rr1_lut16_p3_x8()
72 const int32x4_t vl4567 = vcombine_s32(vl45, vl67); in xnn_f32_velu_ukernel__neonfma_rr1_lut16_p3_x8()
|
| D | velu-neon-rr2-lut16-p3-x8.c | 70 …int32x2_t vl67 = vld1_dup_s32((const int32_t*) ((uintptr_t) xnn_table_exp2minus_k_over_16 + (uint3… in xnn_f32_velu_ukernel__neon_rr2_lut16_p3_x8() local
72 …vl67 = vld1_lane_s32((const int32_t*) ((uintptr_t) xnn_table_exp2minus_k_over_16 + (uint32_t) (vid… in xnn_f32_velu_ukernel__neon_rr2_lut16_p3_x8()
73 const int32x4_t vl4567 = vcombine_s32(vl45, vl67); in xnn_f32_velu_ukernel__neon_rr2_lut16_p3_x8()
|