/external/XNNPACK/src/f32-raddstoreexpminusmax/gen/ |
D | neon-rr2-lut64-p2-x16.c | 81 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidxEF]); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x16() local 93 vlEF = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x16() 94 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x16()
|
D | neonfma-rr1-lut64-p2-x16-acc4.c | 83 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidxEF]); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x16_acc4() local 95 vlEF = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x16_acc4() 96 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x16_acc4()
|
D | neonfma-rr1-lut64-p2-x16-acc2.c | 81 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidxEF]); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x16_acc2() local 93 vlEF = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x16_acc2() 94 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x16_acc2()
|
D | neonfma-rr1-lut64-p2-x16.c | 80 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidxEF]); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x16() local 92 vlEF = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x16() 93 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x16()
|
D | neon-rr2-lut64-p2-x16-acc2.c | 82 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidxEF]); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x16_acc2() local 94 vlEF = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x16_acc2() 95 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x16_acc2()
|
D | neon-rr2-lut64-p2-x16-acc4.c | 84 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidxEF]); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x16_acc4() local 96 vlEF = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x16_acc4() 97 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x16_acc4()
|
D | neon-rr2-lut64-p2-x20.c | 88 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidxEF]); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x20() local 102 vlEF = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x20() 103 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x20()
|
D | neon-rr2-lut64-p2-x20-acc2.c | 89 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidxEF]); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x20_acc2() local 103 vlEF = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x20_acc2() 104 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x20_acc2()
|
D | neonfma-rr1-lut64-p2-x20.c | 87 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidxEF]); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x20() local 101 vlEF = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x20() 102 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x20()
|
D | neonfma-rr1-lut64-p2-x20-acc2.c | 88 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidxEF]); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x20_acc2() local 102 vlEF = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x20_acc2() 103 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x20_acc2()
|
D | neonfma-rr1-lut64-p2-x20-acc5.c | 91 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidxEF]); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x20_acc5() local 105 vlEF = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x20_acc5() 106 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_raddstoreexpminusmax_ukernel__neonfma_rr1_lut64_p2_x20_acc5()
|
D | neon-rr2-lut64-p2-x20-acc5.c | 92 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2_k_over_64[(uint32_t) vidxEF]); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x20_acc5() local 106 vlEF = vld1_lane_f32(&xnn_table_exp2_k_over_64[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x20_acc5() 107 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_raddstoreexpminusmax_ukernel__neon_rr2_lut64_p2_x20_acc5()
|
/external/XNNPACK/src/f32-vsigmoid/gen/ |
D | vsigmoid-neonfma-rr1-lut64-p2-div-x16.c | 78 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxEF]); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut64_p2_div_x16() local 90 vlEF = vld1_lane_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut64_p2_div_x16() 91 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut64_p2_div_x16()
|
D | vsigmoid-neonfma-rr1-lut2048-p1-div-x16.c | 77 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) vidxEF]); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x16() local 89 vlEF = vld1_lane_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x16() 90 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x16()
|
D | vsigmoid-neonfma-rr1-lut2048-p1-nr2recps-x16.c | 77 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) vidxEF]); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut2048_p1_nr2recps_x16() local 89 vlEF = vld1_lane_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut2048_p1_nr2recps_x16() 90 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut2048_p1_nr2recps_x16()
|
D | vsigmoid-neonfma-rr1-lut64-p2-nr1recps1fma-x16.c | 78 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxEF]); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x16() local 90 vlEF = vld1_lane_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x16() 91 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut64_p2_nr1recps1fma_x16()
|
D | vsigmoid-neonfma-rr1-lut2048-p1-nr2fma-x16.c | 77 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) vidxEF]); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut2048_p1_nr2fma_x16() local 89 vlEF = vld1_lane_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut2048_p1_nr2fma_x16() 90 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut2048_p1_nr2fma_x16()
|
D | vsigmoid-neonfma-rr1-lut64-p2-nr2fma-x16.c | 78 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxEF]); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x16() local 90 vlEF = vld1_lane_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x16() 91 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut64_p2_nr2fma_x16()
|
D | vsigmoid-neonfma-rr1-lut2048-p1-nr1recps1fma-x16.c | 77 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) vidxEF]); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut2048_p1_nr1recps1fma_x16() local 89 vlEF = vld1_lane_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut2048_p1_nr1recps1fma_x16() 90 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut2048_p1_nr1recps1fma_x16()
|
D | vsigmoid-neonfma-rr1-lut2048-p1-div-x20.c | 82 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) vidxEF]); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x20() local 98 vlEF = vld1_lane_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x20() 99 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x20()
|
D | vsigmoid-neonfma-rr1-lut64-p2-nr2recps-x16.c | 78 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxEF]); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x16() local 90 vlEF = vld1_lane_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x16() 91 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_vsigmoid_ukernel__neonfma_rr1_lut64_p2_nr2recps_x16()
|
D | vsigmoid-neon-rr2-lut2048-p1-nr2recps-x16.c | 78 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) vidxEF]); in xnn_f32_vsigmoid_ukernel__neon_rr2_lut2048_p1_nr2recps_x16() local 90 vlEF = vld1_lane_f32(&xnn_table_exp2minus_k_over_2048[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_vsigmoid_ukernel__neon_rr2_lut2048_p1_nr2recps_x16() 91 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_vsigmoid_ukernel__neon_rr2_lut2048_p1_nr2recps_x16()
|
D | vsigmoid-neon-rr2-lut64-p2-nr2recps-x16.c | 79 float32x2_t vlEF = vld1_dup_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) vidxEF]); in xnn_f32_vsigmoid_ukernel__neon_rr2_lut64_p2_nr2recps_x16() local 91 vlEF = vld1_lane_f32(&xnn_table_exp2minus_k_over_64[(uint32_t) (vidxEF >> 32)], vlEF, 1); in xnn_f32_vsigmoid_ukernel__neon_rr2_lut64_p2_nr2recps_x16() 92 const float32x4_t vlCDEF = vcombine_f32(vlCD, vlEF); in xnn_f32_vsigmoid_ukernel__neon_rr2_lut64_p2_nr2recps_x16()
|
/external/XNNPACK/src/f32-velu/gen/ |
D | velu-neon-rr2-lut16-p3-x16.c | 93 …int32x2_t vlEF = vld1_dup_s32((const int32_t*) ((uintptr_t) xnn_table_exp2minus_k_over_16 + (uint3… in xnn_f32_velu_ukernel__neon_rr2_lut16_p3_x16() local 95 …vlEF = 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_x16() 96 const int32x4_t vlCDEF = vcombine_s32(vlCD, vlEF); in xnn_f32_velu_ukernel__neon_rr2_lut16_p3_x16()
|
D | velu-neonfma-rr1-lut16-p3-x16.c | 92 …int32x2_t vlEF = vld1_dup_s32((const int32_t*) ((uintptr_t) xnn_table_exp2minus_k_over_16 + (uint3… in xnn_f32_velu_ukernel__neonfma_rr1_lut16_p3_x16() local 94 …vlEF = 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_x16() 95 const int32x4_t vlCDEF = vcombine_s32(vlCD, vlEF); in xnn_f32_velu_ukernel__neonfma_rr1_lut16_p3_x16()
|