/external/XNNPACK/src/f32-gemm/gen/ |
D | 1x8-neon-dup-ld64.c | 59 const float32x4_t va0c1 = vdupq_lane_f32(va0, 1); in xnn_f32_gemm_ukernel_1x8__neon_dup_ld64() local 60 vacc0x0123 = vmlaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_gemm_ukernel_1x8__neon_dup_ld64() 61 vacc0x4567 = vmlaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_gemm_ukernel_1x8__neon_dup_ld64()
|
D | 1x8-neonfma-dup-ld64.c | 59 const float32x4_t va0c1 = vdupq_lane_f32(va0, 1); in xnn_f32_gemm_ukernel_1x8__neonfma_dup_ld64() local 60 vacc0x0123 = vfmaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_gemm_ukernel_1x8__neonfma_dup_ld64() 61 vacc0x4567 = vfmaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_gemm_ukernel_1x8__neonfma_dup_ld64()
|
D | 1x8-psimd-splat.c | 58 const psimd_f32 va0c1 = psimd_splat1_f32(va0); in xnn_f32_gemm_ukernel_1x8__psimd_splat() local 63 vacc0x0123 = psimd_qfma_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_gemm_ukernel_1x8__psimd_splat() 64 vacc0x4567 = psimd_qfma_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_gemm_ukernel_1x8__psimd_splat()
|
D | 4x8-neon-dup-ld64.c | 95 const float32x4_t va0c1 = vdupq_lane_f32(va0, 1); in xnn_f32_gemm_ukernel_4x8__neon_dup_ld64() local 99 vacc0x0123 = vmlaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_gemm_ukernel_4x8__neon_dup_ld64() 103 vacc0x4567 = vmlaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_gemm_ukernel_4x8__neon_dup_ld64()
|
D | 4x8-neonfma-dup-ld64.c | 95 const float32x4_t va0c1 = vdupq_lane_f32(va0, 1); in xnn_f32_gemm_ukernel_4x8__neonfma_dup_ld64() local 99 vacc0x0123 = vfmaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_gemm_ukernel_4x8__neonfma_dup_ld64() 103 vacc0x4567 = vfmaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_gemm_ukernel_4x8__neonfma_dup_ld64()
|
D | 4x8-psimd-splat.c | 97 const psimd_f32 va0c1 = psimd_splat1_f32(va0); in xnn_f32_gemm_ukernel_4x8__psimd_splat() local 105 vacc0x0123 = psimd_qfma_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_gemm_ukernel_4x8__psimd_splat() 109 vacc0x4567 = psimd_qfma_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_gemm_ukernel_4x8__psimd_splat()
|
D | 4x8-neonfma-dup-ld128.c | 97 const float32x4_t va0c1 = vdupq_lane_f32(vget_low_f32(va0), 1); in xnn_f32_gemm_ukernel_4x8__neonfma_dup_ld128() local 101 vacc0x0123 = vfmaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_gemm_ukernel_4x8__neonfma_dup_ld128() 105 vacc0x4567 = vfmaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_gemm_ukernel_4x8__neonfma_dup_ld128()
|
D | 4x8-neon-dup-ld128.c | 97 const float32x4_t va0c1 = vdupq_lane_f32(vget_low_f32(va0), 1); in xnn_f32_gemm_ukernel_4x8__neon_dup_ld128() local 101 vacc0x0123 = vmlaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_gemm_ukernel_4x8__neon_dup_ld128() 105 vacc0x4567 = vmlaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_gemm_ukernel_4x8__neon_dup_ld128()
|
D | 4x2-neonfma-lane-ld64.c | 98 const float32x2_t va0c1 = vdup_lane_f32(va0, 1); in xnn_f32_gemm_ukernel_4x2__neonfma_lane_ld64() local 102 vacc0x01 = vfma_f32(vacc0x01, va0c1, vb01c1); in xnn_f32_gemm_ukernel_4x2__neonfma_lane_ld64()
|
/external/XNNPACK/src/f32-gemm/gen-inc/ |
D | 1x8-neon-dup-ld64.c | 61 const float32x4_t va0c1 = vdupq_lane_f32(va0, 1); in xnn_f32_gemminc_ukernel_1x8__neon_dup_ld64() local 62 vacc0x0123 = vmlaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_gemminc_ukernel_1x8__neon_dup_ld64() 63 vacc0x4567 = vmlaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_gemminc_ukernel_1x8__neon_dup_ld64()
|
D | 1x8-neonfma-dup-ld64.c | 61 const float32x4_t va0c1 = vdupq_lane_f32(va0, 1); in xnn_f32_gemminc_ukernel_1x8__neonfma_dup_ld64() local 62 vacc0x0123 = vfmaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_gemminc_ukernel_1x8__neonfma_dup_ld64() 63 vacc0x4567 = vfmaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_gemminc_ukernel_1x8__neonfma_dup_ld64()
|
D | 1x8-psimd-splat.c | 60 const psimd_f32 va0c1 = psimd_splat1_f32(va0); in xnn_f32_gemminc_ukernel_1x8__psimd_splat() local 65 vacc0x0123 = psimd_qfma_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_gemminc_ukernel_1x8__psimd_splat() 66 vacc0x4567 = psimd_qfma_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_gemminc_ukernel_1x8__psimd_splat()
|
D | 4x8-neon-dup-ld64.c | 97 const float32x4_t va0c1 = vdupq_lane_f32(va0, 1); in xnn_f32_gemminc_ukernel_4x8__neon_dup_ld64() local 101 vacc0x0123 = vmlaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_gemminc_ukernel_4x8__neon_dup_ld64() 105 vacc0x4567 = vmlaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_gemminc_ukernel_4x8__neon_dup_ld64()
|
D | 4x8-neonfma-dup-ld64.c | 97 const float32x4_t va0c1 = vdupq_lane_f32(va0, 1); in xnn_f32_gemminc_ukernel_4x8__neonfma_dup_ld64() local 101 vacc0x0123 = vfmaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_gemminc_ukernel_4x8__neonfma_dup_ld64() 105 vacc0x4567 = vfmaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_gemminc_ukernel_4x8__neonfma_dup_ld64()
|
D | 4x8-neonfma-dup-ld128.c | 99 const float32x4_t va0c1 = vdupq_lane_f32(vget_low_f32(va0), 1); in xnn_f32_gemminc_ukernel_4x8__neonfma_dup_ld128() local 103 vacc0x0123 = vfmaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_gemminc_ukernel_4x8__neonfma_dup_ld128() 107 vacc0x4567 = vfmaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_gemminc_ukernel_4x8__neonfma_dup_ld128()
|
D | 4x8-psimd-splat.c | 99 const psimd_f32 va0c1 = psimd_splat1_f32(va0); in xnn_f32_gemminc_ukernel_4x8__psimd_splat() local 107 vacc0x0123 = psimd_qfma_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_gemminc_ukernel_4x8__psimd_splat() 111 vacc0x4567 = psimd_qfma_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_gemminc_ukernel_4x8__psimd_splat()
|
D | 4x8-neon-dup-ld128.c | 99 const float32x4_t va0c1 = vdupq_lane_f32(vget_low_f32(va0), 1); in xnn_f32_gemminc_ukernel_4x8__neon_dup_ld128() local 103 vacc0x0123 = vmlaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_gemminc_ukernel_4x8__neon_dup_ld128() 107 vacc0x4567 = vmlaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_gemminc_ukernel_4x8__neon_dup_ld128()
|
/external/XNNPACK/src/f32-igemm/gen/ |
D | 1x8-neon-dup-ld64.c | 72 const float32x4_t va0c1 = vdupq_lane_f32(va0, 1); in xnn_f32_igemm_ukernel_1x8__neon_dup_ld64() local 73 vacc0x0123 = vmlaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_igemm_ukernel_1x8__neon_dup_ld64() 74 vacc0x4567 = vmlaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_igemm_ukernel_1x8__neon_dup_ld64()
|
D | 1x8-neonfma-dup-ld64.c | 72 const float32x4_t va0c1 = vdupq_lane_f32(va0, 1); in xnn_f32_igemm_ukernel_1x8__neonfma_dup_ld64() local 73 vacc0x0123 = vfmaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_igemm_ukernel_1x8__neonfma_dup_ld64() 74 vacc0x4567 = vfmaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_igemm_ukernel_1x8__neonfma_dup_ld64()
|
D | 1x8-psimd-splat.c | 71 const psimd_f32 va0c1 = psimd_splat1_f32(va0); in xnn_f32_igemm_ukernel_1x8__psimd_splat() local 76 vacc0x0123 = psimd_qfma_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_igemm_ukernel_1x8__psimd_splat() 77 vacc0x4567 = psimd_qfma_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_igemm_ukernel_1x8__psimd_splat()
|
D | 4x8-neon-dup-ld64.c | 117 const float32x4_t va0c1 = vdupq_lane_f32(va0, 1); in xnn_f32_igemm_ukernel_4x8__neon_dup_ld64() local 121 vacc0x0123 = vmlaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_igemm_ukernel_4x8__neon_dup_ld64() 125 vacc0x4567 = vmlaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_igemm_ukernel_4x8__neon_dup_ld64()
|
D | 4x8-neonfma-dup-ld64.c | 117 const float32x4_t va0c1 = vdupq_lane_f32(va0, 1); in xnn_f32_igemm_ukernel_4x8__neonfma_dup_ld64() local 121 vacc0x0123 = vfmaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_igemm_ukernel_4x8__neonfma_dup_ld64() 125 vacc0x4567 = vfmaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_igemm_ukernel_4x8__neonfma_dup_ld64()
|
D | 4x8-psimd-splat.c | 119 const psimd_f32 va0c1 = psimd_splat1_f32(va0); in xnn_f32_igemm_ukernel_4x8__psimd_splat() local 127 vacc0x0123 = psimd_qfma_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_igemm_ukernel_4x8__psimd_splat() 131 vacc0x4567 = psimd_qfma_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_igemm_ukernel_4x8__psimd_splat()
|
D | 4x8-neonfma-dup-ld128.c | 119 const float32x4_t va0c1 = vdupq_lane_f32(vget_low_f32(va0), 1); in xnn_f32_igemm_ukernel_4x8__neonfma_dup_ld128() local 123 vacc0x0123 = vfmaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_igemm_ukernel_4x8__neonfma_dup_ld128() 127 vacc0x4567 = vfmaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_igemm_ukernel_4x8__neonfma_dup_ld128()
|
D | 4x8-neon-dup-ld128.c | 119 const float32x4_t va0c1 = vdupq_lane_f32(vget_low_f32(va0), 1); in xnn_f32_igemm_ukernel_4x8__neon_dup_ld128() local 123 vacc0x0123 = vmlaq_f32(vacc0x0123, va0c1, vb0123c1); in xnn_f32_igemm_ukernel_4x8__neon_dup_ld128() 127 vacc0x4567 = vmlaq_f32(vacc0x4567, va0c1, vb4567c1); in xnn_f32_igemm_ukernel_4x8__neon_dup_ld128()
|