Lines Matching refs:vxa4
98 const int16x8_t vxa4 = vmovl_s8(va4); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm() local
113 vacc4x0123 = vmlal_lane_s16(vacc4x0123, vget_low_s16(vxb01234567c0), vget_low_s16(vxa4), 0); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
114 vacc4x4567 = vmlal_lane_s16(vacc4x4567, vget_high_s16(vxb01234567c0), vget_low_s16(vxa4), 0); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
128 vacc4x0123 = vmlal_lane_s16(vacc4x0123, vget_low_s16(vxb01234567c1), vget_low_s16(vxa4), 1); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
129 vacc4x4567 = vmlal_lane_s16(vacc4x4567, vget_high_s16(vxb01234567c1), vget_low_s16(vxa4), 1); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
143 vacc4x0123 = vmlal_lane_s16(vacc4x0123, vget_low_s16(vxb01234567c2), vget_low_s16(vxa4), 2); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
144 vacc4x4567 = vmlal_lane_s16(vacc4x4567, vget_high_s16(vxb01234567c2), vget_low_s16(vxa4), 2); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
158 vacc4x0123 = vmlal_lane_s16(vacc4x0123, vget_low_s16(vxb01234567c3), vget_low_s16(vxa4), 3); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
159 vacc4x4567 = vmlal_lane_s16(vacc4x4567, vget_high_s16(vxb01234567c3), vget_low_s16(vxa4), 3); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
176 vacc4x0123 = vmlal_lane_s16(vacc4x0123, vget_low_s16(vxb01234567c4), vget_high_s16(vxa4), 0); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
177 vacc4x4567 = vmlal_lane_s16(vacc4x4567, vget_high_s16(vxb01234567c4), vget_high_s16(vxa4), 0); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
191 vacc4x0123 = vmlal_lane_s16(vacc4x0123, vget_low_s16(vxb01234567c5), vget_high_s16(vxa4), 1); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
192 vacc4x4567 = vmlal_lane_s16(vacc4x4567, vget_high_s16(vxb01234567c5), vget_high_s16(vxa4), 1); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
206 vacc4x0123 = vmlal_lane_s16(vacc4x0123, vget_low_s16(vxb01234567c6), vget_high_s16(vxa4), 2); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
207 vacc4x4567 = vmlal_lane_s16(vacc4x4567, vget_high_s16(vxb01234567c6), vget_high_s16(vxa4), 2); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
221 vacc4x0123 = vmlal_lane_s16(vacc4x0123, vget_low_s16(vxb01234567c7), vget_high_s16(vxa4), 3); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
222 vacc4x4567 = vmlal_lane_s16(vacc4x4567, vget_high_s16(vxb01234567c7), vget_high_s16(vxa4), 3); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
238 const int16x8_t vxa4 = vmovl_s8(va4); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm() local
253 vacc4x0123 = vmlal_lane_s16(vacc4x0123, vget_low_s16(vxb01234567c0), vget_low_s16(vxa4), 0); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
254 vacc4x4567 = vmlal_lane_s16(vacc4x4567, vget_high_s16(vxb01234567c0), vget_low_s16(vxa4), 0); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
270 vacc4x0123 = vmlal_lane_s16(vacc4x0123, vget_low_s16(vxb01234567c1), vget_low_s16(vxa4), 1); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
271 … vacc4x4567 = vmlal_lane_s16(vacc4x4567, vget_high_s16(vxb01234567c1), vget_low_s16(vxa4), 1); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
287 … vacc4x0123 = vmlal_lane_s16(vacc4x0123, vget_low_s16(vxb01234567c2), vget_low_s16(vxa4), 2); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
288 … vacc4x4567 = vmlal_lane_s16(vacc4x4567, vget_high_s16(vxb01234567c2), vget_low_s16(vxa4), 2); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
304 … vacc4x0123 = vmlal_lane_s16(vacc4x0123, vget_low_s16(vxb01234567c3), vget_low_s16(vxa4), 3); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
305 … vacc4x4567 = vmlal_lane_s16(vacc4x4567, vget_high_s16(vxb01234567c3), vget_low_s16(vxa4), 3); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
321 … vacc4x0123 = vmlal_lane_s16(vacc4x0123, vget_low_s16(vxb01234567c4), vget_high_s16(vxa4), 0); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
322 … vacc4x4567 = vmlal_lane_s16(vacc4x4567, vget_high_s16(vxb01234567c4), vget_high_s16(vxa4), 0); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
338 … vacc4x0123 = vmlal_lane_s16(vacc4x0123, vget_low_s16(vxb01234567c5), vget_high_s16(vxa4), 1); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
339 … vacc4x4567 = vmlal_lane_s16(vacc4x4567, vget_high_s16(vxb01234567c5), vget_high_s16(vxa4), 1); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
355 … vacc4x0123 = vmlal_lane_s16(vacc4x0123, vget_low_s16(vxb01234567c6), vget_high_s16(vxa4), 2); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()
356 … vacc4x4567 = vmlal_lane_s16(vacc4x4567, vget_high_s16(vxb01234567c6), vget_high_s16(vxa4), 2); in xnn_qc8_gemm_minmax_fp32_ukernel_6x8__neonv8_mlal_lane_prfm()