197 …const float16x8_t vb01234567c3 = vld1q_f16(w); w = (const void*) ((uintptr_t) w + sizeof(float16x8…  in xnn_f16_gemm_minmax_ukernel_8x8__neonfp16arith_ld64()  local
200         vacc0x01234567 = vfmaq_lane_f16(vacc0x01234567, vb01234567c3, va0, 3);  in xnn_f16_gemm_minmax_ukernel_8x8__neonfp16arith_ld64()
201         vacc1x01234567 = vfmaq_lane_f16(vacc1x01234567, vb01234567c3, va1, 3);  in xnn_f16_gemm_minmax_ukernel_8x8__neonfp16arith_ld64()
202         vacc2x01234567 = vfmaq_lane_f16(vacc2x01234567, vb01234567c3, va2, 3);  in xnn_f16_gemm_minmax_ukernel_8x8__neonfp16arith_ld64()
203         vacc3x01234567 = vfmaq_lane_f16(vacc3x01234567, vb01234567c3, va3, 3);  in xnn_f16_gemm_minmax_ukernel_8x8__neonfp16arith_ld64()
204         vacc4x01234567 = vfmaq_lane_f16(vacc4x01234567, vb01234567c3, va4, 3);  in xnn_f16_gemm_minmax_ukernel_8x8__neonfp16arith_ld64()
205         vacc5x01234567 = vfmaq_lane_f16(vacc5x01234567, vb01234567c3, va5, 3);  in xnn_f16_gemm_minmax_ukernel_8x8__neonfp16arith_ld64()
206         vacc6x01234567 = vfmaq_lane_f16(vacc6x01234567, vb01234567c3, va6, 3);  in xnn_f16_gemm_minmax_ukernel_8x8__neonfp16arith_ld64()
207         vacc7x01234567 = vfmaq_lane_f16(vacc7x01234567, vb01234567c3, va7, 3);  in xnn_f16_gemm_minmax_ukernel_8x8__neonfp16arith_ld64()
218         vacc0x01234567 = vfmaq_f16(vacc0x01234567, va0c3, vb01234567c3);  in xnn_f16_gemm_minmax_ukernel_8x8__neonfp16arith_ld64()
219         vacc1x01234567 = vfmaq_f16(vacc1x01234567, va1c3, vb01234567c3);  in xnn_f16_gemm_minmax_ukernel_8x8__neonfp16arith_ld64()
220         vacc2x01234567 = vfmaq_f16(vacc2x01234567, va2c3, vb01234567c3);  in xnn_f16_gemm_minmax_ukernel_8x8__neonfp16arith_ld64()
221         vacc3x01234567 = vfmaq_f16(vacc3x01234567, va3c3, vb01234567c3);  in xnn_f16_gemm_minmax_ukernel_8x8__neonfp16arith_ld64()
222         vacc4x01234567 = vfmaq_f16(vacc4x01234567, va4c3, vb01234567c3);  in xnn_f16_gemm_minmax_ukernel_8x8__neonfp16arith_ld64()
223         vacc5x01234567 = vfmaq_f16(vacc5x01234567, va5c3, vb01234567c3);  in xnn_f16_gemm_minmax_ukernel_8x8__neonfp16arith_ld64()
224         vacc6x01234567 = vfmaq_f16(vacc6x01234567, va6c3, vb01234567c3);  in xnn_f16_gemm_minmax_ukernel_8x8__neonfp16arith_ld64()
225         vacc7x01234567 = vfmaq_f16(vacc7x01234567, va7c3, vb01234567c3);  in xnn_f16_gemm_minmax_ukernel_8x8__neonfp16arith_ld64()