68     float16x8_t vacc2x89ABCDEF = vacc0x89ABCDEF;  in xnn_f16_gemm_minmax_ukernel_4x16__neonfp16arith_ld64()  local
89         vacc2x89ABCDEF = vfmaq_lane_f16(vacc2x89ABCDEF, vb89ABCDEFc0, va2, 0);  in xnn_f16_gemm_minmax_ukernel_4x16__neonfp16arith_ld64()
103         vacc2x89ABCDEF = vfmaq_f16(vacc2x89ABCDEF, va2c0, vb89ABCDEFc0);  in xnn_f16_gemm_minmax_ukernel_4x16__neonfp16arith_ld64()
116         vacc2x89ABCDEF = vfmaq_lane_f16(vacc2x89ABCDEF, vb89ABCDEFc1, va2, 1);  in xnn_f16_gemm_minmax_ukernel_4x16__neonfp16arith_ld64()
130         vacc2x89ABCDEF = vfmaq_f16(vacc2x89ABCDEF, va2c1, vb89ABCDEFc1);  in xnn_f16_gemm_minmax_ukernel_4x16__neonfp16arith_ld64()
143         vacc2x89ABCDEF = vfmaq_lane_f16(vacc2x89ABCDEF, vb89ABCDEFc2, va2, 2);  in xnn_f16_gemm_minmax_ukernel_4x16__neonfp16arith_ld64()
157         vacc2x89ABCDEF = vfmaq_f16(vacc2x89ABCDEF, va2c2, vb89ABCDEFc2);  in xnn_f16_gemm_minmax_ukernel_4x16__neonfp16arith_ld64()
170         vacc2x89ABCDEF = vfmaq_lane_f16(vacc2x89ABCDEF, vb89ABCDEFc3, va2, 3);  in xnn_f16_gemm_minmax_ukernel_4x16__neonfp16arith_ld64()
184         vacc2x89ABCDEF = vfmaq_f16(vacc2x89ABCDEF, va2c3, vb89ABCDEFc3);  in xnn_f16_gemm_minmax_ukernel_4x16__neonfp16arith_ld64()
206         vacc2x89ABCDEF = vfmaq_f16(vacc2x89ABCDEF, va2, vb89ABCDEF);  in xnn_f16_gemm_minmax_ukernel_4x16__neonfp16arith_ld64()
220     vacc2x89ABCDEF = vmulq_f16(vacc2x89ABCDEF, vscale);  in xnn_f16_gemm_minmax_ukernel_4x16__neonfp16arith_ld64()
230     vacc2x89ABCDEF = vminq_f16(vacc2x89ABCDEF, vmax);  in xnn_f16_gemm_minmax_ukernel_4x16__neonfp16arith_ld64()
240     vacc2x89ABCDEF = vmaxq_f16(vacc2x89ABCDEF, vmin);  in xnn_f16_gemm_minmax_ukernel_4x16__neonfp16arith_ld64()
251       vst1q_f16(c2 + 8, vacc2x89ABCDEF);  in xnn_f16_gemm_minmax_ukernel_4x16__neonfp16arith_ld64()
272         vacc2x01234567 = vacc2x89ABCDEF;  in xnn_f16_gemm_minmax_ukernel_4x16__neonfp16arith_ld64()