134         const float32x4_t vk20c0x4567 = vld1q_f32(w + 28);  in xnn_f32_conv_hwc_ukernel_3x3s2p1c3x8__neon_2x2()  local
138         vo0x0c4567 = vmlaq_lane_f32(vo0x0c4567, vk20c0x4567, vget_low_f32(vi2x0), 1);  in xnn_f32_conv_hwc_ukernel_3x3s2p1c3x8__neon_2x2()
139         vo1x0c4567 = vmlaq_lane_f32(vo1x0c4567, vk20c0x4567, vget_low_f32(vi4x0), 1);  in xnn_f32_conv_hwc_ukernel_3x3s2p1c3x8__neon_2x2()
143         vo0x1c4567 = vmlaq_lane_f32(vo0x1c4567, vk20c0x4567, vget_high_f32(vi2x1), 1);  in xnn_f32_conv_hwc_ukernel_3x3s2p1c3x8__neon_2x2()
144         vo1x1c4567 = vmlaq_lane_f32(vo1x1c4567, vk20c0x4567, vget_high_f32(vi4x1), 1);  in xnn_f32_conv_hwc_ukernel_3x3s2p1c3x8__neon_2x2()
580         const float32x4_t vk20c0x4567 = vld1q_f32(w + 28);  in xnn_f32_conv_hwc_ukernel_3x3s2p1c3x8__neon_2x2()  local
584         vo0c4567 = vmlaq_lane_f32(vo0c4567, vk20c0x4567, vget_low_f32(vi2x0), 1);  in xnn_f32_conv_hwc_ukernel_3x3s2p1c3x8__neon_2x2()
585         vo1c4567 = vmlaq_lane_f32(vo1c4567, vk20c0x4567, vget_low_f32(vi4x0), 1);  in xnn_f32_conv_hwc_ukernel_3x3s2p1c3x8__neon_2x2()
868         const float32x4_t vk20c0x4567 = vld1q_f32(w + 28);  in xnn_f32_conv_hwc_ukernel_3x3s2p1c3x8__neon_2x2()  local
872         vo0c4567 = vmlaq_lane_f32(vo0c4567, vk20c0x4567, vget_low_f32(vi2x0), 1);  in xnn_f32_conv_hwc_ukernel_3x3s2p1c3x8__neon_2x2()
873         vo1c4567 = vmlaq_lane_f32(vo1c4567, vk20c0x4567, vget_low_f32(vi4x0), 1);  in xnn_f32_conv_hwc_ukernel_3x3s2p1c3x8__neon_2x2()