For AXI Blocking Mode, the minimum latency is reduced by six cycles. The resource cost of Blocking mode is also reduced versus v4.0. This saving is proportional to the total number of input bits. NonBlocking behavior has also been changed relative to v4.0 to further reduce the resource cost of moving to AXI from pre-AXI versions such as v3.1. Latency in v6.0 is one cycle less than for v4.0 in NonBlocking mode except for the zero latency case, which is still zero.