Date: Mon, 29 Sep 2003 19:01:12 -0500 (CDT) From: Yale Patt To: ee360n-15280@sunfire2.ece.utexas.edu Subject: a student asks... Cc: lynch@sunfire2.ece.utexas.edu, santhosh@sunfire2.ece.utexas.edu Dr Patt and the TAs, State 12 (PC <- BaseR) can be implemented in two differebt ways on the data path. I have listed both implementations below: 1. Get the BaseR from Reg File, pass it through ALU, send it through the bus to PCMUX, and finally place it in PC. 2. Get the BaseR from Reg File, send it through the ADDR1MUX to the adder, at the same time select 0 at ADDR2MUX and send to adder, send BaseR + 0 to PCMUX, and finally place the result in PC. Obviously, these two implementations results in different micro-instructions. That is true. Is there any adv. or disadv. in using either implementatoin? In the case in point, no difference. However, sometimes there is. Let me ask you if you can find the advantage one has over the other. One of the good things about the microarchitecture is that you can specify individual control signals so that sometimes you can do more than one microoperation at the same time. What other thing might you want to do that is possible with one of your two solutions but not possible with the other? Thank you <>