DVD - Lecture 5d: STA Example

[@socialogic9777]

Sir, I'm confused when deciding RAT. RAT at a node is min[ RAT(successor node) - delay( node to successor node ]. From this relation, the minimum delay from node to successor node should be considered. In the lecture you have taken the worst case i.e max delay. kindly, correct me

[@AdiTeman]

Hi Socialogic, Yes, I see why this could be a bit confusing. The "min" is around the entire expression (as you wrote). So we're looking for the RAT that has the minimum value, in other words, what is the **earliest** time that the data needs to arrive at that node to make it to the end in time. That means that we are looking for "how much time do we need from this point and on to finish our operation". The thing that sets this is the longest delay from this point to the end. So we are looking for the maximum delay from this point to the end. Because the RAT takes the end as a reference - looking backwards - we are looking for the minimum of all these calculations. But we take the maximum arcs from here to the end. So the slides are correct. You can see this visually on slide 23 (www.eng.biu.ac.il/temanad/files/2018/12/Lecture-5-STA.pdf). To quote the slide: "RAT: longest logic delay to the capture edge of the clock (dependent on cycle time) Hope that helps understand it better.