Priority arbiters

The paper presents asynchronous design solutions to the problem of Priority Arbitration which is defined in the following form. A system consists of multiple, physically concurrent, processes with a shared resource. The discipline of resource allocation is a function of parameters of the active requests, which are assigned to the requests either statically or dynamically. This function can be defined in an (arbitrary) combinatorial way (contrary to conventional, `topological´, mappings, such as that used in a daisy-chain arbiter). The proposed designs are quasi-speed-independent. Furthermore, the priority logic, in the dynamic case, has the following architectural feature: it is a tree structure in which the control flow is maximally decoupled from the data-path by means of an early propagation of the `valid´-`invalid´ signals, concurrently, with processing the priority data. This lends to significant reduction in the overall arbitration delay when the number of active requests is low