A Survey of Some Theoretical Aspects of Multiprocessing
ACM Computing Surveys (CSUR)
ACM Computing Surveys (CSUR)
Construction with parallel derivatives of the closure of a parallel program schema
STOC '74 Proceedings of the sixth annual ACM symposium on Theory of computing
The microprogramming of pipelined processors
ISCA '77 Proceedings of the 4th annual symposium on Computer architecture
Comments on "The Identification of Maximal Parallelism in Straight-Line Microprograms"
IEEE Transactions on Computers
A Theory of Asynchronous Control Networks
IEEE Transactions on Computers
A Programmable Logic Approach for VLSI
IEEE Transactions on Computers
The Identification of Maximal Parallelism in Straight-Line Microprograms
IEEE Transactions on Computers
| Hi-index | 14.98 |
Microprogramming has become a widely used technique which brings versatility to the control unit of a digital system. However, since all microcommands contained in a microinstruction are changed simultaneously, this form of control requires the coding of all possible combinations of parallel commands as separate microinstructions. This causes the resulting microprograms to become space- and timewise inefficient and makes it very difficult to control and/or synchronize independent processes. A new approach to modeling and implementing the flow of control in hardware systems is presented in this paper. Here a control flow is realized by a set of independent control operators that pass control to each other according to the precedence relationships defined by the control function which is implemented. The detailed architecture of a VLSI chip to implement a control unit as a hard- programmable control unit (HPCU) is also given. An HPCU contains many independent control operators and allows mask-programmable connections to be made among the operators to implement any given control function. Examples included in the paper demonstrate the effectiveness of an HPCU in implementing complex parallel control flows.