muFP(ID:2261/muf001)

VLSI functional language 


for micro FP (actually spelt with the greek letter)

Functional language for hardware design, predecessor to Ruby[1].

Structures:
Related languages
FP => muFP   Evolution of
muFP => Ruby   Evolution of
References:
  • Sheeran, M. "muFP - an algebraic VLSI design language," Technical Monograph PRG-39, Oxford University Computing Laboratory, University of Oxford, Sep. 1984. view details
  • Sheeran, Mary: "muFP, a language for VLSI design" view details
          in Proceedings of the 1984 ACM Symposium on LISP and functional programming, , August 06-08, 1984, Austin, Texas, view details
  • Moreno, Jaime H. "A Proposal for the Systematic Design of Arrays for Matrix Computations" Computer Science Department University of California, Los Angeles TEchnical Report No. CSD-870019 May 1987 view details Abstract: We propose to develop a general and systematic methodology for the design of matrix solvers, based on the dependence graph of the algorithms. A fully-parallel graph is transformed to incorporate issues such as data broadcasting and synchronization, interconnection structure, Ii0 bandwidth, number and utilization of PEs, throughput, delay, and the capability to solve problems larger than the size of the array. The objective is to devise a methodology which handles and relates features of the algorithm and the implementation, in a unified manner. This methodology assists a designer in selecting transformations to an algorithm from a set of feasible ones, and in evaluating the resulting implementations. This research is motivated by the lack of an adequate design methodology for matrix computations. Standard structures (systolic arrays) have been used for these implementations, but they might be non-optimal for a particular algorithm. Reported systems have used ad-hoc design approaches. Some design methodologies have been proposed, but they do not address many important issues. A preliminary version of the proposed methodology has been applied to algorithms for matrix multiplication and LU--decomposition. The approach produces structures which correspond to proposed systolic arrays for these computations, as well as structures which exhibit better efficiency than those arrays. The results show that different transformations on a graph may lead to entirely different computing structures. The selection of an adequate transformation is thus directed by the specific restrictions and performance objectives imposed on the implementation. The designer can identify and manipulate the parameters that are more relevant to a given application.
          in Proceedings of the 1984 ACM Symposium on LISP and functional programming, , August 06-08, 1984, Austin, Texas, view details