Functions executed by a multi-functional processing unit (PU) correspond to clusters of operations in the specification, which are represented as Signal Flow Graphs (SFGs). Because of high-throughput demands, the operations of each SFG are executed in parallel. Since only one of the SFGs is executed at the same time, operations belonging to dtzerent SFGs can be executed on the same operator. In this paper, we concentrate on the most important part of the mapping of several SFGs onto one PU, which is the assignment of the SFGs’ operations to the Pu’s operators, given a number of allocated operators. The problem is to find an operator assignment that minimizes the silicon area that is occupied by the PU’s interconnection consisting of multiplexers and wires. Here, we present an approach based on local search algorithms such as iterative improvement and simulated annealing. Although these algorithms are known to be generally applicable, we show that detailed knowledge of the operator assignment problem is required to obtain good results within acceptable CPU time limits for large problem instances.
[1]
E. de Niet,et al.
An experimental digital VCR with 40 mm drum, single actuator and DCT-based bit-rate reduction
,
1988
.
[2]
A. van der Wed.
PHIDEO: A Silicon Compiler for High Speed Algorithms
,
1991
.
[3]
Jan M. Rabaey,et al.
Hardware selection and clustering in the HYPER synthesis system
,
1992,
[1992] Proceedings The European Conference on Design Automation.
[4]
Fadi J. Kurdahi,et al.
Register-Transfer Synthesis of Pipelined Data Paths
,
1994
.
[5]
Srinivas Devadas,et al.
Algorithms for hardware allocation in data path synthesis
,
1989,
IEEE Trans. Comput. Aided Des. Integr. Circuits Syst..