TEMPT: technology mapping for the exploration of FPGA architectures with hard-wired connections
- 2 January 2003
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 361-367
- https://doi.org/10.1109/dac.1992.227777
Abstract
TEMPT is a technology mapping algorithm aimed at exploring field-programmable gate array (FPGA) architectures with hard-wired connections. TEMPT maps a network of basic blocks to a netlist of hard-wired logic blocks (HLBs), in which each HLB consists of several basic hard-wire blocks connected in an arbitrary tree topology, and optimizes either speed or area. TEMPT is as effective as the Xilinx 4000 CLB mapper, PPR, when minimizing CLBs to implement a set of MCNC benchmarks. Using TEMPT it was shown empirically that many HLBs were significantly faster than FPGAs without hard-wired links. Several HLBs were demonstrated that exhibited superior logic density to the Xilinx 4000 CLB.<>Keywords
This publication has 8 references indexed in Scilit:
- Optimization of field-programmable gate array logic block architecture for speedPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- Third-generation architecture boosts speed and density of field-programmable gate arraysPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- Technology mapping of lookup table-based FPGAs for performancePublished by Institute of Electrical and Electronics Engineers (IEEE) ,1991
- Technology mapping for electrically programmable gate arraysPublished by Association for Computing Machinery (ACM) ,1991
- Chortle-crf: Fast technology mapping for lookup table-based FPGAsPublished by Association for Computing Machinery (ACM) ,1991
- Architecture of field-programmable gate arrays: the effect of logic block functionality on area efficiencyIEEE Journal of Solid-State Circuits, 1990
- An architecture for electrically configurable gate arraysIEEE Journal of Solid-State Circuits, 1989
- DAGON: technology binding and local optimization by DAG matchingPublished by Association for Computing Machinery (ACM) ,1987