Power optimization of combinational modules using self-timed precomputation
See where this sits in the topic map →Summary AI-generated
- TL;DR
- Precomputation is an effective power management technique that prevents certain inputs from loading into registers, significantly reducing switching activity within a circuit.
- Problem
- While precomputation effectively saves power, it can potentially introduce a delay penalty in the circuit.
- Method
- This paper presents a self-timed approach for precomputing combinational logic circuits to maximize power savings without relying on a clock signal.
- Results
- By judiciously placing latches within the combinational logic circuit, the approach achieves significant power reductions without increasing the maximum delay.
- Contributions
- Not specified in the abstract.
- Limitations
- Not specified in the abstract.
- Takeaways
- Experiments on arithmetic modules confirm that power dissipation can be greatly reduced with only marginal increases in circuit area and almost zero delay increase.
- Applications
- Not specified in the abstract.
- Topics
- Not specified in the abstract.
- For industry
- Not specified in the abstract.
- Why it matters
- Not specified in the abstract.
Abstract
Precomputation has recently been proposed as a very effective power management technique. Precomputation works by preventing some of the inputs from being loaded into the input registers, thus significantly reducing the switching activity in the circuit. In this paper we present a self-timed approach for the precomputation of combinational logic circuits. This technique allows for maximum power savings without the need of a clock signal. However we may incur in some delay penalty. We describe how to achieve significant power reductions without increasing the maximum delay, by choosing a judicious placement of the latches in the combinational logic circuit. Experimental results are presented for arithmetic modules, confirming that power dissipation can be greatly reduced with marginal increases in circuit area and almost zero delay increase.