Hierarchical Dynamic Power-Gating in FPGAs

Author(s):  
Rehan Ahmed ◽  
Steven J. E. Wilton ◽  
Peter Hallschmid ◽  
Richard Klukas
Keyword(s):  
2017 ◽  
Vol 52 (1) ◽  
pp. 50-63 ◽  
Author(s):  
Minki Cho ◽  
Stephen T. Kim ◽  
Carlos Tokunaga ◽  
Charles Augustine ◽  
Jaydeep P. Kulkarni ◽  
...  

2012 ◽  
Vol 182-183 ◽  
pp. 1440-1445
Author(s):  
Xi Tian ◽  
Fei Qiao ◽  
Zai Wang Dong ◽  
Yu Jun Liu ◽  
Yu Ting Zhao

A novel design methodology for multipliers to reducing both active leakage and dynamic power using dynamic power gating is presented, where sleep transistors are inserted between the real and virtual ground rails of various parts of the multiplier which could be selectively turned on/off. On-chip sleep signals are generated from one input signal of the multiplier which has larger dynamic range. By detecting the magnitude of the input signal, the idle parts of the multiplier are identified and the power gating schemes are dynamically applied even when the multiplier is performing useful computation. Simulations show that the total power dissipation of the proposed multiplier could be reduced up to 39.3% in a typical DSP application.


2016 ◽  
Vol 26 (03) ◽  
pp. 1750041 ◽  
Author(s):  
Abhishek Nag ◽  
Debanjali Nath ◽  
Sambhu Nath Pradhan

Leakage power reduction of an SRAM-based look-up table (LUT) in field-programmable gate array (FPGA) has been achieved in this work by implementing an efficient and dynamic power gating technique. The logic of gating is based on the theory of automatically shutting down the power supply to the inactive blocks of LUT during runtime, contrary to all previous works which involved manual intervention for the implementation of power gating. Two techniques of power gating are introduced in this work, PG1 and PG2. PG1 results in more power savings than PG2, however, PG2 has an advantage of low area overhead. Simulation has been carried out for all possible input combinations of LUT, designed in Cadence Virtuoso tool at 45[Formula: see text]nm technology. The results indicate a leakage power reduction of up to 50% in PG1 technique, with an average area overhead of 14.15%. The power savings in PG2 is up to 38%, with a minimal increase in area of 1.76%. The power bounce noise is also analyzed for the proposed techniques and reported.


2015 ◽  
Vol 742 ◽  
pp. 741-744 ◽  
Author(s):  
G. Amuthavalli ◽  
R. Gunasundari ◽  
A. Nijandan

As scaling down of CMOS transistor’s channel length is done for miniaturization, the design community primarily focuses on the high performance & power-aware design. The power consumption of any circuit solely holds the performance and the life of it. But static power consumption deteriorates them and dominates the dynamic power consumption because of its leakage components. A modified approach of pulse triggering in the Power Gating technique called MPG (Modified Power Gating) is proposed to reduce the static power consumption (leakage power) of digital subsystems. Sub threshold leakage power of MPG Inverter (INV) and 32-bit Digital Comparator (DC) is analyzed and reduced with 35% to 40% leakage savings compared with conventional and existing techniques by simulating it in Cadence GPDK.


Sign in / Sign up

Export Citation Format

Share Document