TY - GEN
T1 - Nanomechanical Relay-Based Switchblock for FPGA Interconnect
AU - Marot, Victor G R
AU - Pamunuwa, I D B
N1 - Publisher Copyright:
© 2025 IEEE.
PY - 2025/6/27
Y1 - 2025/6/27
N2 - Reprogrammable non-volatile nanomechanical (NEM) switches have single pole double throw functionality alongside zero standby power that allows the creation of four-way switch nodes made of four NEM switches as opposed to 36 or higher for a SRAM-based implementation. The reduced device count per switch node and accompanying reductions in dynamic power, in turn, allows the creation of island-style FPGA switchblocks with more flexibility than traditional switchblocks by enabling switch matrices fully populated with switch nodes. Simulations for a matrix switchblock within a Stratix IV architecture showed an average reduction of over 40% in the channel width and total critical path wire length, as well as a reduction of nearly 50% in total wire length, over a Wilton topology switchblock. The improved routing efficiency results in more efficient mapping, reduced critical path latency, and an overall reduction in the chip size to map a given function. These improvements, in combination with near total elimination of leakage power, makes a compelling case for NEM-CMOS FPGAs, especially for edge computing applications where energy and footprint constraints are far more stringent.
AB - Reprogrammable non-volatile nanomechanical (NEM) switches have single pole double throw functionality alongside zero standby power that allows the creation of four-way switch nodes made of four NEM switches as opposed to 36 or higher for a SRAM-based implementation. The reduced device count per switch node and accompanying reductions in dynamic power, in turn, allows the creation of island-style FPGA switchblocks with more flexibility than traditional switchblocks by enabling switch matrices fully populated with switch nodes. Simulations for a matrix switchblock within a Stratix IV architecture showed an average reduction of over 40% in the channel width and total critical path wire length, as well as a reduction of nearly 50% in total wire length, over a Wilton topology switchblock. The improved routing efficiency results in more efficient mapping, reduced critical path latency, and an overall reduction in the chip size to map a given function. These improvements, in combination with near total elimination of leakage power, makes a compelling case for NEM-CMOS FPGAs, especially for edge computing applications where energy and footprint constraints are far more stringent.
UR - https://2025.ieee-iscas.org/
U2 - 10.1109/ISCAS56072.2025.11043259
DO - 10.1109/ISCAS56072.2025.11043259
M3 - Conference Contribution (Conference Proceeding)
SN - 9798350356847
T3 - IEEE International Symposium on Circuits and Systems proceedings
BT - 2025 IEEE International Symposium on Circuits and Systems (ISCAS)
PB - Institute of Electrical and Electronics Engineers (IEEE)
T2 - 2025 IEEE International Symposium on Circuits and Systems (ISCAS 2025)
Y2 - 25 May 2025 through 28 May 2025
ER -