An exemplary embodiment of the present disclosure provides a multiplexer/demultiplexer, comprising a plurality of unit cells arranged in a lattice, a first domain, a second domain, a third domain, and a controller. Each of the unit cells can comprise a topological-insulative material, a first piezoelectric patch, and a second piezoelectric patch. A first domain can comprise a first portion of the plurality of unit cells. A second domain can comprise a second portion of the plurality of unit cells. A third domain can comprise a third portion of the plurality of unit cells. The controller can be configured to: apply a negative capacitance to the first piezoelectric patches in the first domain; apply a negative capacitance to the second piezoelectric patches in the second domain; and alternately apply a negative capacitance to the first and second piezoelectric patches, respectively, in the third domain.

Some embodiments of the invention provide a novel method and chip design for a forwarding chip, that decouples input-output (IO) technology requirements from the technology used in a high bandwidth switching ASIC. In some embodiments, a main die including a latest generation switching chip is coupled to a set of IO dies (e.g., SerDes dies). The main die, in some embodiments, uses a latest technology (e.g., 7 nm nodes) while the IO dies, in some embodiments, use a more mature technology (e.g., 16 nm nodes). Some embodiments provide multiple IO dies that each provide connectivity to external components to the high bandwidth switching ASIC (e.g., a core ASIC die). The multiple dies are mounted on a silicon interposer, in some embodiments, using microbumps to make the connections between the dies and the silicon interposer. Additional connections to the pad are made from each die including to general purpose input-output (GPIO) connections. In some embodiments, the main die and the IO dies make all connections through microbumps on the silicon interposer and some microbumps connect to external components using through-silicon vias (TSVs). The microbumps of the main die, in some embodiments, are arranged so that they are mirrored on either side of the main die and rotationally invariant under a 180 degree rotation. IO dies, in some embodiments, are mounted in a first orientation to connect to a first side of the main die and a second rotated (by 180 degrees) orientation to connect to a second opposite side of the main die.

Some embodiments of the invention provide a novel method and chip design for a forwarding chip, that decouples input-output (IO) technology requirements from the technology used in a high bandwidth switching ASIC. In some embodiments, a main die including a latest generation switching chip is coupled to a set of IO dies (e.g., SerDes dies). The main die, in some embodiments, uses a latest technology (e.g., 7 nm nodes) while the IO dies, in some embodiments, use a more mature technology (e.g., 16 nm nodes). Some embodiments provide multiple IO dies that each provide connectivity to external components to the high bandwidth switching ASIC (e.g., a core ASIC die). The multiple dies are mounted on a silicon interposer, in some embodiments, using microbumps to make the connections between the dies and the silicon interposer. Additional connections to the pad are made from each die including to general purpose input-output (GPIO) connections. In some embodiments, the main die and the IO dies make all connections through microbumps on the silicon interposer and some microbumps connect to external components using through-silicon vias (TSVs). The microbumps of the main die, in some embodiments, are arranged so that they are mirrored on either side of the main die and rotationally invariant under a 180 degree rotation. IO dies, in some embodiments, are mounted in a first orientation to connect to a first side of the main die and a second rotated (by 180 degrees) orientation to connect to a second opposite side of the main die.

Some embodiments of the invention provide a novel method and chip design for a forwarding chip, that decouples input-output (IO) technology requirements from the technology used in a high bandwidth switching ASIC. In some embodiments, a main die including a latest generation switching chip is coupled to a set of IO dies (e.g., SerDes dies). The main die, in some embodiments, uses a latest technology (e.g., 7 nm nodes) while the IO dies, in some embodiments, use a more mature technology (e.g., 16 nm nodes). Some embodiments provide multiple IO dies that each provide connectivity to external components to the high bandwidth switching ASIC (e.g., a core ASIC die). The multiple dies are mounted on a silicon interposer, in some embodiments, using microbumps to make the connections between the dies and the silicon interposer. Additional connections to the pad are made from each die including to general purpose input-output (GPIO) connections. In some embodiments, the main die and the IO dies make all connections through microbumps on the silicon interposer and some microbumps connect to external components using through-silicon vias (TSVs). The microbumps of the main die, in some embodiments, are arranged so that they are mirrored on either side of the main die and rotationally invariant under a 180 degree rotation. IO dies, in some embodiments, are mounted in a first orientation to connect to a first side of the main die and a second rotated (by 180 degrees) orientation to connect to a second opposite side of the main die.

Some embodiments of the invention provide a novel method and chip design for a forwarding chip, that decouples input-output (IO) technology requirements from the technology used in a high bandwidth switching ASIC. In some embodiments, a main die including a latest generation switching chip is coupled to a set of IO dies (e.g., SerDes dies). The main die, in some embodiments, uses a latest technology (e.g., 7 nm nodes) while the IO dies, in some embodiments, use a more mature technology (e.g., 16 nm nodes). Some embodiments provide multiple IO dies that each provide connectivity to external components to the high bandwidth switching ASIC (e.g., a core ASIC die). The multiple dies are mounted on a silicon interposer, in some embodiments, using microbumps to make the connections between the dies and the silicon interposer. Additional connections to the pad are made from each die including to general purpose input-output (GPIO) connections. In some embodiments, the main die and the IO dies make all connections through microbumps on the silicon interposer and some microbumps connect to external components using through-silicon vias (TSVs). The microbumps of the main die, in some embodiments, are arranged so that they are mirrored on either side of the main die and rotationally invariant under a 180 degree rotation. IO dies, in some embodiments, are mounted in a first orientation to connect to a first side of the main die and a second rotated (by 180 degrees) orientation to connect to a second opposite side of the main die.

Methods and devices for reading and programming a state of a switch device are presented. The programming of the state of the switch device is performed by providing driving pulses to the switch device. The amplitude and the width of the driving pulses are a function of one or more of a) temperature of the switch device, b) desired state of the switch device, and c) operational time of the switch device. The described devices include a device to store the data demonstrating the functional relation between the amplitude and the width of the driving pulses and the temperature of the switch device. Such device can be a lookup table or an arithmetic logic unit (ALU). The disclosed switch devices can be PCM switches.

A protection switching circuit includes a plurality of ferrite modules arranged in a matrix, wherein the matrix includes a plurality of columns and a plurality of rows. The protection switching circuit further includes a primary control module configured to select a specific ferrite module to be polarized and a redundant control module configured to select a specific ferrite module to be polarized, wherein the redundant control module is used when the primary control module is not used. The protection switching circuit further includes a plurality of first switches, wherein the plurality of first switches couples the plurality of columns of the matrix to a first and second charging circuit. The protection switching circuit further includes a plurality of second switches, wherein the plurality of second switches are organized into pairs, wherein each pair in the plurality of second switches couples a respective row of the matrix to a reference potential.

A protection switching circuit includes a plurality of ferrite modules arranged in a matrix, wherein the matrix includes a plurality of columns and a plurality of rows. The protection switching circuit further includes a primary control module configured to select a specific ferrite module to be polarized and a redundant control module configured to select a specific ferrite module to be polarized, wherein the redundant control module is used when the primary control module is not used. The protection switching circuit further includes a plurality of first switches, wherein the plurality of first switches couples the plurality of columns of the matrix to a first and second charging circuit. The protection switching circuit further includes a plurality of second switches, wherein the plurality of second switches are organized into pairs, wherein each pair in the plurality of second switches couples a respective row of the matrix to a reference potential.

According to an embodiment, a semiconductor switch includes a first insulating film on a semiconductor substrate, a first semiconductor layer on the first insulating film, a semiconductor switch circuit on the first semiconductor layer, and a wiring on the first insulating film. The first insulating film being between the wiring and the substrate. The wiring connects the semiconductor switch circuit and a terminal. A polycrystalline semiconductor layer is between the wiring and the first insulating film.

An exemplary embodiment of the present disclosure provides a multiplexer/demultiplexer, comprising a plurality of unit cells arranged in a lattice, a first domain, a second domain, a third domain, and a controller. Each of the unit cells can comprise a topological-insulative material, a first piezoelectric patch, and a second piezoelectric patch. A first domain can comprise a first portion of the plurality of unit cells. A second domain can comprise a second portion of the plurality of unit cells. A third domain can comprise a third portion of the plurality of unit cells. The controller can be configured to: apply a negative capacitance to the first piezoelectric patches in the first domain; apply a negative capacitance to the second piezoelectric patches in the second domain; and alternately apply a negative capacitance to the first and second piezoelectric patches, respectively, in the third domain.