Abstract: Some embodiments of the invention provide a three-dimensional (3D) circuit that is formed by stacking two or more integrated circuit (IC) dies to at least partially overlap and to share one or more interconnect layers that distribute power, clock and/or data-bus signals. The shared interconnect layers include interconnect segments that carry power, clock and/or data-bus signals. In some embodiments, the shared interconnect layers are higher level interconnect layers (e.g., the top interconnect layer of each IC die). In some embodiments, the stacked IC dies of the 3D circuit include first and second IC dies. The first die includes a first semiconductor substrate and a first set of interconnect layers defined above the first semiconductor substrate. Similarly, the second IC die includes a second semiconductor substrate and a second set of interconnect layers defined above the second semiconductor substrate.

Abstract: Direct-bonded native interconnects and active base dies are provided. In a microelectronic architecture, active dies or chiplets connect to an active base die via their core-level conductors. These native interconnects provide short data paths, which forgo the overhead of standard interfaces. The system saves redistribution routing as the native interconnects couple in place. The base die may contain custom logic, allowing the attached dies to provide stock functions. The architecture can connect diverse interconnect types and chiplets from various process nodes, operating at different voltages. The base die may have state elements for drive. Functional blocks aboard the base die receive native signals from diverse chiplets, and communicate with all attached chiplets. The chiplets may share processing and memory resources of the base die. Routing blockages are minimal, improving signal quality and timing. The system can operate at dual or quad data rates.

Abstract: Aspects of the disclosure relate to forming stacked NAND with multiple memory sections. Forming the stacked NAND with multiple memory sections may include forming a first memory section on a sacrificial substrate. A logic section may be formed on a substrate. The logic section may be bonded to the first memory section. The sacrificial substrate may be removed from the first memory section and a second memory section having a second sacrificial substrate may be formed and bonded to the first memory section.

Abstract: Aspects of the disclosure relate to forming a completed stack of layers. Forming the completed stack of layers may include forming a first stack of layers on a first substrate and forming a second stack of layers on a second substrate. The first stack of layers may be bonded to the second stack of layers. The first or second substrate may be removed. Prior to bonding the first stack of layers and the second stack of layer, one or more holes may be etched in the first stack of layers. After removing the second substrate, one or more holes may be etched in the second stack of layers, wherein each of the one or more holes in the second stack of layers extend into a corresponding hole in the one or more holes in the first stack of layers.

Abstract: Direct-bonded native interconnects and active base dies are provided. In a microelectronic architecture, active dies or chiplets connect to an active base die via their core-level conductors. These native interconnects provide short data paths, which forgo the overhead of standard interfaces. The system saves redistribution routing as the native interconnects couple in place. The base die may contain custom logic, allowing the attached dies to provide stock functions. The architecture can connect diverse interconnect types and chiplets from various process nodes, operating at different voltages. The base die may have state elements for drive. Functional blocks aboard the base die receive native signals from diverse chiplets, and communicate with all attached chiplets. The chiplets may share processing and memory resources of the base die. Routing blockages are minimal, improving signal quality and timing. The system can operate at dual or quad data rates.

Abstract: An integrated circuit and a method for designing an IC wherein the base or host chip is bonded to smaller chiplets via DBI technology. The bonding of chip to chiplet creates an uneven or multi-level surface of the overall chip requiring a releveling for future bonding. The uneven surface is built up with plating of bumps and subsequently releveled with various methods including planarization.

Abstract: Direct-bonded native interconnects and active base dies are provided. In a microelectronic architecture, active dies or chiplets connect to an active base die via their core-level conductors. These native interconnects provide short data paths, which forgo the overhead of standard interfaces. The system saves redistribution routing as the native interconnects couple in place. The base die may contain custom logic, allowing the attached dies to provide stock functions. The architecture can connect diverse interconnect types and chiplets from various process nodes, operating at different voltages. The base die may have state elements for drive. Functional blocks aboard the base die receive native signals from diverse chiplets, and communicate with all attached chiplets. The chiplets may share processing and memory resources of the base die. Routing blockages are minimal, improving signal quality and timing. The system can operate at dual or quad data rates.

Abstract: Configurable smart object systems with methods of making modules and contactors are provided. Example systems implement machine learning based on neural networks that draw low power for use in smart phones, watches, drones, automobiles, and medical devices. Example assemblies can be configured from pluggable, interchangeable modules that have compatible ports for interconnecting and integrating functionally dissimilar sensor systems. An example method includes mounting an element of a configurable machine learning assembly on a substrate, creating at least one fold in the substrate, folding the substrate at the fold into a housing of a module of the configurable machine learning assembly, and adding a molding material to the housing to at least partially fill the module of the configurable machine learning assembly. The example module construction may also form contactors on folded edges of the module for making physical and electrical contact with other modules of the smart object machine learning assembly.

Abstract: Configurable smart object systems with grid or frame-based connectors are provided. Example systems implement machine learning based on neural networks that draw low power for use in smart phones, watches, drones, automobiles, and medical devices, for example. Example assemblies can be configured from pluggable, interchangeable modules that have compatible ports with magnetic electrical contacts for interconnecting and integrating functionally dissimilar sensor systems. An example system has a clip attachable to a substrate for securing a smart object module to the substrate, and a housing of the clip with a geometry for aligning electrical contacts of the smart object module with electrical contacts of the substrate. The clip may have a compliant layer to provide spring, resilience, or pressure to securing the smart object module to the substrate. The clip may also integrate features of a secure digital (SD) port and a universal serial bus (USB) port.

Abstract: Some embodiments provide a three-dimensional (3D) circuit structure that has two or more vertically stacked bonded layers with a machine-trained network on at least one bonded layer. As described above, each bonded layer can be an IC die or an IC wafer in some embodiments with different embodiments encompassing different combinations of wafers and dies for the different bonded layers. The machine-trained network in some embodiments includes several stages of machine-trained processing nodes with routing fabric that supplies the outputs of earlier stage nodes to drive the inputs of later stage nodes. In some embodiments, the machine-trained network is a neural network and the processing nodes are neurons of the neural network. In some embodiments, one or more parameters associated with each processing node (e.g., each neuron) is defined through machine-trained processes that define the values of these parameters in order to allow the machine-trained network (e.g.

Abstract: An integrated circuit and a method for designing an IC where the smallest repeatable block is selected, designed and tested to span across multiple die levels. The block is configured to be timing closed at the block level thereby reducing the overall complexity of the design and avoiding the limiting effects of the constrained EDA tools. The block may subsequently be repeated on multiple die to be stacked in an IC.

Abstract: Some embodiments of the invention provide a three-dimensional (3D) circuit that is formed by vertically stacking two or more integrated circuit (IC) dies to at least partially overlap. In this arrangement, several circuit blocks defined on each die (1) overlap with other circuit blocks defined on one or more other dies, and (2) electrically connect to these other circuit blocks through connections that cross one or more bonding layers that bond one or more pairs of dies. In some embodiments, the overlapping, connected circuit block pairs include pairs of computation blocks and pairs of computation and memory blocks. The connections that cross bonding layers to electrically connect circuit blocks on different dies are referred to below as z-axis wiring or connections. This is because these connections traverse completely or mostly in the z-axis of the 3D circuit, with the x-y axes of the 3D circuit defining the planar surface of the IC die substrate or interconnect layers.

Abstract: A microelectronic package may include a substrate having first and second surfaces each extending in first and second directions, a NAND wafer having a memory storage array, a bitline driver chiplet configured to function as a bitline driver, and a wordline driver chiplet configured to function as a wordline driver. The NAND wafer may be coupled to the first surface of the substrate, and the bitline and wordline driver chiplets may each be mounted to a front surface of the NAND wafer. The NAND wafer may have element contacts electrically connected with conductive structure of the substrate. The bitline and wordline driver chiplets may be elongated along the first and second directions, respectively. Front surfaces of the bitline driver chiplet and the wordline driver chiplet may be arranged in a single common plane and may be entirely contained within an outer periphery of the front surface of the NAND wafer.

Abstract: A microelectronic circuit structure comprises a stack of bonded layers comprising a bottom layer and at least one upper layer. At least one of the upper layers comprises an oxide layer having a back surface and a front surface closer to the bottom layer than the back surface, and a plurality of FD-SOI transistors built on the front surface. At least a first back gate line and a second back gate line extend separate from each other above the back surface for independently providing a first back gate bias to a first group of transistors and a second back gate bias to a second different group of transistors.

Abstract: An integrated circuit and a method for designing an IC wherein the base or host chip is bonded to smaller chiplets via DBI technology. The bonding of chip to chiplet creates an uneven or multi-level surface of the overall chip requiring a releveling for future bonding. The uneven surface is built up with plating of bumps and subsequently releveled with various methods including planarization.

Abstract: A microelectronic unit may include an epitaxial silicon layer having a source and a drain, a buried oxide layer beneath the epitaxial silicon layer, an ohmic contact extending through the buried oxide layer, a dielectric layer beneath the buried oxide layer, and a conductive element extending through the dielectric layer. The source and the drain may be doped portions of the epitaxial silicon layer. The ohmic contact may be coupled to a lower surface of one of the source or the drain. The conductive element may be coupled to a lower surface of the ohmic contact. A portion of the conductive element may be exposed at the second dielectric surface of the dielectric layer. The second dielectric surface may be directly bonded to an external component to form a microelectronic assembly.

Abstract: Some embodiments of the invention provide a three-dimensional (3D) circuit that is formed by stacking two or more integrated circuit (IC) dies to at least partially overlap and to share one or more interconnect layers that distribute power, clock and/or data-bus signals. The shared interconnect layers include interconnect segments that carry power, clock and/or data-bus signals. In some embodiments, the shared interconnect layers are higher level interconnect layers (e.g., the top interconnect layer of each IC die). In some embodiments, the stacked IC dies of the 3D circuit include first and second IC dies. The first die includes a first semiconductor substrate and a first set of interconnect layers defined above the first semiconductor substrate. Similarly, the second IC die includes a second semiconductor substrate and a second set of interconnect layers defined above the second semiconductor substrate.

Abstract: Some embodiments provide a three-dimensional (3D) circuit structure that has two or more vertically stacked bonded layers with a machine-trained network on at least one bonded layer. As described above, each bonded layer can be an IC die or an IC wafer in some embodiments with different embodiments encompassing different combinations of wafers and dies for the different bonded layers. The machine-trained network in some embodiments includes several stages of machine-trained processing nodes with routing fabric that supplies the outputs of earlier stage nodes to drive the inputs of later stage nodes. In some embodiments, the machine-trained network is a neural network and the processing nodes are neurons of the neural network. In some embodiments, one or more parameters associated with each processing node (e.g., each neuron) is defined through machine-trained processes that define the values of these parameters in order to allow the machine-trained network (e.g.

Abstract: Some embodiments of the invention provide a three-dimensional (3D) circuit that is formed by stacking two or more integrated circuit (IC) dies to at least partially overlap and to share one or more interconnect layers that distribute power, clock and/or data-bus signals. The shared interconnect layers include interconnect segments that carry power, clock and/or data-bus signals. In some embodiments, the shared interconnect layers are higher level interconnect layers (e.g., the top interconnect layer of each IC die). In some embodiments, the stacked IC dies of the 3D circuit include first and second IC dies. The first die includes a first semiconductor substrate and a first set of interconnect layers defined above the first semiconductor substrate. Similarly, the second IC die includes a second semiconductor substrate and a second set of interconnect layers defined above the second semiconductor substrate.

Abstract: Aspects of the disclosure relate to forming a completed stack of layers. Forming the completed stack of layers may include forming a first stack of layers on a first substrate and forming a second stack of layers on a second substrate. The first stack of layers may be bonded to the second stack of layers. The first or second substrate may be removed. Prior to bonding the first stack of layers and the second stack of layer, one or more holes may be etched in the first stack of layers. After removing the second substrate, one or more holes may be etched in the second stack of layers, wherein each of the one or more holes in the second stack of layers extend into a corresponding hole in the one or more holes in the first stack of layers.