Optical multi-die interconnect bridge (OMIB)
Through optical multi-die interconnect bridge (OMIB) transmission of data between the processor and memory, the problem of low data transmission efficiency in the prior art is solved and more efficient AI computing hardware performance is achieved.
Patent Information
- Application Number
- CN202380028350.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2023-02-27
- Filing Date
- 2023-03-17
- Publication Date
- 2024-11-05
AI Technical Summary
Data transfer efficiency between existing processors is inefficient, especially in AI computing hardware, where electrical interconnects consume high power, limited pin count, and memory is located in the center of the chip, the processor needs additional transmission to access memory from the edge of the chip. Distance leads to inefficiency and is difficult to meet the needs of AI applications.
Optical multi-die interconnection bridge (OMIB) is adopted to transmit data between the processor and the memory through a photon receiver, and a temperature-stable optical modulator is used to realize bidirectional transmission of optical signals, reduce electrical connections, and improve data transmission efficiency.
It improves the speed and bandwidth of data transmission, reduces power consumption, reduces latency, and can effectively support efficient data transmission of AI computing hardware, meeting the needs of AI applications.
Smart Images

Figure CN118901031A_ABST
Abstract
Description
[0001] Cross-references
[0002] This application claims priority to U.S. Provisional Patent Application Serial No. 63 / 344,585, filed on February 27, 2023, entitled “Optical, Multi-Die Interconnect Bridge (OMIB)”, U.S. Provisional Patent Application Serial No. 63 / 420,330, filed on October 28, 2022, entitled “Thermally Stable Optical Modulation Elements Coupled to Electronic Elements”, and U.S. Patent Application Serial No. 17 / 903,455, filed on September 6, 2022, entitled “Photonic Memory Fabric for System Memory Interconnection”; which patent application claims priority to U.S. Provisional Patent Application Serial No. 63 / 321,453, filed on March 18, 2022, entitled “Photonic Memory Fabric for System Memory Interconnectio”. These priority applications are incorporated herein by reference as if fully set forth in this specification.
[0003] Each publication, patent, and / or patent application mentioned in this specification is herein incorporated by reference in its entirety to the same extent as if each individual publication and / or patent application was specifically and individually indicated to be incorporated by reference. Technical Field
[0004] Embodiments of the present disclosure relate generally to apparatus and methods for use in interconnecting semiconductor dies, and more particularly to apparatus and methods for optical interconnection using integrated circuits and / or external chiplets. Background Art
[0005] The subject matter discussed in this section should not be considered to be prior art simply because it is mentioned in this section. Similarly, issues mentioned in this section or associated with the subject matter provided as background should not be considered to have been previously acknowledged in the prior art. The subject matter in this section merely represents different approaches, which themselves may also correspond to implementations of the claimed technology.
[0006] Integrated circuits (ICs) with processors, especially those used to perform artificial intelligence and machine learning functions, move large amounts of data between one or more processor ICs and one or more memory ICs. Chips can facilitate the interconnection of processor dies, memory dies, and other circuits to increase bandwidth and reduce latency and power consumption in the process. Summary of the invention
[0007] Processing AI workloads typically uses dedicated hardware. Typical hardware bridges two chips with an electrical interconnect. Electrical interconnects consume high power, have pin count limitations, and can only bring data to the edge of the chip. When the memory is located in the central area of the chip, every time the processor sends a request from outside the chip to the edge of the chip to access the memory, the signal needs to travel an additional distance. This is very inefficient and makes it difficult for AI computing hardware to keep up with the demands required by AI applications.
[0008] In a first aspect, an embodiment provides a package comprising a bridging element, a first unidirectional photon path, and a second unidirectional photon path. The bridging element comprises a first interconnection region and a second interconnection region, both of which are configured to be electrically connected to one or more bare chips. The unidirectional photon path is coupled in opposite directions between the two interconnection regions. The third unidirectional photon path and the fourth unidirectional photon path can be bidirectionally coupled between the first interconnection region and the optical interface (OI). The first part of the photon transceiver can be coupled to the first interconnection region to send a first optical signal through the first unidirectional photon path and receive a second optical signal through the second unidirectional photon path. It can also be configured to send a third optical signal through the third unidirectional photon path and receive a fourth optical signal through the fourth unidirectional photon path. The first part of the photon transceiver can be located in a photon integrated circuit (PIC), and the second part of the photon transceiver can be located in an electronic integrated circuit (EIC). The first part and the second part can be coupled by an electrical interconnection having a length of less than 2 mm. The modulator may be an electro-absorption modulator fabricated with germanium, silicon, a germanium alloy, a silicon alloy, an indium phosphide (InP) based III-V material, or a gallium arsenide (GaAs) based III-V material. The EIC may include two or more computing elements surrounding a central region. The second portion is proximate to one of the computing elements.
[0009] A further understanding of the nature and advantages of the specific embodiments disclosed herein may be obtained by reference to the remaining portions of the specification and the appended drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0010] The present invention will be described with reference to the accompanying drawings, in which:
[0011] Figure 1are top views of wafers and dies that may include or be included in one or more microelectronic packages having an OMIB, according to various embodiments.
[0012] Figure 2 An example OMIB with two bridged dies is shown.
[0013] Figure 3 An example package is shown in which two dies may be coupled by electrical interconnects.
[0014] Figure 4 An example package is shown in which two dies may be coupled by electrical interconnects.
[0015] Figure 5A-5E An example package for using an optical engine to provide optical signals for use in a photonic network is shown.
[0016] Figure 6 A perspective view of an example OMIB bridging two dies and coupled to a light engine is shown.
[0017] Fig. 7A A perspective view of an example OMIB bridging two dies and receiving unmodulated light from a core light engine (CLE) is shown.
[0018] Figure 7B An example package in which a CLE may be used is shown.
[0019] Fig. 8A is a schematic diagram of an example OMIB with a photonic link capable of connecting two dies within a package, according to some embodiments.
[0020] Figure 8B An embodiment is shown having two OMIBs coupled between a first die and a second die.
[0021] Fig. 9 is a schematic diagram of an example system for a link using WDM for communications from a die to an external device in accordance with some embodiments.
[0022] Fig.10 According to some embodiments, Fig. 9 An example backchannel for a WDM system in FIG.
[0023] Fig.11 An example unidirectional logical channel comprising multiple photonic links is shown.
[0024] Fig.12 Shows Fig.11 Transmission of an example message of eight words in a unidirectional logical channel.
[0025] Fig.13 An example AMS block is shown in more detail.
[0026] Fig.14A An arrangement of three OMIBs bridging a portion of two dies is shown.
[0027] Fig. 14B An example arrangement of five dies bridged by OMIB is shown.
[0028] Fig. 14C An example arrangement of dies bridged by OMIB and bridges is shown.
[0029] Fig.14D An example offset checkerboard of OMIB bridging dies in two dimensions is shown.
[0030] Fig.15 is a flow chart illustrating an example method of making an OMIB according to various embodiments.
[0031] Fig.16 Components of an example system are shown in which one or more OMIBs may be used.
[0032] Figures 17A-17D An example of an EIC / OMIB combination employing temperature compensation to broaden the temperature range of the modulator in the OMIB is shown.
[0033] Fig.18 An example method 1800 of manufacturing a system is shown.
[0034] In the drawings, similar reference numerals represent functionally similar elements. The systems and methods shown in the drawings and described in the following detailed description can be arranged and designed in a variety of different embodiments. The drawings and detailed description are not intended to limit the scope of the claims. Instead, they merely represent examples of different embodiments of the present invention. DETAILED DESCRIPTION
[0035] Processing AI workloads typically uses dedicated hardware. Typical hardware bridges two chips with an electrical interconnect. Electrical interconnects consume high power, have pin count limitations, and can only bring data to the edge of the chip. When the memory is located in the central area of the chip, every time the processor sends a request from outside the chip to the edge of the chip to access the memory, the signal needs to travel an additional distance. This is very inefficient and makes it difficult for AI computing hardware to keep up with the demands required by AI applications.
[0036] This document discloses an optical multi-die interconnect bridge element (OMIB). The OMIB can be used as a bridge between semiconductor dies, such as electronic integrated circuits (EICs). The bridge can include the OMIB alone, or in combination with a substrate coupled to the OMIB or having the OMIB embedded therein. Using the OMIB for multi-die processing systems solves many problems associated with processing AI workloads, including latency, power, and bandwidth. The photonic receiver can include two parts, such as a first part in the OMIB and a second part in the EIC, the first part including a modulator and / or a photodetector, and the second part including an AMS block, as described below.
[0037] In various arrangements, the OMIB can send or receive photon signals to transmit data. Memory (such as a cache) can be located in the central area of the EIC die within two millimeters (2 mm) of the AMS block, so that the photon transceiver in the OMIB is close to the edge of the memory, which is directly above or below the portion of the die where the memory edge is located. The central area can intersect the center of the EIC die. Computing elements (such as central processing units (CPUs), graphics processing units (GPUs), tensor processing units (TPUs)) can also be advantageously arranged in the central area of the die within 2 millimeters (2 mm) from the location of the photon transceiver, or spatially associated with the memory. Due to the heat generated by the connected chip, the photon IC has avoided reaching the center of the die of the connected chip. The photon chip can have a limited temperature range within which the modulator operates within the specification. One reason why the OMIB in the technology of the present disclosure can reach the center of the die is that it uses a temperature-stabilized modulator as described in U.S. Provisional Patent Application Serial No. 63 / 420,330 entitled "Thermally Stable Optical Modulation Elements Coupled to Electronic Elements".
[0038] As a result, OMIB is faster and uses less power than traditional systems. By optically transferring data to the point of compute rather than to the edge of the die, latency is improved. This allows the die to save electrical pipeline stages and also utilize fewer electrical connections to transfer data from the edge of the die to the interior where the memory is located. The electrical movement of data from the edge of the die to the interior requires a slower and more power-hungry data path. If the exemplary system is used to train an AI model, the benefits of moving data optically to the point of compute are repeated over and over, resulting in significant savings and / or the feasibility of implementing these types of complex AI systems.
[0039] the term
[0040] The present description may use perspective-based descriptions (such as top / bottom, inside / outside, up / down, etc.) Such descriptions are only used to facilitate discussion and are not intended to limit the application of the embodiments described herein to any particular orientation.
[0041] As used herein, the phrase "one of" should be interpreted as one of the listed items. For example, the phrase "one of A, B, and C" should be interpreted as meaning only A, only B, or only C.
[0042] As used herein, the phrases "at least one" and "one or more" should be interpreted as one or more items. For example, the phrase "at least one of A, B, and C" or the phrase "at least one of A, B, or C" should be interpreted to mean any combination of A, B, and / or C.
[0043] Unless otherwise specified, the use of ordinal adjectives "first", "second", "third", etc. to describe objects merely refers to different instances or categories of the objects and does not imply any rank or order.
[0044] The term "coupled" is used in an operational sense and is not limited to direct or indirect coupling. "Coupled to" is generally used in the sense of direct coupling, while "coupled with..." is generally used in the sense of direct or indirect coupling. "Coupling" in an electronic system can refer to a configuration that allows the flow of information, signals, data, or physical quantities (such as electrons) between two elements that are coupled to or coupled to each other. In some cases, the flow can be unidirectional, in other cases, the flow can be bidirectional or multidirectional. Coupling can be galvanic (meaning in this context that there is a direct electrical connection), capacitive, inductive, electromagnetic, optical, or by any other process that is physically permitted.
[0045] The term "connected" is used to indicate a direct connection (such as an electrical, optical, electromagnetic or mechanical connection) between the things being connected without any intervening things or devices.
[0046] The term "configured to" perform one or more tasks is a broad description of a structure, generally referring to "having circuitry that performs one or more tasks during operation". Thus, the item being described may be configured to perform a task even when the unit / circuit / component is not currently turned on or activated. In general, the circuits forming the structure corresponding to "configured to" may include hardware circuits, and may also be controlled by switches, fuses, bonding wires, metal masks, firmware, and / or software. Similarly, for ease of description, various items may be described as performing one or more tasks. Such descriptions should be interpreted as including the phrase "configured to".
[0047] As used herein, the term "based on" is used to describe one or more factors that influence a determination. The term does not exclude the possibility that other factors may influence the determination. That is, a determination may be based only on specified factors or on specified factors and other unspecified factors. Consider the phrase "determining A based on B." The phrase specifies that B is a factor used to determine A or a factor that influences the determination of A. The phrase does not exclude that the determination of A may also be based on some other factor (such as C). The phrase is also intended to cover embodiments in which A is determined based only on B. Therefore, the phrase "based on" is synonymous with the phrase "based at least in part on."
[0048] A "processor" includes any suitable hardware system, mechanism or component that processes data, signals or other information. A processor may include a system having a general-purpose central processing unit, multiple processing units, a dedicated circuit for implementing a function, or other system. Examples of processing systems may include servers, clients, end-user devices, routers, switches, network storage, etc. A "computer" may be any processor that communicates with a memory. The memory may be any suitable processor-readable storage medium (such as a random access memory (RAM), a read-only memory (ROM), a magnetic disk or an optical disk) or other tangible medium suitable for storing instructions executed by a processor.
[0049] The terms "substantially," "close," "approximately," "near," and "about" mean within plus or minus 10% of a specified value, unless expressly specified otherwise.
[0050] As used herein, the following terms or acronyms are at least partially defined as follows:
[0051] AI - Artificial Intelligence
[0052] AMS - Analog / Mixed Signal
[0053] BGA - Ball Grid Array
[0054] A channel is one or more lanes that can be linked together.
[0055] Chiplet - An integrated circuit with simple or specialized functions used in conjunction with other ICs or chips in a multi-chip module.
[0056] CLE-Core Light Engine
[0057] CPU - Central Processing Unit
[0058] CW - Continuous Wave
[0059] DDR Memory - Double Data Rate Memory
[0060] DFB-Distributed Fiber Bragg
[0061] DIMM - Dual In-line Memory Module
[0062] DRV-Driver
[0063] DSP - Digital Signal Processor
[0064] EAM - Electro Absorption Modulator
[0065] EIC-Electronic Integrated Circuit
[0066] EW-Evanescent Wave
[0067] FAU - Fiber Array Unit
[0068] FBG - Fiber Bragg Grating
[0069] FPGA - Field Programmable Gate Array
[0070] GC - Grating Coupler - Coupling between optical fiber and on-chip photonic waveguide
[0071] GPU - Graphics Processing Unit
[0072] GRIN Lens - Gradient Index Lens
[0073] HBM - High Bandwidth Memory
[0074] IC - Integrated Circuit - A monolithic integrated circuit, i.e. a single semiconductor die that can be delivered as a bare die or packaged circuit. For the purposes of this article, the term integrated circuit also includes packaged circuits containing multiple semiconductor dies, stacked dies, or multi-die substrates. Such structures are now common in the industry, produced by the same supply chain, and are usually indistinguishable from monolithic circuits to the average user.
[0075] A channel includes a serializer, a link, and a deserializer.
[0076] LED - Light Emitting Diode
[0077] LDSU - Load / Store Unit
[0078] A link in the context of this patent document is a combination of a modulator, a photon path (in an optical transmission medium) and a photodetector.
[0079] LGA - Land Grid Array
[0080] ML - Machine Learning
[0081] MLA-Microlens Array
[0082] MOD - Modulator
[0083] MZI - Mach-Zehnder Interferometer
[0084] OI - Optical Interface - An interface that uses any means to optically interface between an optical fiber and a photonic IC.
[0085] OMIB - Optical Multi-Die Interconnect Bridge
[0086] PCB - Printed Circuit Board
[0087] PCIe - PCI Express - high-speed serial computer expansion bus
[0088] PCM - Phase Change Memory
[0089] PD - Photodetector, such as photodiode
[0090] PGA - Pin Grid Array
[0091] PIC - Photonic Integrated Circuit
[0092] QR - Quick Response
[0093] A processing device, processor, compute device, or compute element may refer to any device or portion of a device that processes electronic data from registers and / or memory to transform the electronic data into other electronic data that may be stored in registers and / or memory.
[0094] RAM - Random Access Memory, some types are Static RAM (SRAM), Dynamic RAM (DRAM), Magnetic RAM (MRAM), Resistive RAM (RRAM), Conductive Bridge RAM (CBRAM) etc.
[0095] RDL - Redistribution Layer - An additional metal layer on an integrated circuit that allows pads to be used elsewhere on the die.
[0096] SCM - Storage Class Memory
[0097] SLD - Super Luminescent Diode
[0098] SSD - Solid State Drive
[0099] SOA - Semiconductor Optical Amplifier
[0100] TPU - Tensor Processing Unit
[0101] UCIe - Universal Chip Interconnect Express - open specification for die-to-die interconnect and serial bus between chiplets
[0102] VCSEL - Vertical Cavity Surface Emitting Laser
[0103] A waveguide is an implementation of a unidirectional path for photons in an optical transmission medium.
[0104] WDM - Wavelength Division Multiplexing
[0105] Implementation
[0106] Embodiments herein are directed to addressing power, latency or pin count issues by providing various OMIB configurations and various packages including at least one OMIB. Figure 1 1 is a top view of a wafer 100 and a die 102 that include or are included in one or more microelectronic packages with an OMIB according to various embodiments. The wafer 100 may be composed of a semiconductor material and may include one or more die 102 having an IC structure formed on a surface of the wafer 100. Each die 102 may be a unit of a semiconductor product including a suitable IC. After the manufacture of the semiconductor product is completed, the wafer 100 may undergo a singulation process in which the die are separated from each other to provide discrete "chips" of the semiconductor product. The die 102 and at least one other die may be included in a microelectronic package with an OMIB.
[0107] A second wafer (not shown) may be used in a second process to form an OMIB in a manner similar to the manufacture of die 102. The OMIB is a photonic integrated circuit (PIC) and has optical components manufactured therein, in contrast to die 102, which is an electronic integrated circuit (EIC) and typically has electronic components manufactured therein. The OMIB may be embedded in a packaging substrate. The packaging substrate may be considered to be a cored or coreless substrate. The packaging substrate may include one or more layers of dielectric material, organic or inorganic. The packaging substrate may also include one or more conductive elements (such as vias, pads, traces, microstrips, striplines, etc.). The conductive elements may be inside or on the surface of the packaging substrate. Typically, the conductive elements may allow signals to be routed through the packaging substrate or between elements coupled to the packaging substrate. In some embodiments, the packaging substrate may be, for example, a printed circuit board (PCB), an interposer, a motherboard, or some other type of substrate.
[0108] Figure 2An example OMIB with two bridged dies is shown. The OMIB 200 formed in the second stage of the manufacturing process can share dimensions with each die 102 (horizontal or vertical) formed in the first stage of the manufacturing process. The die can be as large as possible within the size constraints of the reticle used to manufacture it. Figure 2 In the embodiment, die 102 has a width corresponding to the maximum reticle width MRW1 of the EIC reticle and a height corresponding to the maximum reticle height MRH1 of the EIC reticle, and OMIB 200 has a width corresponding to the maximum reticle width MRW2 of the PIC reticle. This may be for convenience and cost saving during manufacturing, and is not a limitation of various embodiments.
[0109] In some embodiments, the wafer 100 or die 102 may include a memory device, a computing device, or both (examples include, but are not limited to, a random access memory (RAM) device (such as a static RAM (SRAM) device, a magnetic RAM (MRAM) device, a resistive RAM (RRAM) device, or a conductive bridge RAM (CBRAM) device), a logic device (e.g., an AND gate, an OR gate, a NAND gate, a NOR gate, or an exclusive OR (EXOR) gate), NAND flash memory, solid-state drive (SSD) CMOS memory, thin film transistor-based memory, phase change memory (PCM), storage class memory (SCM), magnetoresistive memory (MRAM), Resistive RAM, DRAM, high bandwidth memory (HBM), DDR-based DRAM, DIMM memory, CPU, GPU, MPU, tensor engine, load / store unit (LDSU), neural compute engine, dot product and / or convolution engine, field programmable gate array (FPGA), AI accelerator, or any other suitable circuit element. Multiple of these devices can be combined on a single die 102. For example, die 102 can include a memory including multiple memory arrays, one or more processors, other logic circuits, communication circuits, and power management functions, and execute instructions stored in the memory array, or otherwise interact with the memory array using a processor on die 102.
[0110] Figure 3 An example package 300 is shown in which a die 310 and a die 320 may be coupled by electrical interconnects. The package 300 includes an OMIB 330 that acts as a bridge between the die 310 and the die 320. An optional substrate 340 has been coupled to the OMIB 330 such that the OMIB 330 is embedded in the substrate 340. Figure 3340, but the OMIB 330 may not be flush with the surface of the substrate 340, but may at least partially protrude from the substrate 340 or be embedded in the inner layer of the substrate 340. The various elements depicted may be on the surface of the OMIB 330 or the die 310 and / or the die 320, or at least partially protrude from the OMIB 330 or the die 310 and / or the die 320. In addition, in other embodiments, the specific number of elements, the number of interconnects, the number of dies and other devices may be different. In some embodiments, the package may include an adhesive, a sealant, an underfill, a mold, an overmold, or some other structural material or dielectric material. In some embodiments, certain elements may be located at different positions within the signal path relative to other elements. An optical interface (OI) 350 is included on the surface of the OMIB 330. The optical interface 350 may be coupled to the light engine 370 to receive unmodulated light through the optical fiber 360, and / or coupled to the external device optical interface to exchange modulated and / or unmodulated light in either direction. In addition to the optical interface 350, there may be one or more electrical interfaces on the surface of the OMIB 330 to interface with external electronic devices.
[0111] Various embodiments may use different types of light engines. The light engines may be external or integrated into the OMIB. Example embodiments may use the following light engines to transfer optical signals to and from the package that includes the OMIB 330: Laser diodes - These are highly coherent light sources that produce narrow beams of light and are widely used in photonic chips for data communications and sensing applications; Light Emitting Diodes (LEDs) - A diode that emits light when an electric current is passed through it (they are a popular choice for photonic chips due to their low cost, compact structure, and long life); Super Luminescent Diodes (SLDs) - SLDs are similar to LEDs, but they emit a wider broadband spectrum (SLDs are used in applications such as optical amplification, wavelength division multiplexing, and fiber optic sensing); Vertical Cavity Surface Emitting Lasers (VCSELs) - These lasers emit light perpendicular to the surface of the chip and are well suited for photonics applications (VCSELs are widely used in data communications and sensing applications such as 3D sensing, LiDAR, and data center interconnects); Silicon Photonics - Devices that use the light-carrying properties of silicon to create light sources on silicon chips (Silicon Photonics has the potential to revolutionize photonics by reducing the size, cost, and power consumption of photonic components). These are some of the most commonly used light sources in packages that include the OMIB 330. The choice of light engine depends on the specific requirements of the implementation (eg, wavelength, power, and modulation speed).
[0112] The light engine can be local to the OMIB 330 or connected to the OMIB 330 via optical fiber 360. When integrated onto the PIC, a distributed fiber Bragg (DFB) laser or quantum dot laser can be attached during processing or integrated into the local technology where possible. When remote, any packaged continuous wave (CW) laser with a power and spectrum suitable for the modulation technology can be used. In one embodiment, the light source is a set of DFB lasers attached to a silicon interposer and connected to the PIC via optical fiber.
[0113] Optical interface 350 is used to terminate optical fibers at the edge or top of OMIB 330 for optical input and / or output to an external process or device that is communicatively coupled to package 300 through the optical fiber. The choice of OI depends on the specific requirements of the implementation (e.g., wavelength of light, coupling efficiency, and cost). Optical interface 350 and any other optical interface in or on OMIB 330 can include any device for optical interface between optical fiber and photonic IC, such as: edge coupler, grating coupler (CG), graded refractive index (GRIN) lens coupler, fiber Bragg grating (FBG) coupler, microlens array (MLA) coupler, evanescent wave (EW) coupler, adiabatic coupler, wavelength division multiplexing (WDM) coupler, prism coupler, butt coupler, end-fire coupler, and V-groove coupler.
[0114] In one embodiment, the optical interface 350 includes a fiber array unit (FAU) for optically connecting the OMIB to a light source and / or an optical I / O unit. A FAU is a device used in optical communication systems that combines or separates optical signals from multiple optical fibers into a single optical signal or multiple optical signals. FAUs can be used in a variety of applications (such as wavelength division multiplexing (WDM), parallel optical interconnects, and optical sensing). Two main types of fiber array units can be used: linear and circular. Linear FAUs combine or separate optical signals along a straight line, while circular fiber array units combine or separate optical signals in a circular configuration. These two types of fiber array units are typically made of precision molded optical plastic or ceramic materials and can have from a few to hundreds of optical fibers arranged in a specific pattern. The choice of FAU depends on the specific requirements of the application (such as the number of optical fibers, the arrangement of the optical fibers, the wavelength of the light used, and the desired coupling efficiency). Each die may have an associated analog / mixed signal (AMS) block associated with a portion of the OMIB to which it is coupled and one or more transistors or supporting circuits to route electrical signals to the transistor or some other IC component.
[0115] OMIB 330 may also include one or more electrical interfaces (EIs). OIs and EIs may be attached to the surface of the OMIB, or between the bridged die 102, or offset in an area in the x or y direction where the die 102 does not extend. EIs have functions similar to OIs, except that they give OMIBs the ability to transmit data to and from the die by electrical connections rather than optical connections, and may also connect external I / O blocks to the OMIB, for example, by wires from the EI to external I / O blocks. Different types of electrical connections are possible, including wires, RDLs, etc. Electrical connections typically use known standards or bus protocols to electrically send and receive data between the EI and any die bridged by the OMIB or the EI and external I / O blocks by electrical interconnections. The electrical path between the EI and the die may be routed through the OMIB, through the substrate, or through both (or in the case of RDL, through a layer of the OMIB and / or substrate).
[0116] Figure 4 An example package 400 is shown in which a die 410 and a die 420 may be coupled via electrical interconnects. The package 400 is similar to Figure 3 package 300, with similar component numbers, in addition, Figure 4 Shown are a first photonic transceiver 432 and a second photonic transceiver 434. A portion of the first transceiver 432 is in die 410 and a portion of the second photonic transceiver 434 is in die 420. Another portion of the first transceiver 432 is in OMIB 430 and another portion of the second transceiver 434 is in OMIB 430. Figure 4 Bidirectional photon paths 481, 482, and 483 are shown. Bidirectional photon path 483 connects first photon transceiver 432 and second photon transceiver 434. Bidirectional photon path 483 may include two or more unidirectional photon paths in opposite directions to create a bidirectional path between the two transceivers, thereby enabling optical communication and / or data movement through photon channels in both directions between the two dies 410 and 420. Bidirectional photon path 481 links photon transceiver 432 to optical interface 450, and bidirectional photon path 482 links photon transceiver 434 to optical interface 450. Optical interface 450 is coupled to optical engine 470 via optical fiber 460. Bidirectional photon paths 481-483 may form a photon network inside OMIB 430. Optical interface 450 may include any device for interfacing between optical fiber and photon IC. Substrate 440 is optional.
[0117] The photon path may be implemented in an optical transmission medium. The optical transmission medium may include a waveguide on a PIC, an optical fiber, or other optical transmission medium (such as free space optics or glass etched waveguides), or some combination of the foregoing. Examples of optical modulators include, but are not limited to, electro-absorption modulators (EAMs), micro-ring resonators, or any suitable optical component with sufficient thermal stability.
[0118] Figure 5A -E shows an example package for using an optical engine to provide optical signals for use in a photonic network. The optical engine may include a laser diode, optics, and / or control electronics. The laser diode is the source of the laser light, and the optics are used to shape and control the output of the laser. The control electronics (not shown) provide the necessary power and modulation to the laser diode to generate the desired output. Figure 5A-5E In the embodiment of the present invention, the optical engine provides the optical signal to the optical splitter, modulator (MOD) adjacent to the AMS block in the first die through the optical interface (such as FAU and GC) and finally to the photodetector (PD) adjacent to the AMS block of the second die. Fig. 7A In the case described), a FAU is not required and the light can be directed directly into the GC. Figure 5A-5E The photonic network in OMIB realizes multiple light paths. Figure 5A-5E The GC and FAU are shown together forming an optical interface, but embodiments may use any means to interface between an optical fiber and a photonic IC.
[0119] Figure 5A An example 500 of connections within an OMIB is shown, showing the optical path starting at light engine 570 . Figure 5A The first die 510 is divided into a general part 510B and an AMS part 510A, wherein the general part 510B may include various processing, storage and communication functions, and the AMS part 510A includes analog / mixed signal circuits for interfacing with the OMIB 530. It also includes an optional substrate 540 and an optical engine 570. The AMS part 510A of the first die 510 may include a driver (DRV1) and a transimpedance amplifier (TIA1). The AMS part 520A of the second die 520 may include a driver (DRV2) and a transimpedance amplifier (TIA2).
[0120] The optical engine 570 transmits light to the optical splitter SP within the OMIB 530 through the fiber array unit FAU and the grating coupler GC (or any other OI configured to receive light on the OI input and pass the received light on the OI output). The optical splitter SP distributes light to the modulator MOD1 and the modulator MOD2 through two different photon paths 531 and 532. In some embodiments, the optical splitter or optical splitter tree distributes light on more than two different photon paths to feed additional modulators. The photon paths can be implemented with any suitable optical transmission medium and can include a mix of waveguides and optical fibers.
[0121] The modulator MOD1 modulates the light it receives from the optical splitter SP with information from the driver DRV1, and transmits the modulated light to the photodetector PD2 through the photon path 533. The photodetector PD2 converts the received light into an electrical signal of the second die 520. The modulator MOD2 modulates the light it receives from the optical splitter SP with information from the driver DRV2, and transmits the modulated light to the photodetector PD1 through the photon path 534. The photodetector PD1 converts the received light into an electrical signal of the first die 510. Together with the serializer (not shown) in the first die 510, the driver DRV1, the transimpedance amplifier TIA2, and the deserializer (not shown) in the second die 520, the modulator MOD1, the photon path 533, and the photodetector PD1 form a data channel from the first die 510 to the second die 520.
[0122] OMIB 530, photonic IC includes a first interconnect region, a second interconnect region, and an offset region. The first interconnect region includes a pad pattern located above MOD1 and PD1, which pad pattern matches the pad pattern on the first die 510 located below DRV1 and TIA1, or is otherwise configured to form electrical interconnections between the respective components. The second interconnect region includes a pad pattern located above PD2 and MOD2, which pad pattern matches the pad pattern on the second die 520 located below TIA2 and DRV2, or is configured to form electrical interconnections between the respective components. Figure 6The offset region further shown in the figure includes an optical interface (such as a GC / FAU). Two or more pads of the pad pattern on the first die 510 are physically and electrically coupled to two or more pads of the pad pattern in the first interconnect region, and two or more pads on the pad pattern on the second integrated circuit are physically and electrically coupled to two or more pads of the pad pattern in the second interconnect region. One or more of the two or more pads on the first die 510 can be located at a position more than 100 microns (100 μm) from the edge of the first die 510. One or more of the two or more pads on the second die 520 can be located at a position more than 100 microns (100 μm) from the edge of the second die 520. The distance between the (lower) surface of the first die 510 and the (top) surface of the OMIB 530 is less than 2 mm, and in many cases less than 50 microns. The distance between the (lower) surface of the second die 520 and the (upper) surface of the OMIB 530 is less than 2 mm, and in many cases less than 50 microns.
[0123] Figure 5B The same example is shown, only the signal path is shown. The path of the unmodulated light is omitted. Figure 5B A signal path is shown with a bold line as follows. The interface I / F1 in the first die 510 transmits a digital signal carrying information from the common part 510B of the first die 510 to the driver DRV1. The driver DRV1 converts the digital signal into an analog electrical signal suitable for driving the photon modulator, and transmits the analog electrical signal to the modulator MOD1 in the OMIB 530. The modulator MOD1 sends the modulated light with the information to be transmitted to the photodetector PD2 through the photon path 533. The photodetector PD2 converts the modulated light into a current signal and transmits the current signal to the transimpedance amplifier TIA2, which amplifies the signal and converts it into a voltage signal suitable for the digital interface I / F2. The interface I / F2 converts the voltage into a digital signal suitable for processing in the common part 520B of the second die 520. Similarly, information to be transmitted from the common portion 520B of the second die 520 to the common portion 510B of the first die 510 propagates through I / F2, DRV2, MOD2, photon path 534, photodetector PD1, transimpedance amplifier TIA1, and interface I / F1.
[0124] Photon paths 533 and 534 include waveguides or other suitable optical transmission media to carry optical signals from the modulator to the photodetector. The modulator in OMIB 530 is coupled to a driver in the AMS portion of the first die or the second die through a copper pillar or other suitable electrical interconnect. The photodetector in OMIB 530 is coupled to a transimpedance amplifier in the AMS portion of the first die or the second die through a copper pillar or other suitable electrical interconnect. The photon paths in OMIB 530 can be unidirectional so that a pair of photon paths in opposite directions can be included in a single bidirectional information channel.
[0125] The electrical interconnection is shown as forming a coupling (or abutting coupling) between the elements in the AMS component and the corresponding elements in the OMIB 530. In one embodiment, the interconnection is a copper column of no more than 2 mm. In other embodiments, the electrical interconnection can be a solder bump formed by a material such as tin, silver or copper. If the solder bump is used for interconnection, the solder bump can be a flip chip bump. In other embodiments, the interconnection can be an element of a ball grid array (BGA), a pin of a pin grid array (PGA), an element of a planar grid array (LGA) or some other types of interconnection. Generally, the interconnection can physically and electrically couple the AMS block to the OMIB 530. For example, one or more interconnections can be physically coupled to the pads of the die and the substrate 540 and / or the pads of the OMIB 530, and allow electrical signals to be transmitted between them. The interconnection 525 may not have a uniform size, shape or spacing. A finer interconnection spacing may be required to allow a denser communication path between the elements coupled to the OMIB. In an embodiment, the size, shape, spacing or type of one or more interconnects may be different from that depicted in the figures, or different from other interconnects. The specific type, size, shape or spacing of the interconnects may be based on one or more factors (such as usage, materials used, design considerations, and manufacturing considerations).
[0126] Figure 5C An example of an inter-OMIB connection is shown. The path of unmodulated light is omitted. The first die 510 and the second die 520 can use the OMIB 580 and the optical fiber 561 to perform optical communication with the external device optical interface 571. For example, the first die 510 can transmit to the external device optical interface 571 through I / F1, DRV1, MOD1, photon path 581, optional multiplexer MUX (in the case of desired wavelength division multiplexing), GC, FAU and optical fiber 561. Conversely, the first die 510 can receive from the external device optical interface 571 through optical fiber 561, FAU, GC, photon path 583, optional demultiplexer DEMUX (in the case of desired wavelength division multiplexing), PD1, TIA1 and I / F1.
[0127] Figure 5D An example of a stacked die arrangement is shown. The AMS circuits (DRV1 and TIA1) are included in the first AMS die 511A, and the general / digital circuits are included in the first general die 511B. The first AMS die 511A is stacked on top of the OMIB 580 via electrical interconnect 526A, and the first general die 511B is stacked on top of the first AMS die 511A via electrical interconnect 526B. The OMIB 580 interfaces with the optical engine 570 and / or external device optical interface 571 via optical fiber 560, and may also interface with the second die 520.
[0128] Figure 5E An arrangement is shown in which the first universal die 512B and the first AMS die 512A are stacked and placed side by side on the OMIB 590, respectively. An electrical interconnect, such as a bus, enables signals to travel to and from the first universal die 512B and the first AMS die 512A (e.g., using UCIe, PCIe, or any suitable protocol). The electrical interconnect may include an electrical path 527A from TIA1 to I / F1 and an electrical path 527B from I / F1 to DRV1.
[0129] Figure 6 A perspective view 600 of an example OMIB 630 bridging two dies (a first die 610 and a second die 620) and coupled to a light engine 670 is shown. The substrate 640 is optional. The first die 610 and the second die 620 each have an AMS block (a first AMS portion 610A and a second AMS portion 620A) in an area overlapping the OMIB 630, such that the AMS block abuts the OMIB 630 and / or is stacked on the OMIB 630 in a configuration that couples the first die 610 to the OMIB 630 and also couples the second die 620 to the OMID 630 through electrical interconnects. The interconnects can be made of copper pillars or any other suitable electrical interconnects. The offset portion of the OMIB creates an offset region where the OI and / or an external light engine 670 can be positioned. The external light engine 670 provides a light source for the photonic network through the OI. The external device optical interface 671 provides bidirectional communication with the die via an optical fiber between the OI and the external device optical interface 671. Figure 6 , the top surfaces of the OMIB 630 and the substrate 640 are shown as being substantially coplanar, although as previously described, the OMIB 630 need not be flush with the surface of the substrate 640. The bottom surface of the first die 610 or the second die 620 can be physically and electrically coupled to the top surface of the OMIB 630. If the first portion of the transceiver (the modulator and / or PD in the OMIB 630) is aligned with the second portion of the transceiver (the AMS portion 610A or the AMS portion 620A), the OMIB 630 and the first die 610 or the second die 620 form an abutting coupling.
[0130] Fig. 7A A perspective view of an example OMIB 730 bridging two dies and receiving unmodulated light from a core light engine (CLE) is shown. Substrate 740 is optional. The first die 710 and the second die 720 each include an AMS block (710A, 720A) in an area overlapping the OMIB 730, such that the AMS block is above and / or stacked on the OMIB 730, which is configured to couple the two dies to the OMIB 730 through suitable electrical interconnects (e.g., where the modulator is directly above the driver and the TIA is directly above the PD). The CLE is located in an offset region of the OMIB 730. The CLE includes a laser that can be placed parallel to the surface of the OMIB 730, a first optical component (e.g., a mirror 701 that reflects the laser beam downward or to one side), and a second optical component (e.g., a lens 702 that focuses the laser beam to a smaller area), wherein the first optical component and the second optical component can be configured to turn the light substantially ninety (90) degrees so that the light from the laser can be directed into the OMIB. Inside the OMIB, the light can enter the GC that also includes a 90 degree mirror so that the light can enter a waveguide parallel to the surface of the OMIB. The external device optical interface 771 communicates bidirectionally with the die via an optical fiber between the OI and the external device optical interface 771.
[0131] Figure 7B An example package 700 in which a CLE 775 may be used is shown. The substrate 741 is optional. The optical signal is provided from the CLE 775 to the grating coupler CG1 (or any other interface device between the optical fiber and the photonic IC) and the splitter SP through the OMIB 731, from which it is directed to the first die 711 and the second die 721. The OMIB 731 also provides an optical path in both directions between the dies. A second grating coupler CG2 and a FAU, or other associated fiber optic connectors using any other interface device between the optical fiber and the photonic IC, can provide access to and from the external device optical interface 771. In some embodiments, the CLE 775 may be included in a CLE array. In other embodiments, the CLE 775 may include a semiconductor optical amplifier (SOA) or an SOA array to amplify the power of unmodulated light.
[0132] Fig. 8Ais a schematic diagram of an example OMIB 830 with a photonic link capable of connecting two dies within a package according to some embodiments. The OMIB 830 provides optical communication from each modulator (MOD1-4) in the first die 810 to each photodetector (PD1-4) in the second die 820. The waveguide system in the OMIB can bridge and / or enable the connection between the transmission unit (not shown) in the first die 810 and the receiving unit (not shown) in the second die 820. The optical engine 870, which can be located inside or outside the OMIB 830, outputs carrier light of a single wavelength λa1. The splitter tree SPT can split the light of wavelength λa1 on multiple optical paths (four of which have been drawn) leading to the modulators MOD1-4. The modulators modulate the light having a single wavelength λa1 and transmit the modulated light to the photodetectors PD1-4 through their respective waveguides. Fig. 8A Four links are shown, each link including a modulator, a photonic path, and a photodetector. Each link can be part of a channel, which also includes a serializer and a deserializer. Multiple channels can be bonded together in a single channel, which can be unidirectional or bidirectional.
[0133] Figure 8B An embodiment is shown with two OMIBs 831 and 832 coupled between a first die 811 and a second die 821. An optical engine 870 outputs carrier light to a splitter tree SPT that can split the light between multiple optical paths leading to modulators MOD1-4 associated with one or more photon channels. For example, a first bidirectional photon channel can include links MOD1 / PD1 and MOD2 / PD2, and a second bidirectional photon channel can include links MOD3 / PD3 and MOD4 / PD4.
[0134] Fig. 9 is a schematic diagram of an example system for a link using WDM for communication from a die to an external device according to some embodiments. In this example, a first die 910 and an external device are coupled via a grating coupler GC and FAU in OMIB 930 (or any other device for optical interface between optical fiber and photonic IC), optical fiber 991, and external device optical interface 971.
[0135] The optical engine 970 can be internal or external and provides light at multiple wavelengths (e.g., between 2 and 16 wavelengths), such as the four wavelengths λb1, λb2, λb3, λb4 shown, to the OMIB 930. The optical splitter tree SPT (e.g., similar to Fig. 8AThe SPT of the optical modulator (SPT) can divide multiple wavelengths of light between multiple different channels or different links (for example, two links, only one of which is fully shown) located in different OMIBs or different PICs. Carrier light with wavelengths λb1, λb2, λb3, λb4 is provided to the optical modulator MOD1-4, which can modulate four light beams with different parts of the information to be transmitted, and the WDM multiplexer (MUX) combines the four light beams into a single light beam including four different wavelengths. The single light beam is then transmitted to the external device optical interface 971 through the GC, FAU and optical fiber 991.
[0136] Fig.10 According to some embodiments, Fig. 9 An example back channel of a WDM system in OMIB 930. The external device optical interface 971 provides an optical signal to a WDM demultiplexer (DEMUX) in OMIB 930 through an optical fiber 991, a FAU, and a GC. The WDM demultiplexer separates the optical signal into four independent modulated signals with wavelengths λb1, λb2, λb3, and λb4. These four signals are provided to the photodetectors PD1-4 through respective optical waveguides, and the photodetectors PD1-4 are electrically connected to the first die 910 to provide information from the external device to the first die 910.
[0137] Although the embodiments discussed above are directed to a WDM multiplexer that displays four optical links in one direction and receives four different wavelengths, in other embodiments, two or more optical links and a WDM multiplexer that receives two or more different wavelengths may be used. Therefore, the WDM demultiplexer will output two or more different wavelengths corresponding to these alternative embodiments.
[0138] Using intra-OMIB and inter-OMIB photonic channels, e.g., as described above, typically including one or more links in each direction, processors in an EIC in a single package can be connected into an electro-optical network. The resulting network topology typically depends on the choice of die pairs coupled via the associated photonic channels; various example topologies are known in the art. Note that while this document generally refers to bidirectional photonic channels, which enable network structures to provide greater flexibility for implementing ML and other computational models compared to unidirectional photonic channels, electro-optical networks can in principle also be formed using unidirectional photonic channels, and such networks can retain any of the benefits discussed herein (e.g., power savings due to transmission of photonic data over longer distances).
[0139] As previously described, a photonic channel includes at least two sets of one or more links that are unidirectional, capable of forming a bidirectional channel. Examples of such channels include, but are not limited to, a photonic channel between two die when bridged by an OMIB and a photonic channel between an OI on the OMIB and an external device optical interface. The nature of the external device optical interface may vary, as long as it has the optical capability to receive messages sent from the OMIB and / or send messages that can be received and used by the OMIB or any die using the OMIB as a bridge.
[0140] Messages can be in the form of packets of varying sizes. Fig.11 An example unidirectional logical channel 1100 including multiple photonic links is shown. The logical channel 1100 has an input link 1101, a transmit bonding engine 1102, photonic paths 1103A-D, a receive bonding engine 1104, and an output link 1105. The transmit bonding engine 1102 can split incoming data packets on the active photonic links and add sequence information. The receive bonding engine 1104 reassembles the data packets using the sequence information. In the depicted example, data arrives at the receive bonding engine via four links.
[0141] Fig.12 Shows Fig.11 1100. The message includes words W0-W7. The sending binding engine 1102 splits these words on four active photonic links. For example, the first link conveys words W0 and W4. The second link conveys words W1 and W5, and so on. The sending binding engine 1102 adds sequence information to each partial message, and the receiving binding engine 1104 reassembles the complete message at its output. Binding can allow high bandwidth, low latency channels to be built from multiple lower bandwidth links, and can provide the ability to avoid invalid links.
[0142] Fig.13 An example AMS block is shown in more detail. Fig.13 The AMS transmit block 1310A of the first die and the AMS receive block 1320A of the second die are shown. To simplify the illustration, the AMS receive block 1315 of the first die and the AMS transmit block 1328 of the second die are not described in detail. Each of the AMS blocks 1310A and 1320A can be connected to the OMIB 1330 through one or more electrical interconnects (only the interconnect 1314 of the AMS transmit block 1310A and the interconnect 1321 of the AMS receive block 1320A are shown herein). Fig.13As shown, each link in the photonic channel includes an optical transmission unit Tx, an optical receiving unit Rx, and an optical transmission medium (e.g., an optical waveguide or optical fiber) connecting the transmission unit to the receiving unit, in this case, the AMS transmission block 1310A is coupled to the AMS receiving block 1320A, and the AMS transmission block 1335 is coupled to the AMS receiving block 1336. The transmission unit (e.g., the transmission unit 1331) includes an optical modulator (e.g., modulator 1333) and an electrical serializer (e.g., serializer 1313), the optical modulator transmits the message onto the optical signal by modulating the carrier light output by the optical engine 1370, and the electrical serializer converts the electronic message received in the form of parallel data words into a signal suitable for driving the optical modulator. The receiving unit (e.g., receiving unit 1332) includes a photodetector (e.g., photodetector 1334) for converting an optical signal received through a transmission medium back into an electrical signal, an electronic device including a transimpedance amplifier (e.g., TIA 1322) and a gain control (e.g., gain control 1323) for normalizing a signal level, a limiter (e.g., limiter 1324) for extracting a bit stream, and a deserializer (e.g., deserializer 1325) for converting a received message back into parallel data. The AMS transmission block 1310A also includes a transmission binding engine 1312, which can split the message from the first interface 1311, as shown in reference Fig.11 As detailed, the AMS receiving block 1320A includes receiving a message from the binding engine 1326 to reassemble the second interface 1327, as shown in reference Fig.12 The OMIB 1330 may also include a receive block 1336 that interfaces with the AMS receive block 1315, a transmit block 1335 that interfaces with the AMS transmit block 1328, an optical engine 1370, and an optical interface OI2.
[0143] In one embodiment, information is modulated at 56 Gb / s in the form of non-return-to-zero (NRZ) code, but more spectrally efficient modulation schemes such as PAM-4 or PAM-8 or higher-order pulse amplitude modulation may be used to achieve higher bandwidth and lower latency links.
[0144] Fig.14AAn arrangement of three OMIBs bridging a portion of two dies is shown. OMIB 1430A bridges two computing elements 1411 in a first die 1410, a first portion of a central region 1412 in the first die 1410, two computing elements 1421 in a second die 1420, and a first portion of a storage region 1422 in the second die 1420. OMIB 1430B bridges two computing elements 1421 in the second die 1420 and a second portion of a storage region 1422 in the second die 1420. OMIB 1430C bridges two computing elements 1411 in the first die 1410, a second portion of a central region 1412 in the first die 1410, two computing elements 1421 in the second die 1420, and a third portion of a storage region 1422 in the second die 1420. The central area 1412 is located in the center of the first die 1410 to balance the delay between the computing element 1411 and the central area 1412. The central area 1412 may include memory, cache memory, other shared memory, on-chip network crossbar switches, switches, routing mechanisms, and memory controllers. The memory area 1422 is located in the center of the second die 1420 to balance the delay between the computing element 1421 and the memory area 1422. The memory area 1422 may include cache memory, other shared memory, and a memory controller. The die portion of the OMIB bridge has 10 computing elements and 2 memory areas. Additional OMIBs (not shown) may be located on opposite sides of the first die 1410 and the second die 1420 to bridge additional dies, computing elements, and portions of the memory area. This arrangement enables data packets to be optically routed from the external device optical interface 1471 to the AMS block, which is close to the memory being used or accessed, and close to the location where the computing element 1411 or 1421 can perform one or more calculations on the data in the memory area. The ability of OMIB to carry instructions from any computing element or external device, for example, using data packets carried as photonic signals via OMIB to the interior of one of the dies 1410 or 1420, to load or store data in memory areas 1412 or 1422 and / or to process data using computing elements 1411 or 1421 near the central area 1412 or 1422, solves many problems of stranded memory and / or latency introduced when optical signals cannot extend beyond the edge of the die 1410 or 1420.
[0145] The AMS transmit and receive blocks can send optical signals from any external device optical interface connected to the OMIB via an inter-OMIB link, or from an AMS transmit and receive block with an inter-OMIB connection within a bridge, to the central area of any die (e.g., to a memory controller to access memory in the central area). The OIs of each OMIB can also be connected by optical fiber, although this is not required. A first side (e.g., right side) of the AMS transmit block 1413 or the AMS receive block 1414 is aligned with a first side (e.g., right side) of a central area (e.g., central area 1412). The first side of the central area is close to and / or contacts a computing element 1411 such as a CPU, GPU, TPU, etc. A second side (e.g., left side) of the AMS transmit block 1413 or the AMS receive block 1414 is aligned with a left side of the OMIB. The alignment of the sides is approximate and does not need to be exact, but typically the alignment is in at least two dimensions to allow adjacency between optical and electrical elements in the OMIB and the AMS block, respectively.
[0146] Fig. 14B An example arrangement of five dies 1410-1414 bridged by an OMIB is shown. Using this arrangement of package 1440, bridges can be infinitely added in two dimensions as needed. In this configuration, each OMIB has two interconnect regions for attaching dies. L2 cache or other memory areas can be located and / or manufactured in the central portion of the die so that the memory area can be optically reached inside the die (rather than at the edge), which results in less latency. Optical channels can also be provided between OMIBs, such as connecting two OMIBs via optical fiber. The OI can be located in an offset region of the OMIB. Fig. 14B The arrangement can be extended infinitely in the x and y directions. In addition, Fig. 14B The arrangement allows routing over multiple hops in the diagonal direction, thus significantly further reducing routing delays.
[0147] Fig. 14C An example arrangement of dies 1413-1417 is shown bridged by the OMIB and bridge 1419. In some implementations, the OMIB and bridge are combined in a single die.
[0148] As shown below, one OMIB can bridge four dies. For example, the OMIB is used to bridge the dies vertically and horizontally, which makes the 4 interconnect areas on each OMIB correspond to the 4 AMS blocks on the respective dies. The OMIB in this arrangement can provide channels in six directions, including two horizontal directions, two vertical directions, and two diagonal directions.
[0149] Fig.14DAn offset checkerboard of OMIB bridge die examples (1410, 1418) in two dimensions is shown. Each OMIB has four interconnect regions and can bridge four dies using six links (2 vertical links, 2 horizontal links, and 2 diagonal links), representing 6 intra-die connections on each OMIB. Some dies (such as die 1418) may include memory regions. The AMS block may interface with the memory regions as well as any processing elements and other logic. Note that Fig.14D May not be drawn to scale. An OMIB of similar size to the die being bridged may reach an AMS block near the center of the die. Furthermore, the AMS block may be relatively much smaller than shown in the example, and each die and each OMIB may have more than four interconnect regions.
[0150] Fig.15 is a flow chart of an example method 1500 for making an OMIB according to various embodiments. The method 1500 includes the following operations.
[0151] Operation 1510 - Fabricating a bridge comprising a photonic link from a first interconnect region to a second interconnect region, wherein the photonic link comprises a first electrical interconnect, a modulator coupled to the first electrical interconnect, an optical transmission medium coupled to the modulator, a photodetector coupled to the optical transmission medium, and a second electrical interconnect. In some embodiments, the modulator is configured to be temperature stabilized by applying a stabilizing voltage to the modulator, wherein the stabilizing voltage is related to a die temperature, and wherein the stabilizing voltage causes a change in electrical absorption in the modulator.
[0152] Operation 1520 - Positioning interconnect regions to enable electrical interconnect and / or abutment coupling with AMS blocks in the die.
[0153] Operation 1530 - Making intra-OMIB connections between interconnect regions.
[0154] Operation 1540 - Making an inter-OMIB connection between the two interconnect regions and the optical interface.
[0155] Fig.16The components of the example system 1600 that can use one or more OMIBs are shown. Each component or component group may include an AMS block and an interface to couple with one or more OMIBs. For example, system 1600 may include a display device interface circuit, such as a connector and a driver to which a display device 1606 may be coupled. In this case, it is not necessary for display device 1606 to be integrated into system 1600 or to become a component of system 1600. Similarly, system 1600 may not have audio input device 1624 or audio output device 1608, but may have an audio input or output device interface circuit (such as a connector and a supporting circuit) for coupling to an external audio input device or an external audio output device.
[0156] The system 1600 may include one or more processing devices 1602. The processing device 1602 may include one or more digital signal processors (DSPs), application specific integrated circuits (ASICs), central processing units (CPUs), graphics processing units (GPUs), tensor processing units (TPUs), AI accelerators, fixed gate programmable arrays (FPGAs), load / store units (LDSUs), neural compute engines (NCEs), dot product and / or convolution engines, server processors, or any other suitable processing devices. System 1600 may include memory 1604, which itself may include one or more memory devices, such as volatile memory, non-volatile memory, flash memory, solid-state memory, and / or hard disk drives, including but not limited to: random access memory (RAM) devices (such as static RAM (SRAM) devices, magnetic RAM (MRAM) devices, resistive RAM (RRAM) devices, or conductive bridge RAM (CBRAM) devices), logic devices (e.g., AND gates, OR gates, NAND gates, or NOR gates), NAND flash memory, solid-state drive (SSD) memory, NOR flash memory, CMOS memory, thin film transistor-based memory, phase change memory (PCM), storage class memory (SCM), magnetoresistive memory (MRAM), resistive RAM, DRAM, high bandwidth memory (HBM), DDR-based DRAM, DIMM memory. In some embodiments, memory 1604 may include memory that shares a die with processing device 1602. The memory may be used as a buffer memory and may include embedded dynamic RAM or spin transfer torque magnetic RAM.
[0157] In some implementations, system 1600 can include a communication chip 1612. For example, communication chip 1612 can be configured to manage wireless communications for transferring data to and from the device.
[0158] The communication chip 1612 can implement any of a variety of wireless standards or protocols, including but not limited to Institute of Electrical and Electronics Engineers (IEEE) standards, including Wi-Fi (IEEE 802.11 series), IEEE 802.16 standards (e.g., IEEE802.16-2005 amendment), Long Term Evolution (LTE) project and any amendments, updates and / or revisions (e.g., Advanced LTE project, Ultra Mobile Broadband (UMB) project (also known as "3GPP2"), and others). IEEE 802.16-compatible broadband wireless access (BWA) networks are commonly referred to as WiMAX networks, which is an abbreviation standing for Worldwide Interoperability for Microwave Access and is a certification mark for products that have passed IEEE 802.16 standard conformance and interoperability testing. The communication chip 1612 can operate according to Global System for Mobile Communications (GSM), General Packet Radio Service (GPRS), Universal Mobile Telecommunications System (UMTS), High Speed Packet Access (HSPA), Evolved HSPA (E-HSPA), or LTE networks. The communication chip 1612 can operate according to Enhanced Data for GSM Evolution (EDGE), GSM EDGE Radio Access Network (GERAN), Universal Terrestrial Radio Access Network (UTRAN), or Evolved UTRAN (E-UTRAN). The communication chip 1612 can operate according to Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Digital Enhanced Cordless Telecommunications (DECT), Evolution Data Optimized (EV-DO) and its derivatives, and any other wireless protocols designated as 3G, 4G, 5G and above. In other embodiments, the communication chip 1612 can operate according to other wireless protocols. The system 1600 may include an antenna 1622 to facilitate wireless communication and / or receive other wireless communications (e.g., AM or FM radio transmissions).
[0159] In some embodiments, the communication chip 1612 can manage wired communications, such as electrical, optical, or any other suitable communication protocol (e.g., Ethernet or USB). As described above, the communication chip 1612 can include multiple communication chips. For example, the first communication chip 1612 can be dedicated to short-range wireless communications such as Wi-Fi or Bluetooth, and the second communication chip 1612 can be dedicated to long-range wireless communications such as EDGE, GPRS, CDMA, WiMAX, LTE, EV-DO. In some embodiments, the first communication chip 1612 can be dedicated to wireless communications, and the second communication chip 1612 can be dedicated to wired communications.
[0160] System 1600 may include battery / power supply circuit 1614. Battery power supply circuit 1614 may include one or more energy storage devices (e.g., batteries or capacitors) and / or circuits for coupling components of system 1600 to an external energy source (e.g., AC line power). System 1600 may include display device 1606 (or corresponding interface circuitry, as described above). Display device 1606 may include any visual indicator, such as a head-up display, a computer monitor, a projector, a touch screen display, a liquid crystal display (LCD), a light emitting diode display, a flat panel display, a virtual reality headset, an augmented reality headset, etc. System 1600 may include audio output device 1608 (or corresponding interface circuitry, as described above). Audio output device 1608 may include any device that generates an audible indicator, such as a speaker, headphones, earbuds, a vibrating element, a piezoelectric crystal, etc. System 1600 may include audio input device 1624 (or corresponding interface circuitry, as described above). Audio input device 1624 may include any device that generates a signal representative of sound, such as a microphone, a microphone array, or a digital musical instrument (eg, a musical instrument with a pickup or a Musical Instrument Digital Interface (MIDI) output).
[0161] The system 1600 may include a positioning device 1618 (or corresponding interface circuitry) such as based on a global positioning system (GPS), Galileo, GLONASS, Beidou, IRNSS, NavIC, and / or QZSS. As known in the art, the positioning device 1618 can communicate with a satellite-based system and can receive the position of the system 1600. The system 1600 may include another output device 1610 (or a corresponding interface circuitry, as described above). Examples of other output devices 1610 may include an audio codec, a video codec, a printer, a wired or wireless transmitter for providing information to other devices, or an additional storage device. The system 1600 may include another input device 1620 (or a corresponding interface circuitry, as described above). Examples of another input device 1620 include an accelerometer, a gyroscope, a compass, an image capture device, a keyboard, a cursor control device (such as a mouse), a stylus, a touchpad, a barcode reader, a quick response (QR) code reader, any sensor, or a radio frequency identification (RFID) reader.
[0162] System 1600 can have any desired form factor, such as a handheld or mobile device (e.g., a cell phone, a smart phone, a tablet computer, a laptop computer, an Internet of Things (IoT) device, a netbook computer, an ultrabook computer, a mobile Internet device, a music player, a personal digital assistant (PDA), an ultra-mobile personal computer, and others), a desktop electronic device, a server device or other networked computing component, a printer, a scanner, a monitor, a set-top box, an entertainment control unit, a vehicle control unit, a digital camera, a digital video recorder, or a wearable electronic device. In some embodiments, system 1600 can be any other electronic device that processes data.
[0163] Thermal behavior
[0164] The close spacing and proximity of the EIC and PIC die (less than 2 mm, typically within 50 microns) creates thermal challenges for OMIB. Depending on the type of modulator used, the temperature range of operation within specification can be less than 30 degrees Celsius. However, the die temperature of the EIC can vary greatly depending on ambient temperature, load conditions, supply voltage, and other factors. Modulator types include Mach-Zehnder interferometers (MZIs), ring modulators, and electro-absorption modulators (EAMs). Ring modulators have a very narrow temperature range (less than 1 degree Celsius), while MZIs and EAMs can have an operating range of more than 30 degrees Celsius. Both modulator temperature and bias voltage affect the frequency (wavelength) at which the modulator efficiency reaches its peak. The wavelength of the laser can be unaffected by these parameters, or be affected differently. Therefore, changes in temperature will result in a difference between the laser wavelength and the efficiency peak wavelength of the modulator, affecting the modulation depth.
[0165] Embodiments may include a modulator that is inherently optimal over a desired temperature range. Alternatively, embodiments may provide temperature compensation including temperature sensing or prediction capabilities. Temperature compensation may be fully integrated in the OMIB, or partially integrated in the OMIB and partially integrated in the connection die.
[0166] Figures 17A-17D An example of an EIC / OMIB combination employing temperature compensation to broaden the temperature range of the modulator in the OMIB is shown. Fig.17AAn EIC 1710 is shown, which is physically and electrically coupled to a PIC 1720 (e.g., OMIB) through metal interconnects (such as bumps or copper pillars). The EIC 1710 includes a driver DRV 1711 and a temperature element TE 1712. The TE 1712 may include a temperature sensor, a temperature predictor, or both (the temperature predictor predicts the die or PIC temperature based on known processor load conditions in its software or firmware). In some embodiments, the temperature sensor is located less than two millimeters (2 mm) from the DRV 1711. In other embodiments, the temperature sensor is located less than 50 microns (50 μm) from the DRV 1711. The driver DRV 1711 transmits a high data rate modulated signal to a modulator 1721 in the PIC 1720. The TE 1712 provides a low frequency (compared to the data rate) temperature-dependent bias voltage to the MOD 1721. If the temperature dependence of the bias voltage is inversely matched to the temperature dependence of the modulator over an extended temperature range, it will eliminate the effect of temperature on modulator performance over the extended temperature range. In other embodiments, the TE 1712 includes a lookup table to convert measured or predicted temperature to a temperature compensated modulator bias voltage.
[0167] exist Fig. 17B In the embodiment of the present invention, the EIC 1730 is physically and electrically connected to the PIC 1740 (e.g., OMIB). The EIC 1730 includes a driver DRV 1731, a temperature element TE 1732, and an adder 1733. The PIC 1740 includes a modulator MOD 1741. The adder 1733 adds a low-frequency temperature-dependent bias voltage from the TE 1732 to a high data rate modulated signal from the DRV 1731 and forwards it to the MOD 1741. In some embodiments, the TE 1732 includes a lookup table to convert a measured or predicted temperature into a temperature compensated modulator bias voltage.
[0168] exist Fig. 17C In the embodiment of the present invention, the EIC 1750 is physically and electrically connected to the PIC 1760 (e.g., OMIB). The EIC 1750 includes a driver 1751, and the PIC 1760 (e.g., OMIB) includes a modulator MOD 1761 and a temperature sensor TS1762. The TS1762 is close to the MOD 1761 and provides a temperature-dependent bias voltage to the MOD 1761. Over an extended temperature range, the temperature dependency of the bias voltage is inversely matched to the temperature dependency of the MOD 1761, thereby extending its operating temperature range.
[0169] exist Fig.17DIn the embodiment of the present invention, the EIC 1770 is physically and electrically connected to the PIC 1780 (e.g., OMIB). The PIC 1780 includes a modulator MOD 1781 and a temperature sensor TS 1782. The TS has a temperature sensor output configured to be coupled to an external circuit of the PIC 1780 (e.g., EIC 1770). The EIC 1770 includes a driver DRV 1771, a temperature controller TC 1772, and an adder 1773. The TC 1772 is configured to receive a first temperature-dependent signal from an external circuit of the EIC 1770 (e.g., from the TS 1782 in the PIC 1780). The TC 1772 converts the first temperature-dependent signal into a temperature-dependent bias voltage, whose temperature dependence is opposite to the temperature dependence of the MOD 1781 within the extended temperature range. The TC 1772 can perform the conversion using temperature-linear and temperature-nonlinear analog circuits. Alternatively or additionally, TC 1772 may use digital circuits to perform conversion to access a memory 1774 storing a temperature profile. TC 1772 may be included in a temperature element TE (such as Fig. 17B In some embodiments, TC 1772 includes a lookup table to convert the measured or predicted temperature into a temperature compensated modulator bias voltage.
[0170] Some modulators can provide stable operation over a wide modulation temperature range of optical components, for example, when integrated into an OMIB, even without adding temperature compensation. The modulator can utilize the Franz-Keldysh effect to achieve electrically induced changes in optical absorption. A variety of materials can be used in the modulator, including germanium and its alloys, silicon and its alloys, III-V materials (such as materials based on indium phosphide (InP) or gallium arsenide (GaAs) material systems). For example, one or more embodiments described herein relate to chip hardware that includes features and functions of providing thermally stable optical modulation elements coupled to electronic components (e.g., drivers in an AMS transmit block). In one or more embodiments, the hardware is a device including an electronic integrated circuit (EIC) and a photonic integrated circuit (PIC). The PIC can be electrically interconnected with the EIC in a coupled or adjacently coupled manner. Each transmitting unit may include a thermally stable optical modulator located in a portion of the PIC. Data may be moved optically in the PIC via an optical carrier located between one of the thermally stable optical modulators in the first portion of the OMIB and one of the receiving units in the second portion of the OMIB or interconnected thereto. In one or more embodiments, the thermally stable optical modulator operates over a temperature range greater than thirty degrees Celsius. In this example, the thermally stable optical modulator may also include a material selected from germanium, silicon, germanium alloys, silicon alloys, indium phosphide (InP) based III-V materials, and gallium arsenide (GaAs) based III-V materials. In one or more embodiments, the optical modulator is an electro-absorption modulator (EAM) that uses the Franz-Keldysh effect for electrically induced charge in optical absorption.
[0171] In one or more embodiments, the thermally stable optical modulator is an EAM that operates within a temperature range of less than thirty degrees Celsius. In this example, the thermally stable optical modulator may include (e.g., contain) a material selected from germanium, silicon, germanium alloys, silicon alloys, III-V materials based on indium phosphide (InP), and III-V materials based on gallium arsenide (GaAs). In one or more embodiments, the thermally stable optical modulator uses a quantum confined Stark effect (QCSE) to achieve an electrically induced change in optical absorption. In one or more embodiments, the thermally stable optical modulator has an output with a high optical modulation amplitude. In this example, the thermally stable optical modulator may include a material selected from germanium, silicon, germanium alloys, silicon alloys, III-V materials based on indium phosphide (InP), and III-V materials based on gallium arsenide (GaAs). In one or more embodiments, the thermally stable optical modulator uses a quantum confined Stark effect (QCSE) to achieve an electrically induced change in optical absorption. In one or more embodiments, the thermally stable optical modulator is configured to operate stably over a wide temperature range. In this example, the thermally stable optical modulator may include a material selected from germanium, silicon, germanium alloys, silicon alloys, indium phosphide (InP) based III-V materials, and gallium arsenide (GaAs) based III-V materials. In one or more embodiments, the modulator uses the Franz-Keldysh effect to achieve electrically induced changes in optical absorption.
[0172] manufacture
[0173] Fig.18An example method 1800 for manufacturing a system is shown. The method 1800 includes step 1810, which includes manufacturing a system including a first die (integrated circuit), a second die (integrated circuit), a photon IC providing a bridging function, a light source, and a data channel. The photon IC includes a first interconnect region, a second interconnect region, and an offset region. The pad pattern on the first integrated circuit matches the pad pattern in the first interconnect region, and the pad pattern on the second integrated circuit matches the pad pattern in the second interconnect region. Two or more pads on the first integrated circuit are physically and electrically coupled to two or more pads in the first interconnect region, and two or more pads on the second integrated circuit are physically and electrically coupled to two or more pads in the second interconnect region. The light source is optically coupled to a first optical interface (first OI) in the offset region. The data channel includes a serializer and a driver in the first integrated circuit, a modulator, a photon path, and a photodetector in the photon IC, and a transimpedance amplifier and a deserializer in the second integrated circuit. The distance between the surface of the first integrated circuit and the surface of the photon IC is less than 2 mm, and the distance between the surface of the second integrated circuit and the surface of the photon IC is less than 2 mm. In many cases, the distance is less than fifty microns (50 μm). At least one of the two or more pads on the first integrated circuit may be located more than 100 microns (100 μm) from an edge of the first integrated circuit. At least one of the two or more pads on the second integrated circuit may be located more than 100 microns (100 μm) from an edge of the second integrated circuit. The modulator may be an electro-absorption modulator (EAM). The modulator may be a Mach-Zehnder interferometer (MZI). The first die or photonic IC may include a temperature sensor or temperature predictor that may transmit a temperature-dependent bias signal.
[0174] Precautions
[0175] Although described with respect to specific embodiments thereof, these specific embodiments are merely illustrative and not restrictive. The description may cite specific structural embodiments and methods, and is not intended to limit the technology to the specifically disclosed embodiments and methods. Other features, elements, methods, and embodiments may be used to practice the technology. The embodiments are described to illustrate the technology, not to limit its scope as defined by the claims. Those of ordinary skill in the art recognize various equivalent variations of the above description.
[0176] For example, many examples in this document use grating couplers to couple optical fibers to photonic ICs. However, many embodiments work well using other methods of optical interface between optical fibers and photonic ICs, such as reference Figure 3Although many examples show only the OMIB bridging the first die and the second die, the OMIB can bridge any number of dies. Although most examples show the AMS functionality included in the first die and / or the second die, the AMS functionality can be partially or fully included in a separate stacked IC, such as Figure 5D-Figure 5E All such embodiments are within the scope and ambit of the disclosed technology.
[0177] All features disclosed in the specification (including claims, abstract and drawings), and all steps in any disclosed method or process, may be combined in any combination, except for combinations where at least some of such features and / or steps are mutually exclusive. Unless expressly stated otherwise, each feature disclosed in the specification (including claims, abstract and drawings) may be replaced by an alternative feature serving the same, equivalent or similar purpose.
[0178] Although described with respect to specific embodiments thereof, these specific embodiments are merely illustrative and not restrictive. For example, many operations can be implemented using off-the-shelf equipment on a printed circuit board (PCB), in a system on a chip (SoC), an application-specific integrated circuit (ASIC), a programmable processor, a coarse-grained reconfigurable architecture (CGRA), or in a programmable logic device (such as a field programmable gate array (FPGA)), thereby avoiding the need for at least a portion of any dedicated hardware. The implementation may be as a single chip, or as a multi-chip module (MCM) that encapsulates multiple semiconductor dies in a single package. All such variations and modifications are considered to be within the scope of the technology disclosed herein, the nature of which is determined by the foregoing description.
[0179] Any technology suitable for manufacturing electronic devices can be used to implement the circuits of a specific embodiment, including complementary metal oxide semiconductors, fin field effect transistors, bipolar complementary metal oxide semiconductors, bipolar JFETs, MOS, NMOS, PMOS, HBTs, MESFETs, etc. Different semiconductor materials (such as silicon, germanium, SiGe, GaAs, InP, GaN, SiC, graphene, etc.) can be used. The circuit can have single-ended or differential inputs, and single-ended or differential outputs. The terminals of the circuit can be used as inputs, outputs, or both, or in a high impedance state, or they can be used to receive power, ground references, reference voltages, reference currents, or other. Although the physical processing of the signals can be presented in a specific order, the order can be changed in different specific embodiments. In some specific embodiments, multiple elements, devices, or circuits shown sequentially in this specification can operate in parallel.
[0180] Specific embodiments may be implemented using programmed general purpose digital computers, application specific integrated circuits, programmable logic devices, field programmable gate arrays, optical, chemical, biological, quantum or nanoengineered systems, etc. Other components and mechanisms may be used. In general, the functions of a specific embodiment may be implemented by any means known in the art. Distributed, networked systems, components and / or circuits may be used. The communication or transmission of data may be wired, wireless or any other means.
[0181] It should also be understood that one or more of the elements shown in the drawings / figures may also be implemented in a more separate or integrated manner, or even removed or rendered inoperable in some cases, as may be useful depending on the particular application.
[0182] Therefore, although specific embodiments are described herein, modifications, various changes and substitutions are intended in the foregoing disclosure, it should be understood that in some cases, certain features of specific embodiments will be employed without corresponding use of other features without departing from the scope and spirit. Therefore, many modifications may be made to adapt a particular situation or material to the basic scope and spirit.
Claims
1. A package, include: a bridge element having a first interconnect region and a second interconnect region for electrically interconnecting to one or more dies; a first photon path from the first interconnected region to the second interconnected region; as well as A second photon path from the second interconnected region to the first interconnected region.
2. The package according to claim 1, further comprising: include: a third photonic path from said first interconnect region to an optical interface (OI); as well as A fourth photon path from the OI to the first interconnect region.
3. The package according to claim 1, further comprising: include: A first portion of a photonic transceiver is coupled to the first interconnect region, the first portion of the photonic transceiver transmitting a first optical signal through the first photonic path and receiving a second optical signal through the second photonic path.
4. The package according to claim 3, in, The first part of the photonic transceiver comprises a modulator and / or a photodetector.
5. The package according to claim 3, in: The first portion of the photonic transceiver is also configured to transmit a third optical signal through the third photonic path and to receive a fourth optical signal through the fourth photonic path.
6. The package according to claim 3, in: The first portion of the photonic transceiver is located in a photonic integrated circuit (PIC), and the second portion of the photonic transceiver is located in an electronic integrated circuit (EIC), and the first portion and the second portion are coupled by an electrical interconnect having a length of less than two millimeters (2 mm).
7. The package according to claim 6, in: The first portion includes an optical modulator for modulating the first optical signal or a third optical signal in response to a modulation signal from the second portion through the electrical interconnect, wherein the third optical signal traverses a third optical path.
8. The package according to claim 7, in: The modulator is an electro-absorption modulator made of a material selected from germanium, silicon, germanium alloy, silicon alloy, indium phosphide (InP) based III-V material and gallium arsenide (GaAs) based III-V material.
9. The package according to claim 7, in: The EIC is a component of the first die coupled to the first interconnect region, the EIC comprising a central region, wherein the second portion is located in or within two (2) millimeters of the central region.
10. The package according to claim 9, in: The central area includes one or more of a memory, a cache, an on-chip network crossbar, a switch, or a routing mechanism.
11. The package according to claim 9, in: The central region intersects a center of the first die.
12. The package according to claim 9, in: The EIC also includes a plurality of computing elements arranged around the central region, wherein the second portion is located in or within two (2) millimeters of one of the computing elements.
13. The package according to claim 1, further comprising: include: A substrate coupled to the bridging element, wherein the bridging element is embedded in the substrate such that a top surface of the bridging element is substantially coplanar with a top surface of the substrate.
14. The package according to claim 12, in: The OI is offset from the first interconnect region or the second interconnect region in at least one dimension on the top surface.
15. The package according to claim 12, in: The first interconnect region extends to the top surface for abutting coupling of the first portion with the second portion on the bottom surface of the first die.
Citation Information
Cited By
Co-packaged optical chip optical interconnection structure based on micro-LED
CN121142733A
