Method, system and apparatus for integrating volatile and non-volatile memory bit cells

By combining the design of volatile and nonvolatile memory arrays in integrated circuits, efficient copying and transfer of data between volatile and nonvolatile memory is solved in the prior art memory devices in reducing size. The durability and density problems during the time improve the performance of the memory.

CN112005302BActive Publication Date: 2024-11-05ARM LTD
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN201980027401.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2018-04-23
Filing Date
2019-03-22
Publication Date
2024-11-05
Estimated Expiration
2039-03-22

AI Technical Summary

Technical Problem

Existing nonvolatile memory devices encounter durability and density problems when reducing sizes, and traditional technologies such as FeRAM and MRAM have high costs and high power consumption, making it difficult to replace flash memory.

Method used

Using an integrated circuit design combining volatile and nonvolatile memory arrays, the efficient copying and transfer of data between volatile and nonvolatile memory is achieved through sharing the data bus and external data ports, reducing power consumption and Waiting time.

Benefits of technology

It improves the efficiency of memory state copying and transfer, reduces power consumption and wait time, enhances the density and durability of the memory, and solves the durability and density problems of traditional technologies when reducing size.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112005302B_ABST
    Figure CN112005302B_ABST
Patent Text Reader

Abstract

The present application discloses a method, system and apparatus for integrating volatile and non-volatile memory bit cells. An integrated circuit device includes: at least a first non-volatile memory array (304); at least a first volatile memory array (302); at least one non-volatile memory data bus (328) for transferring data from non-volatile bit cells in a read operation / transferring data to non-volatile bit cells in a write operation; at least one volatile memory data bus (326) for transferring data to volatile bit cells in a write operation / transferring data from volatile bit cells in a read operation; a shared data bus structure (330); and at least one external data port (332) for transferring values ​​between the shared bus structure (330) and an external terminal of the integrated circuit device.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] Disclosed are techniques for utilizing memory devices. Background Art

[0002] Non-volatile memory is a type of memory in which the memory cell or element does not lose its state after the power supplied to the device is removed. For example, the earliest computer memories made of ferrite rings that could be magnetized in two directions were non-volatile. As semiconductor technology evolved to higher levels of miniaturization, ferrite devices were replaced by more well-known volatile memories such as DRAM (dynamic random access memory) and SRAM (static RAM).

[0003] A type of nonvolatile memory, electrically erasable programmable read-only memory (EEPROM) devices have large cell areas and may require large voltages (e.g., from 12.0 volts to 21.0 volts) on the transistor gate to write or erase. In addition, the erase or write time is typically on the order of tens of microseconds. A limiting factor of EEPROM is the limited erase / write cycle count of no more than 600,000 or 10 5 -10 6 The semiconductor industry has eliminated the need for transfer gate switch transistors between the EEPROM and the nonvolatile transistors by sectorizing the memory array so that a "page" (e.g., a subarray) of the EEPROM can be erased at a time in a flash memory device called a flash memory device. In a flash memory device, the ability to maintain random access (erase / write individual bits) is sacrificed for speed and higher bit density.

[0004] Recently, FeRAM (ferroelectric RAM) has provided low power, relatively high write / read speeds, and endurance of over 10 billion read / write cycles. Similarly, magnetic memory (MRAM) has provided high write / read speeds and endurance, but with a high cost premium and higher power consumption. For example, none of these technologies can reach the density of flash memory devices. Therefore, flash memory remains the non-volatile memory of choice. However, it is generally recognized that flash memory technology may not be easily scalable below 65 nanometers (nm); therefore, new non-volatile memory devices that can be scaled to smaller sizes are being actively sought.

[0005] Technologies considered to replace flash memory devices have included memories based on certain materials that exhibit resistance changes associated with phase changes of the material (determined at least in part by the long-range ordering of atoms in the crystal structure). In one type of variable resistance memory, called a phase change memory (PCM / PCRAM) device, the resistance change occurs as the memory element briefly melts and then cools to either a conductive crystalline state or a non-conductive amorphous state. Typical materials vary and may include GeSbTe, where Sb and Te are interchangeable with other elements on the periodic table having the same or similar properties. However, because these resistance-based memories depend on physical structural phenomena for the transition between conductive and insulating states (e.g., melting at up to 600 degrees Celsius and returning to a solid state that cannot be adequately controlled for useful memory in many applications), they have not proven to be commercially useful.

[0006] Another class of variable resistance memories includes materials that respond to an initial high "forming" voltage and current to activate the variable resistance function. These materials may include, for example, materials with different stoichiometries of x, y, z, and of Transition metal oxides (TMOs), such as CuO, CoO, VO x 、NiO、TiO 2 、 2 O 5 ; and some perovskites, such as Cr; SrTiO 3 . Several of these memory types exist and fall into the classification of resistive RAM (ReRAM) or conductive bridge RAMS (CBRAM) to distinguish them from chalcogenide type memories. It is speculated that the resistive switching in these RAMs is at least partially due to the formation of narrow conductive paths or filaments connecting the top and bottom conductive terminals by an electroforming process, but the existence of such conductive filaments remains a controversial issue. Since the operation of ReRAM / CBRAM can be strongly temperature-dependent, the resistive switching mechanism in ReRAM / CBRAM can also be highly temperature-dependent. Additionally, because the formation and movement of the filaments are random, these systems can operate randomly. Other types of ReRAM / CBRAM can also exhibit unstable quality. Further, the resistive switches in ReRAM / CBRAM tend to fatigue during many memory cycles. That is, after the memory state has changed many times, the resistance difference between the conductive state and the insulating state can change significantly. In commercial memory devices, such changes can cause the memory to exceed specifications and make it unusable. Summary of the invention

[0007] Briefly, one particular embodiment is directed to a device formed in an integrated circuit device, comprising: at least a first nonvolatile memory array, the at least first nonvolatile memory array comprising a plurality of nonvolatile memory bit cells; at least a first volatile memory array, the at least first volatile memory array comprising a first plurality of volatile memory bit cells; at least one nonvolatile memory data bus, the at least one nonvolatile memory data bus configurable to transfer data from the nonvolatile bit cells in a read operation and to transfer data to the nonvolatile bit cells in a write operation; at least one volatile memory data bus configurable to transfer data to the volatile bit cells in a write operation and transferring data from the volatile bit cells in a read operation; a shared data bus structure; and at least one external data port, the at least one external data port being used to transfer input values ​​and output values ​​between the shared bus structure and external terminals of the integrated circuit device, wherein the shared data bus structure is coupled to the at least one non-volatile memory data bus and the at least one volatile memory data bus to enable data transfer between the at least one external data port and the at least one non-volatile memory data bus or at least one volatile memory data bus or a combination thereof, and to enable data transfer between the at least one non-volatile memory data bus and the at least one volatile memory data bus.

[0008] Another specific embodiment relates to a method comprising: applying a first access signal to at least a first word line among a plurality of word lines to enable access to one or more volatile memory bit cells in a first volatile memory array formed in an integrated circuit device and one or more non-volatile memory bit cells in a non-volatile memory array formed in the integrated circuit device; applying a first signal to an external terminal of the integrated circuit device to access at least some of the volatile memory bit cells in the first volatile memory array; and applying a second signal to an external terminal of the integrated circuit device to access at least some of the non-volatile memory bit cells in the non-volatile memory array, wherein the external terminal is coupled to a data port of the integrated circuit device, the data port being configured to access the first volatile memory array or the non-volatile memory array or a combination thereof.

[0009] Another specific embodiment relates to a method, the method comprising: applying a first signal to an external terminal of an integrated circuit device to access a volatile memory bit cell in a volatile memory array formed in the integrated circuit device; and applying a second signal to an external terminal of the integrated circuit device to access a non-volatile memory bit cell in a non-volatile memory array formed in the integrated circuit device, wherein at least one of the volatile memory bit cells and at least one of the non-volatile memory bit cells are connected to a word line among a plurality of word lines formed in the integrated circuit device, and wherein the external terminal is coupled to a data port of the integrated circuit device, the data port being configured to access the volatile memory array or the non-volatile memory array or a combination thereof.

[0010] It should be understood that the foregoing embodiments are merely example embodiments, and that claimed subject matter is not necessarily limited to any specific aspect of these example embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The claimed subject matter is particularly pointed out and distinctly claimed in the concluding portion of the specification. However, as to organization and / or method of operation, and objects, features, and / or advantages thereof, if any, Figure 1 The present invention may be better understood by reference to the following detailed description, in which:

[0012] Figure 1 is a schematic diagram of a computing device according to an embodiment;

[0013] Figure 2 , Figure 3A and Figure 3B is a schematic diagram of a memory system including a volatile memory bit cell and a non-volatile memory bit cell according to an embodiment;

[0014] Figure 4 is a diagram illustrating the timing of operations for copying states between volatile memory bit cells and non-volatile memory bit cells according to an embodiment;

[0015] Figure 5 and Figure 6 is a schematic diagram of a memory system integrating a volatile memory bit cell according to an embodiment and a non-volatile memory bit cell according to an embodiment on a common word line according to an embodiment;

[0016] Fig. 7A and Figure 7B is a diagram illustrating the timing of operations for copying states between volatile memory bit cells and non-volatile memory bit cells according to an embodiment;

[0017] Figure 8is a schematic diagram of a memory system integrating a volatile memory bit cell and a non-volatile memory bit cell according to an embodiment;

[0018] Fig. 9 is a schematic diagram illustrating an addressing scheme for an integrated volatile memory bit cell and a non-volatile memory bit cell according to an embodiment;

[0019] Fig.10 is a schematic diagram of integrating a smaller non-volatile memory bit cell array with a larger volatile memory bit cell array according to an embodiment;

[0020] Fig.11 is a schematic diagram of integrating a smaller volatile memory bit cell array with a larger non-volatile memory bit cell array according to an embodiment;

[0021] Fig.12 is a schematic diagram illustrating the interleaving of volatile memory bit cells and non-volatile memory bit cells according to an embodiment;

[0022] FIG. 13A to FIG. 13E is a schematic diagram of a bit cell circuit including a volatile memory element and a non-volatile memory element according to an embodiment;

[0023] Fig.14A Graph showing current density versus voltage for a CES device according to an embodiment.

[0024] Fig. 14B is a schematic diagram of an equivalent circuit of a CES device according to an embodiment; and

[0025] Fig.15 is a schematic diagram of a three-dimensional integrated circuit structure integrating a volatile memory bit cell and a non-volatile memory bit cell according to an embodiment. DETAILED DESCRIPTION

[0026] In the following detailed description, reference is made to the accompanying drawings, which form a part of the following detailed description, wherein the same reference numerals may indicate the same parts throughout, which are identical, similar and / or similar. It should be appreciated that the drawings are not necessarily drawn to scale, for example, for the sake of simplicity and / or clarity of illustration. For example, the dimensions of some aspects may be enlarged relative to other aspects. Further, it should be understood that other embodiments may be utilized. In addition, structural and / or other changes may be made without departing from the claimed subject matter. References throughout this specification to "claimed subject matter" refer to the subject matter intended to be covered by one or more claims or any part thereof, and are not necessarily intended to refer to a complete set of claims, to a specific combination of a set of claims (e.g., method claims, device claims, etc.), or to a specific claim. It should also be noted that directions and / or references such as upper, lower, top, bottom, etc., for example, may be used to facilitate the discussion of the drawings, without intending to limit the application of the claimed subject matter. Therefore, the following detailed description should not be considered to limit the claimed subject matter and / or equivalents.

[0027] References throughout this specification to an embodiment, an embodiment, an example, an embodiment, and / or the like mean that at least one embodiment and / or embodiment of the claimed subject matter includes specific features, structures, characteristics, and / or the like described with respect to a specific embodiment and / or embodiment. Therefore, the appearance of such phrases, for example, in various places throughout this specification is not necessarily intended to refer to the same embodiment and / or embodiment or to any one specific embodiment and / or embodiment. In addition, it should be understood that the specific features, structures, characteristics, and / or the like described can be combined in various ways in one or more embodiments and / or embodiments, and therefore, within the scope of the predetermined claims. Generally, of course, with respect to the specification of the patent application, these and other issues have the possibility of varying in the specific context of usage. In other words, throughout this disclosure, the specific context of description and / or usage provides helpful guidance on the reasonable inferences to be drawn; however, in the same manner, "in this context" generally refers to the context of the present disclosure without further qualification.

[0028] According to an embodiment, a computing device or computing platform may incorporate a volatile memory system and a non-volatile memory system to perform computing operations. In one embodiment, a memory device may include a "volatile" memory device that can maintain a specific memory state when power is applied to the volatile memory device, but may lose the specific memory state if power is removed. In another embodiment, a memory device may include a "non-volatile" memory that can maintain a specific memory state even after power is removed from the memory device.

[0029] A volatile memory system or a non-volatile memory system may maintain memory states to represent values, symbols, parameters, and / or conditions as memory states such as a "bit cell." In this context, a "bit cell" or "bit cell circuit" as referred to herein includes a circuit or portion of a circuit that is capable of representing a value, symbol, or parameter as a state. For example, a bit cell may include one or more memory devices that are capable of representing a value, symbol, or parameter as a memory state of the memory device. In a particular embodiment, a volatile memory device may be comprised of "volatile memory" bit cells that may lose a detectable memory state after power is removed from the volatile memory bit cells. Similarly, a non-volatile memory device may be comprised of "non-volatile memory" bit cells that are capable of maintaining a detectable memory state after power is removed from the non-volatile memory bit cells.

[0030] According to an embodiment, a computing device or computing platform may include both a non-volatile memory device and a volatile memory device. In a specific embodiment, such a computing device or computing platform may copy or transfer a memory state or a stored value read from a volatile memory device to a non-volatile memory. Similarly, such a computing platform or computing device may copy a memory state read from a non-volatile memory to a volatile memory device. Copying a memory state or transferring a stored value between a volatile memory device and a non-volatile memory device may require latency and power consumption, thereby affecting the performance of the computing platform or device. Additionally, copying a memory state or transferring a stored value between a volatile memory device and a non-volatile memory device may affect the memory bus resources used to transfer values ​​between physical devices. The specific embodiments described herein relate to the coupling of a volatile memory bit cell and a non-volatile memory bit cell to reduce the power consumption and latency associated with copying a memory state between a volatile memory device and a non-volatile memory device.

[0031] Figure 1 1 is a schematic diagram of a computing device 100 according to an embodiment. The processor / controller 104 may execute a process or program (e.g., under the control of computer-readable instructions) to perform various tasks, including, for example, storing values ​​in or reading values ​​from addressable portions of the memory 108. In particular embodiments, the processor / controller 104 may communicate with the memory controller 106 via the bus 102 according to a predefined interface. The processor / controller 104 may provide commands (e.g., specifying a physical memory address) to the memory controller 106 to write values ​​to or read values ​​from addressable portions of the memory 108.

[0032] The memory array 108 may include one or more volatile or nonvolatile memory devices, including, for example, a memory array containing volatile and nonvolatile memory bit cell elements as described herein. The processor / controller 104, the memory controller 106, and the memory 108 may be formed as separate components or integrated in a system on a chip (SoC) along with other components not shown (e.g., sensors, user interfaces, I / O devices). In addition, for example, the processor / controller 104, the memory controller 106, and the memory array 108 may be formed by any of a number of different process technologies, including, for example, the correlated electronic materials (CEM) processes discussed herein, complementary metal oxide semiconductor (CMOS) processes, or other processes for forming nonvolatile memory bit cells or volatile memory bit cells.

[0033] According to an embodiment, the memory 108 may include a volatile memory device including volatile memory bit cells and a non-volatile memory device including non-volatile memory bit cells. Such volatile memory bit cells may include bit cells formed according to any of several circuit structures used to form volatile memory bit cells such as the following: SRAM bit cells, DRAM bit cells, to provide just a few examples. Such non-volatile memory bit cells may be formed according to any of several non-volatile memory bit cells such as the following: flash memory bit cells, correlated electronic memory bit cells, phase change memory (PCM) bit cells, magnetic memory bit cells, to provide just a few examples. As described below in a specific embodiment, the non-volatile memory bit cells and volatile memory bit cells formed in the memory 108 may be integrated to enable the use of any of several different techniques to copy memory states between the non-volatile memory bit cells and the volatile memory bit cells.

[0034] Figure 2 is a schematic diagram illustrating a process of copying states or transferring stored values ​​between or among volatile memory bit cells and non-volatile memory bit cells, for example within memory 108 . Figure 3A304. The integrated circuit device 300 is a schematic diagram of a specific implementation in which memory states may be copied or stored values ​​may be transferred between or among volatile memory bit cells 302 and non-volatile memory bit cells 304 via a memory bus (e.g., a fixed bit length memory bus) coupled between the volatile memory bit cells 302 and the non-volatile memory bit cells 304. In addition to the volatile memory bit cells 302 and the non-volatile memory bit cells 304, the integrated circuit device 300 may also include a shared data bus structure 330 and an external data port 322. According to an embodiment, the shared data bus structure 330 may facilitate copying or transferring stored values ​​between or among the volatile memory bit cells 302 and the non-volatile memory bit cells 304. The shared data bus structure 330 may also facilitate transferring stored data values ​​between the external data port 332 and the volatile memory bit cells 302 or the non-volatile memory bit cells 304.

[0035] According to an embodiment, the volatile memory bitcell 302 and the non-volatile memory bitcell 304 may be formed in a single integrated circuit device 300, wherein the volatile memory bitcell 302 may be formed in one or more volatile memory arrays and the non-volatile memory bitcell 304 may be formed in one or more non-volatile memory bitcells.

[0036] The integrated circuit device 300 may include a plurality of external signal pins, such as, for example, signal pins VM sei (volatile memory select), VM addr (volatile memory address), VM R / W (volatile memory read / write), VM cntl (volatile memory control), and VM clk (volatile memory clock). Likewise, the integrated circuit device 300 may include a plurality of external signal pins, such as, for example, signal pins NVM sel (non-volatile memory select), NVM addr (non-volatile memory address), NVM R / W (non-volatile memory read / write), NVM cntl (non-volatile memory control), and NVM clk (non-volatile memory clock). The integrated circuit device 300 may further include signaling pins 322 and 324 to at least partially form a single data port that can be configured to transfer data between an external device (not shown) and the volatile memory bit cell 302 or the non-volatile memory bit cell 304. In addition, the integrated circuit device 300 may include shared control signals, such as a shared clock signal clk, a shared address signal addr, etc. In a specific example, the shared clock signal clk may control a memory cycle (e.g., for a read operation and / or a write operation) applied to the volatile memory bit cell 302 or the non-volatile memory bit cell, or both. In addition, the shared address signal addr may be used to access a bit cell formed in the volatile memory bit cell 302 or the non-volatile memory bit cell, or both.

[0037] In a particular embodiment, the integrated circuit device 300 including the volatile memory bit cells 302, the non-volatile memory bit cells 304, and the shared bus structure 330 may be formed according to a digital circuit design within a single register transfer level (RTL) boundary that defines a synchronous digital circuit (e.g., in terms of the flow of digital signals between or among registers and the operations performed on such digital signals). In a particular embodiment, the digital circuit design within a single RTL boundary may be defined according to a hardware description language (HDL) such as, for example, Verilog or VHDL based on a high-level representation of the circuit.

[0038] The integrated circuit 300 further includes at least one volatile memory data bus 326, which is configurable to transfer data signals to the volatile memory bit cells 302 in a write operation and to transfer stored values ​​obtained from the volatile bit cells 302 in a read operation. Likewise, the integrated circuit 300 further includes at least one non-volatile memory data bus 328, which is configurable to transfer stored values ​​to the non-volatile bit cells 304 in a write operation and to transfer data obtained from the non-volatile bit cells 304 in a read operation. Figure 3BA specific embodiment includes a shared data bus structure 330 and an external data port 332 Figure 3A A schematic diagram of a specific implementation of an integrated circuit device 300 is shown in . Here, the value of the signal 306 provided to the multiplexers 308 and 310 may indicate whether to copy the state from the volatile memory cell 302 to the non-volatile memory cell 304 or to copy the state from the non-volatile memory cell 304 to the volatile memory bit cell 302.

[0039] In this context, a "read operation" as referred to herein refers to an operation implemented by a circuit to detect the memory state of one or more bit cells. Further in this context, a "write operation" as referred to herein refers to an operation implemented by a circuit to place one or more bit cells in a specific memory state. For example, a write operation may include generating a "programming signal" having specific properties (e.g., voltage and / or current) that may be applied to one or more portions of a bit cell to place the bit cell in a specific memory state (e.g., a memory state detectable in a subsequent read operation).

[0040] In one embodiment, the memory state or stored value of one or more volatile memory bit cells 302 may be copied or transferred to one or more non-volatile memory bit cells 304. In this context, values ​​stored in one or more first memory bit cells may be "transferred" to one or more second memory bit cells by placing the one or more second memory bit cells in a particular memory state so as to store or represent the values ​​stored in the one or more first memory bit cells according to a particular mapping between stored values ​​and memory states. In the particular embodiment of FIG. 3 , transferring the stored values ​​from the one or more volatile memory bit cells 302 to the one or more non-volatile memory bit cells 304 may include one or more read operations applied to the one or more volatile memory bit cells 302 to detect the memory state of the one or more volatile memory bit cells 302, followed by one or more write operations applied to the one or more non-volatile memory bit cells 304. Similarly, transferring stored values ​​from one or more non-volatile memory bit cells 304 to one or more volatile memory bit cells 302 may include one or more read operations applied to one or more non-volatile memory bit cells 304 to detect the memory state of one or more non-volatile memory bit cells 304, followed by one or more write operations applied to one or more volatile memory bit cells 302.

[0041] like Figure 3BAs shown in the specific embodiment of the present invention, the external data port 332 may include external pins 322 and 324. In one embodiment, the volatile memory data bus 326 may be configurable to transfer data received at the external signal pin 322 to the volatile memory bit cell 302 in a write operation according to the first state of the multiplexer 308. In addition, at least one volatile memory data bus 326 may be configurable to signal to transfer data retrieved from the volatile memory bit cell 302 to the external signal pin 324 in a read operation according to the first state of the multiplexer 312. Similarly, in another embodiment, at least one non-volatile memory data bus 328 may be configurable to transfer data received at the external signal pin 322 to the non-volatile memory bit cell 304 in a write operation according to the first state of the multiplexer 310. In addition, at least one non-volatile memory data bus 328 may be configurable to transfer data retrieved from the non-volatile memory bit cell 304 to the external signal pin 324 in a read operation according to the second state of the multiplexer 312. Thus, by setting the states of multiplexers 308 , 310 , and 312 , the integrated circuit device 300 may configure the external data port 322 formed by external signal pins 322 and 324 to transfer data between an external device and the volatile memory bit cell 302 or the nonvolatile memory bit cell 304 .

[0042] In another embodiment, the at least one volatile memory data bus 326 and the at least one volatile memory data bus 328 may be configured to transfer stored values ​​between the volatile memory bit cells 302 and the non-volatile memory bit cells 304 independently of the data ports formed by the external signal pins 322 and 324. In one particular implementation, the volatile memory bit cells 302, the non-volatile memory bit cells 304, the at least one volatile memory data bus 326, and the at least one volatile memory data bus 328 may be configured to transfer stored values ​​between the volatile memory bit cells 302 and the non-volatile memory bit cells independently of the data ports formed by the external signal pins 322 and 324 by applying a combination of signal conditions (e.g., including signal conditions affected by voltage levels, current levels, signal timing, etc.) to external signal pins (e.g., VM R / W, VM addr, VM sel, NVM R / W, NVM addr, NVM sel, etc.) of the integrated circuit 300. Applying such voltages to the external signal pins of the integrated circuit 300 may, for example, place the multiplexer 308 in the second state, thereby enabling the at least one volatile memory data bus 326 to receive the stored value transferred from the non-volatile memory bit cell 304 in a read operation. Here, the stored value transferred from the non-volatile memory bit cell 304 received in the read operation may be stored in the volatile memory bit cell 302 in a subsequent write operation. Similarly, applying the combination of voltages to the external signal pins of the integrated circuit 300 may place the multiplexer 310 in the second state, thereby enabling the at least one volatile memory data bus 310 to receive the data transferred from the volatile memory bit cell 302 in a read operation. Here, the stored value transferred from the volatile memory bit cell 302 received in the read operation may be stored in the non-volatile memory bit cell 304 in a subsequent write operation.

[0043] In one embodiment, the volatile memory data bus 326 and the non-volatile memory data bus 328 can have the same bus width (e.g., byte or word) to transfer the same amount of data or retrieved stored values ​​over a memory cycle. For example, the buses 314 and 316 can include the same number of conductors, each capable of sending a signal representing a single bit or symbol. In an alternative embodiment, the volatile memory data bus 326 and the non-volatile memory data bus 328 can have different bus widths. For example, the volatile memory data bus 326 can have a bus width that is an integer multiple of the bus width of the non-volatile memory data bus 328. In this example, the multiplexer 310 can split the data transferred from the volatile memory bit cell 302 in a single memory cycle for storage in the non-volatile memory element 304 in a write operation during multiple memory cycles. Likewise, if the bus width of the nonvolatile memory data bus 328 is an integer multiple of the bus width of the volatile memory data bus 326, the multiplexer 308 can split the data transferred from the nonvolatile memory bit cells 304 in a single memory cycle for storage in the nonvolatile memory elements 302 in write operations during multiple memory cycles.

[0044] Figure 4 The timing of the operation of copying the state from the volatile memory bit cell (e.g., the volatile memory bit cell 302) to the non-volatile memory bit cell (e.g., the non-volatile memory bit cell 304) is illustrated. It should be understood that Figure 4 The specific timing of the operations illustrated in are merely examples of timing, and other variations of timing may be employed without departing from the claimed subject matter. For example, specific variations may involve active low word lines and / or pre-signals, true pre-charge versus post-charge schemes, and the like. In an embodiment, volatile memory bit cells and non-volatile memory bit cells may be accessed for write operations in response to applying a voltage signal on a word line to couple a bit cell circuit to one or more bit lines. A sense amplifier (not shown) may be maintained in an equilibrium mode until a word line signal is valid. For example, such a sense amplifier may not become active until a certain signal level on the bit line is reached. Additionally, a write driver circuit (not shown) may be enabled prior to activating the word line.

[0045] In a first clock cycle, a voltage on a word line coupled to one or more volatile memory bit cells (e.g., one or more volatile memory bit cells 302) may be raised to enable a read operation to be applied to one or more VM bit cells. A signal VM WL high active may indicate that access to the associated volatile memory bit cell is enabled if the associated signal is high, while a signal NVM WL high active may indicate that access to the associated non-volatile memory bit cell is enabled if the associated signal is high. A signal VM Pre low active in a lower state may indicate precharging of the bit line of the volatile memory bit cell (e.g., before the signal VM WL high active is raised). Similarly, a signal NVM Pre low active in a lower state may indicate precharging of the bit line of the non-volatile memory bit cell (e.g., before the signal NVM WL high active is raised). As shown, the leading edge of "Pre1" may precede the leading edge of the voltage signal on the word line. The memory state or stored value detected in the read operation may then be copied or transferred to one or more bit cells in a write operation to one or more of the non-volatile memory bit cells 304. Similar read and write operations may occur in a pipelined manner as shown in subsequent clock cycles. Thus, it may be observed that transferring a state from a volatile memory bit cell (e.g., volatile memory bit cell 302) to a non-volatile memory bit cell (e.g., non-volatile memory bit cell 304) may involve latency.

[0046] You can use the following Figure 5 , Figure 6 , Figure 8 and Fig.12 The features shown in Figure 3A and Figure 3B Various aspects of the integrated circuit device 300 are shown in FIG. Figure 5 is a schematic diagram of a memory such as memory 108 including a volatile memory bit cell array 504 and a non-volatile memory bit cell array 506. Word lines 508 can be used to access bit cells in the non-volatile memory bit cell array 506 and word lines 520 can be used to access bit cells in the volatile memory bit cell array 504 for read and write operations. In this context, a "word line" includes a conductor used to send a signal to select a specific bit cell or group of bit cells to be accessed in a read operation or a write operation. In a specific example embodiment, the voltage of the signal on the word line can be raised or lowered to select or deselect a specific bit cell or group of bit cells to be connected to the corresponding bit line or group of bit lines during a read or write operation. However, it should be understood that this is merely an example of a word line and the claimed subject matter is not limited in this regard.

[0047] In order to enable copying of memory states (or transfer of corresponding stored values) between volatile memory bit cells in volatile memory bit cell array 504 and non-volatile memory bit cells 506, a particular word line 508 may be used to access one or more bit cells in non-volatile memory bit cell array 506 and a corresponding word line 520 may be used to access one or more bit cells in volatile memory bit cell array 504.

[0048] In one embodiment, decoder circuit 510 may include a "shared decoder circuit" because, among other things, decoder circuit 510 may assert a voltage signal on word line 520 to access a bit cell in volatile memory bit cell array 504 and assert a voltage signal on word line 508 to access a bit cell in non-volatile memory bit cell array 506. In an alternative embodiment, decoder circuit 510 may assert a voltage signal on word line 520 to access a bit cell in volatile memory bit cell array 504, and a second optional decoder 512 may assert a voltage signal on word line 508 to access a bit cell in non-volatile memory bit cell array 506. According to an embodiment, assertion of word line 520 may connect a corresponding bit cell in volatile memory 504 to a bit line (not shown) that is in turn connected to I / O circuit 516. I / O circuit 516 may include a sense amplifier circuit (not shown) for detecting the memory state of the volatile memory bit cell connected to the bit line in a read operation. The I / O circuit 516 may also include a write driver circuit (not shown) to generate a programming signal to affect the memory state of a volatile memory bit cell connected to a bit line in a write operation. Similarly, the I / O circuit 518 may include a sense amplifier circuit (not shown) for detecting the memory state of a non-volatile memory bit cell connected to a bit line in a read operation. The I / O circuit 518 may also include a write driver circuit (not shown) to generate a programming signal to affect the memory state of a non-volatile memory bit cell connected to a bit line in a write operation. In this context, a "bit line" includes a conductor that can be connected to at least a portion of a bit cell circuit during a write operation to send a signal to change the memory state of the bit cell circuit, or send a signal representing the memory state of the bit cell circuit during a read operation. According to an embodiment, a bus 550 coupled between the I / O circuits 516 and 518 can facilitate copying memory states (or transferring corresponding stored values) between one or more volatile memory bit cells 504 and non-volatile memory bit cells 506. For example, bus 550 may include an address portion that identifies a target bit cell to be affected by a write operation and a data portion (e.g., having a data bus width) that is used to send one or more signals indicating a memory state obtained in a read operation to be written to the target bit cell in the write operation.

[0049] In a particular embodiment, the data bus width and word address width used to access the bit cells in the volatile memory bit cell array 504 (e.g., at the I / O circuit 516) can be the same as the data bus width and word address width used to access the bit cells in the non-volatile memory bit cell array 506 (e.g., at the I / O circuit 518). However, the corresponding word lines 520 and 508 can be connected through the buffer 514 formed between the respective portions of the bit cells in the volatile memory bit cell array 504 and the respective portions of the bit cells in the non-volatile memory bit cell array 506. In an embodiment, in order to enable the decoder circuit 510 to operate as a shared decoder circuit as discussed above, the buffer 514 can reshape the voltage signal generated by the decoder 510 on the corresponding word line 520 to be applied to the word line 508 for accessing the bit cell in the non-volatile memory array 506. Here, for example, an access signal may be applied to word line 508 in response to decoder 510 applying the access signal to word line 520 coupled to bit line 508 through buffer 514. In an alternative embodiment, optional decoder 512 may generate signals on word line 508 for accessing bit cells in non-volatile memory bit array 506, whereas decoder 510 generates signals on word line 520 for accessing bit cells in volatile memory bit cell array 504. Buffer 514 may also perform a latching function to enable pipelining of operations between bit cells of volatile memory bit cell array 504 and bit cells of non-volatile memory bit cell array 506. For example, buffer 514 may affect word line signals to enable pipelining of read and write operations to copy memory states between bit cells of volatile memory bit cell array 504 and bit cells of non-volatile memory bit cell array 506.

[0050] As discussed above, in one embodiment, memory states may be copied (or corresponding stored values ​​may be transferred) between bit cells of the volatile memory bit cell array 504 and bit cells of the non-volatile memory bit cell array 506. In a particular implementation, the states of bit cells in the volatile memory bit cell array 504 coupled to a particular word line 520 may be copied or written to bit cells in the non-volatile memory bit cell array 506 coupled to a particular word line 508 (coupled to the particular word line 520 via the buffer 514). Fig. 7A508 to be applied to a bit cell in the nonvolatile memory bit cell array 506 at the trailing edge of the clock pulse in the first clock cycle. Here, a signal R1 may be asserted on a word line 520 to be applied to a selected bit cell in the nonvolatile memory bit cell array 504 for a read operation after the leading edge of a clock pulse in a first clock cycle. During a read operation, a circuit in the I / O circuit 516 connected to the selected bit cell via a bit line may detect the memory state of the selected bit cell. After the state in the selected bit cell in the array 504 is detected from the read operation, a signal W1 may then be asserted on a word line 508 to be applied to a bit cell in the nonvolatile memory bit cell array 506 at the trailing edge of the clock pulse in the first clock cycle. Here, a bit cell in the nonvolatile memory bit cell array 506 may be accessed for a write operation to write or copy the detected state of the selected bit cell in the volatile memory bit cell array 504 to the accessed bit cell in the nonvolatile memory bit cell array 506. As noted above, the buffer 514 in the word line 508 can provide a latch at the boundary between the bit cell of the word line 508 coupled to the array 504 being accessed for a read operation and the bit cell of the word line 508 coupled to the array 506 being accessed for a write operation. This can allow a falling edge of the voltage on the word line 520 coupled to the array 504 and the Pre signal (the bit line that restores VM) can also fall. In other embodiments, the buffer 514 can include a level shifter circuit or a latch circuit to raise or lower the voltage on the associated word line. For example, the buffer 514 including a level shifter circuit can respond to a first voltage on the word line 520 (enabling access to a selected volatile memory bit cell of the array 504) by applying a second different voltage on the word line 508 (enabling access to a selected non-volatile memory bit cell of the array 506).

[0051] If you can Fig. 7A It is observed that the leading edge of the word line signal W1 for accessing a bit cell in the nonvolatile memory bit cell array 506 may occur at the trailing edge of the clock pulse in the immediately preceding clock cycle. The VM Pre active low signal may be asserted low after the word line signal R1 is deasserted, or returned low immediately following a write operation that occurred during the assertion of the word line voltage signal R1.

[0052] As above combined Figure 3A and Figure 3BAs noted, the volatile memory bus 326 and the non-volatile memory bus 328 may have different bus widths (e.g., where one of the bus widths is an integer multiple of the other bus width). Similarly, the bus 550 may include a bus width for accessing the volatile memory array 504 that is different from the bus width for accessing the non-volatile memory array 506. In a particular numerical example implementation, the non-volatile memory array 506 may be accessed using a bus width of 64 bits, while the volatile memory array 504 may be accessed using a bus width of 16 bits, such that 64 bits may be transferred between the non-volatile memory array 506 and the volatile memory array 504 over four memory cycles of the volatile memory array 504 and a single memory cycle of the non-volatile memory array 506. It should be understood that this is merely a particular numerical example, and different widths may be implemented (e.g., in the case of non-integer widths, the bus width for accessing the volatile memory array 504 is wider than the bus width for accessing the non-volatile memory array 506). In this particular example, the I / O circuit 516 may include a 4×1 column multiplexer (not shown) to facilitate the transfer of four 16-bit words of the volatile memory array 504 between a single 64-bit word of the non-volatile memory array 506. Similarly, data may be present in a port (not shown) in the I / O circuit 518. Such data in the port may have a width of 16 bits, or may have a width of 64 bits. To support such 64-bit transfers in a single transaction of the non-volatile memory array 506 and four transactions of the volatile memory array 504, the corresponding buffer 514 may maintain or latch a particular access signal on the word line 508 in an active state (e.g., at a constant voltage) while the access signal on the word line 520 cycles between precharge and active phases for four transactions.

[0053] To support the transfer of storage values ​​between the volatile memory bit cells 504 and the nonvolatile memory bit cells 506 in the event that the bus width for accessing the nonvolatile memory bit cells 506 is four times the bus width for accessing the volatile memory bit cells 504, the buffer 514 may further include a latch to maintain an access signal on the decoded word line 508 for the selected nonvolatile memory bit cells 506. For example, such an access signal may be maintained on the decoded word line 508 until data including a full nonvolatile bus width has been read from the selected volatile memory bit cells 504. In the above example where the bus width for accessing the volatile memory bit cells 504 is 16 bits and the bus width for accessing the nonvolatile memory bit cells 506 is 64 bits, sixty-four bits of data may be stored in the volatile memory bit cells 504 on the same word line 520 or row (and therefore the decoder address may not change), but across a column address width of 4 bits. If the column address of the volatile memory bit cell 504 is to be decoded according to an eight-to-one encoding scheme, the column multiplexer can select one of the eight columns while the word line 520 is selected. In three subsequent accesses to the volatile memory bit cell 504 on the selected word line 520, the same row address can be accessed while the column address can cycle from bit 0 to bit 1...ending on bit 3. In this way, access to the volatile memory bit cell 504 can occur in four access cycles while using the same row address. For each such access cycle, 16 bits can be read or written via the bus 550. Although the row address can remain unchanged, the buffer 514 (which can include a latch as discussed above) can maintain the value of the decoded word line 508 so that the row is selected for the selected volatile memory bit cell 506.

[0054] In one embodiment for accessing the selected non-volatile memory bit cell 506 in a write operation in conjunction with four corresponding cycles for accessing the selected volatile memory bit cell 504, the selected non-volatile memory bit cell 506 may be accessed by the write operation in four different cycles. In an alternative embodiment, a write buffer in the I / O circuit 516 (not shown) may accumulate the full 64 bits before enabling the word line 506 for the write operation applied to the selected non-volatile memory bit cell 506. The control for this word line 508 may include a combination of a self-timing signal and a clock signal depending on the specific implementation. Thus, the features of the buffer 514 may support the case where the bus width for accessing the volatile memory bit cell 504 and the bus width for accessing the non-volatile memory bit cell 506 are not equal.

[0055] exist Figure 5 and Figure 6In a particular embodiment of the present invention, a volatile memory bit cell is shown adjacent to a decoder circuit and a non-volatile memory bit cell is shown adjacent to (or on both sides of) a decoder and a non-volatile memory bit cell. Figure 5 and Figure 6 In an alternative embodiment of the embodiment of the present invention, the non-volatile memory bit cell can be formed adjacent to (and on both sides of) both the decoder circuit and the volatile memory bit cell. Here, the buffer circuit can also be used to reshape the access signal applied to the word line connected to the volatile memory bit cell (e.g., instead of reshaping the access signal applied to the word line connected to the non-volatile memory bit cell).

[0056] Figure 6 6 is a schematic diagram of an alternative embodiment in which a word line 608 may be used to access bit cells in both the volatile memory bit cell array 604 and the non-volatile memory bit cell array 606 for read and write operations. For example, a voltage on a single word line may be asserted by a decoder 510 to access both the non-volatile memory bit cell 506 and the volatile memory bit cell 504 (e.g., via corresponding buffers 514). Figure 5 As in the embodiment of the present invention, the decoder 610 may include a "shared decoder circuit" to assert a voltage signal on a single word line 608 to access a bit cell in the volatile memory bit cell array 604 and the non-volatile memory bit cell array 606. In addition, the decoder 610 may use the same word address bus to access the bit cells in the volatile memory bit cell array 604 and the non-volatile memory bit cell array 606. The bus 650 coupled between the I / O circuits 616 and 618 can facilitate copying memory states (or transferring corresponding stored values) between one or more volatile memory bit cells 604 and the non-volatile memory bit cells 606. For example, the bus 650 may include: an address portion that identifies a target bit cell to be affected by a write operation; and a data portion (e.g., having a data bus width) that is used to send one or more signals indicating a memory state obtained in a read operation to be written to the target bit cell in the write operation.

[0057] In an embodiment, the data bus width of bus 650 between I / O circuit 616 and I / O circuit 618 may be the same. Alternatively, the data bus width at I / O circuit 616 may be an integer multiple of the bus width at I / O circuit 618, or the data bus width at I / O circuit 618 may be an integer multiple of the bus width at I / O circuit. For simplicity of this discussion, it is assumed that the data bus widths at I / O circuit 616 and I / O circuit 618 are the same (e.g., the same number of bits or bytes). However, it should be understood that the data bus widths at I / O circuit 616 and I / O circuit 618 may be different without departing from the claimed subject matter. For example, if I / O circuit 616 has a data bus width that is an integer multiple of the data bus width of I / O circuit 618, then transferring values ​​between I / O circuit 616 and I / O circuit 618 may require a single access cycle for I / O circuit 616 and an integer multiple of access cycles for I / O circuit 618. However, in other implementations, the use of read or write masks may enable a bus width for I / O circuitry 616 that is not necessarily an integer multiple of the bus width for I / O circuitry 618 .

[0058] exist Figure 6 In a particular embodiment of the present invention, bit cells in the volatile memory bit cell array 604 and bit cells in the non-volatile memory bit cell array 606 can be accessed simultaneously by asserting a single word line 608. In other words, a single word line 608 can be used to access corresponding bit cells in both arrays 604 and 606 without a buffer (e.g., buffer 514) connecting a first bit line to access a bit cell in array 604 and connecting a second bit line to access a bit cell in array 606. Using the same word line decoding scheme to access bit cells in both the volatile memory bit cell array 604 and the non-volatile memory bit cell array 606, Figure 6 Embodiments of may enable tighter coupling between bit cells in array 604 and bit cells in array 606. In certain implementations, Figure 5 or Figure 6 Embodiments of the invention may enable a read-modify-write implementation that would potentially reduce latency in copying memory states between non-volatile memory bit cells (e.g., of arrays 504 or 604) and volatile memory bit cells (e.g., of arrays 506 and 606).

[0059] In some embodiments, copying or transferring a certain amount of multiple memory states or stored values ​​(e.g., a "packet" such as a byte or word) between a non-volatile memory bit cell and a volatile memory bit cell may require a minimum of two memory cycles (e.g., two memory clock cycles). For example, a read operation performed on a portion of a first memory to detect a memory state may consume a first memory cycle and a subsequent operation to write the detected memory state to a portion of a second memory may consume a second memory cycle. Accessing the first memory to be read and the second memory to be written may require additional time to restore the bit line voltage.

[0060] Use Figure 6 , it is possible to assert a voltage signal on the shared word line 608 to perform a read operation to detect a memory state in the first array, and then continuously maintain the voltage signal while performing the operation to write the detected memory state to the second array. Here, while asserting the shared word line 608 to access the bit cells in the volatile memory bit cell array 604 and the bit cells in the non-volatile memory bit cell array 618, the I / O circuit 616 and the I / O circuit 618 can perform read and write operations in the same memory cycle. If the memory state is copied from the selected bit cell in the volatile memory bit cell array 604 to the selected bit cell in the non-volatile memory bit cell array 606, the sense amplifier of the I / O circuit 616 can detect the memory state of the selected bit cell in the volatile memory bit cell array 604 in the same memory clock cycle, and the write driver circuit of the I / O circuit 618 can apply a programming signal to the selected bit cell of the non-volatile memory bit cell array 606 to write the detected memory state. Similarly, if the memory state or stored value is copied or transferred from a selected bit cell in the nonvolatile memory bit cell array 606 to a selected bit cell in the volatile memory bit cell array 604, the bus 650 in the same memory clock cycle sense amplifier of the I / O circuit 618 can detect the memory state of the selected bit cell in the nonvolatile memory bit cell array 606, and the write driver circuit of the I / O circuit 616 can apply a programming signal to the selected bit cell of the volatile memory bit cell array 604 to write a value corresponding to the detected value.

[0061] In accordance with an embodiment, the voltages on the bit lines connecting the bit cells in the volatile memory bit cell array 604 to the I / O circuit 616 and the bit cells in the non-volatile memory bit cell array 606 to the I / O circuit 618 may be restored after a read or write operation. As noted above in the embodiments, the read and write operations for copying memory states between the volatile memory bit cells 604 and the non-volatile memory bit cells may occur in a single clock cycle. Thus, this particular two-part access procedure (a read operation for detecting the memory state of a selected bit cell and a write operation for placing the selected bit cell in the detected memory state) may be performed in a single memory access cycle. Thus, the voltages on the bit lines connecting the affected bit cells and the I / O circuits 616 and 618 may be restored in a single period after a single clock cycle to perform the read and write operations. In the illustrated specific example of Figure 7B In the timing diagram of , a read operation for detecting a memory state of a first memory may occur in a cycle time that is 1.5 times the cycle time of a memory cycle for accessing a non-volatile memory bit cell in array 604, followed by an operation for writing the detected memory state to a second memory (e.g., assuming that the duration of the read operation for detecting the memory state in a bit cell of array 604 and the duration of a subsequent operation for writing the detected memory state to a bit cell in array 606 are approximately the same). Here, Figure 7B The voltage signal on the word line connected to the nonvolatile memory bit cell and the volatile memory bit cell is shown to rise after the leading edge of the first clock pulse and fall after the leading edge of the next clock pulse. While the voltage signal on the word line is rising, a read operation R1 detects the memory state of the volatile memory bit cell and a subsequent write operation W1 can write a value corresponding to the detected state to the nonvolatile memory bit cell. The bit line voltage for the read operation R1 and the write operation W1 can then be restored in a single memory access cycle after the voltage signal on the word line is lowered. In other embodiments, the duration for accessing a bit cell in a nonvolatile memory array can be longer than the duration for accessing a bit cell in a volatile memory array. Therefore, Fig. 7A The specific implementation of may reflect the trade-off between slower cycle time for smaller area and power.

[0062] Figure 6The particular embodiment of 610 shows that the volatile memory bit cell 604 is configured between the decoder 610 and the non-volatile memory bit cell 606, wherein the decoder 610 and the non-volatile memory bit cell 606 are located on either side of the volatile memory bit cell 604. In an alternative embodiment, the decoder 610 can be configured to be placed in a "butterfly configuration" between two physically separate portions of the volatile memory bit cell 604 located on either side of the decoder 610 and two physically separate portions of the non-volatile memory bit cell 606 located on either side of the physically separate portions of the volatile memory bit cell 604. Here, the decoder circuit 610 can apply access signals directly to the word lines connected to the selected bit cells in any one of the separate portions of the volatile memory bit cells 604. In addition, the access signals applied to the word lines connected to the specific separate portions of the volatile memory bit cells 604 can access the bit cells in the specific portion of the non-volatile memory bit cells 606.

[0063] In another alternative embodiment (also not shown), the volatile memory bit cell 604, the non-volatile memory bit cell 606, the decoder 610, the I / O circuit 616, and the I / O circuit 618 can be further configured in a "four-quadrant" butterfly configuration. Here, in the aforementioned butterfly configuration, each physically separate portion of the volatile memory bit cell 604 can be further divided or bifurcated into two additional portions, such that a portion of the I / O circuit 616 can be formed between the two portions of the physically separate portion of the volatile memory bit cell 604. Similarly, each physically separate portion of the non-volatile memory bit cell 606 can be further divided or bifurcated into two additional portions, such that a portion of the I / O circuit 618 can be formed between the two portions of the physically separate portion of the non-volatile memory bit cell 606.

[0064] In yet another alternative embodiment (also not shown), the decoder 610 may be configured between the volatile memory bit cells 604 and the non-volatile memory bit cells 606 such that the volatile memory bit cells 604 and the non-volatile memory bit cells 606 are located on either side of the decoder 610 in an "asymmetric butterfly configuration." Here, the decoder circuit 610 may apply a signal to a selected bit cell connected to the volatile memory access bit cell 604 or directly to a selected non-volatile memory bit cell 606.

[0065] Figure 8 8 is a schematic diagram of a memory system integrating volatile memory bit cells and non-volatile memory bit cells according to an embodiment. Here, the volatile memory bit slice 804 may include one or more volatile memory bit cells interleaved with the non-volatile memory bit cells of the non-volatile memory bit slice 806. In an embodiment, the above-mentioned Figure 5 , Figure 6 , Fig. 7A and Figure 7B As discussed, corresponding volatile memory bit cells of the volatile memory bit slice 804 and adjacent or local nonvolatile memory bit cells of the nonvolatile memory bit slice 806 are accessed in response to a single voltage asserted on a word line. For example, memory states or stored values ​​in one or more bit cells of the volatile memory bit slice 804 may be copied or transferred to one or more bit cells of the nonvolatile memory bit slice 806 based on a read operation performed at the I / O circuitry 816 to detect memory states of one or more bit cells of the volatile memory bit cell array 804 and a write operation performed by the I / O circuitry 818 to write values ​​corresponding to the one or more detected memory states.

[0066] In some embodiments, a stored value is transferred between a volatile memory bit cell and a non-volatile memory bit cell (e.g., as described above in conjunction with Figure 5 and Figure 6 discussed) can increase or stress the memory bus requirements. On the other hand, Figure 8 Certain embodiments of the present invention may enable the transfer of stored values ​​between adjacently formed volatile and non-volatile memory bit cells without accessing a memory bus coupled between a uniform volatile memory bit cell array (e.g., volatile memory array 504 or 604) and a uniform non-volatile memory bit cell array (e.g., non-volatile memory array 506 or 606). Figure 8 As shown in , stored values ​​may be transferred between one or more bit cells of a bit slice of volatile memory bit cells 804 and one or more bit cells of a bit slice of non-volatile memory bit cells 806 .

[0067] In one embodiment, the I / O circuitry 816 may include read circuitry connected to one or more common first bit lines to detect values ​​stored in selected volatile memory bit cells of the volatile memory bit slice 804, and the I / O circuitry 818 may include write driver circuitry connected to one or more common second bit lines to generate programming signals to store values ​​corresponding to the detected values ​​in selected nonvolatile memory bit cells of the nonvolatile memory bit slice 806. Here, values ​​stored in one or more volatile memory bit cells of the volatile memory bit slice 804 may be transferred to one or more nonvolatile memory bit cells of the nonvolatile memory bit slice 806 using simple read-modify-write transactions and without accessing external bus devices. Similarly, I / O circuitry 818 may include read circuitry connected to one or more common first bit lines to detect values ​​stored in selected nonvolatile memory bit cells of nonvolatile memory bit slice 806, and I / O circuitry 816 may include write driver circuitry connected to one or more common second bit lines to generate programming signals to store values ​​corresponding to the detected values ​​in selected volatile memory bit cells of volatile memory bit slice 804. Here, values ​​stored in one or more volatile memory bit cells of nonvolatile memory bit slice 806 may be transferred to one or more nonvolatile memory bit cells of volatile memory bit slice 804 using simple read-modify-write transactions without accessing external bus devices.

[0068] In one embodiment, a bit slice of volatile memory bit cells 804 or a bit slice of non-volatile memory bit cells 806 may include a single bit "column slice" of bit cells connected to a single bit line and selectable by a word line, or multiple bits coupled to multiple corresponding bit lines selectable by a single word line. In one embodiment where a bit slice 804 or a bit slice 806 includes multiple bit cells accessible by a single word line, a multiplexer (not shown) may be used to connect the bit line of a selected bit cell (e.g., selected among multiple bit cells accessible by a word line) to a read circuit or a write circuit to enable transfer of a stored value between the selected bit cell and another bit cell as discussed above. In a particular embodiment, a first multiplexer may facilitate transfer of a stored value from a bit slice 804 to a bit slice 806, and a second multiplexer may facilitate transfer of a stored value from a bit slice 806 to a bit slice 804. For example, if the number of bit lines connected to a bit slice 804 is not equal to the number of bit lines connected to a bit slice 806, the widths of the first multiplexer and the second multiplexer may not necessarily be equal.

[0069] Fig. 9is a schematic diagram illustrating an addressing scheme for integrated volatile memory bit cells and non-volatile memory bit cells according to an embodiment. Such an addressing scheme may be implemented, for example, in a decoder circuit (e.g., decoder circuit 510 or decoder 610) for selecting a bit cell to be accessed for a read or write operation. In an embodiment, the physical address A may range from a value n to o. For example, an addressable portion (e.g., a fixed length word) in a volatile memory bit cell array (e.g., volatile memory bit cell array 504 or 604) may be accessed according to an address in a range of values ​​n to m. Similarly, an addressable portion in a non-volatile memory bit cell array (e.g., non-volatile memory bit cell array 506 or 606) may be accessed according to an address in a range of values ​​from m to o or more.

[0070] Depending on the embodiment, a physical address A ranging from a value n to o may represent an encoded address or a decoded address. For example, if the encoded address bus is written as A<0:5>, the six bits may be <0> , <1> ... <5> Decoded or expanded into 64 signals representing an address, such as WL<0:63>. In a particular embodiment, the address space WL<0:63> can be divided into volatile memory and non-volatile memory components. For example, WL<0:15> can be dedicated to accessing non-volatile memory components, while WL<16:63> can be dedicated to accessing volatile memory components.

[0071] In another embodiment where a physical address a ranging from a value n to o represents a decoded address, overlapping portions of the bus may be used to access either volatile memory or non-volatile memory. For example, address WL<0:31> may access NVM, whereas address WL<16:63> may access VM. Address WL<16:31>, when accessed, may be used to read from / write to both VM and NVM.

[0072] Figure 5 and Figure 6 Embodiments of the present invention relate to coupling of a volatile memory bit cell array and a non-volatile memory bit cell array to facilitate copying memory states (or transferring stored values) between the volatile memory bit cell array and the non-volatile memory bit cell array. According to an embodiment, Fig.101 is a schematic diagram of integrating only a portion of a volatile memory bit cell array 1004 with a non-volatile memory bit cell array 1006. In this particular embodiment, the volatile memory bit cell array 1004 can be formed as a cache memory in a computing platform that is partitioned into cache "ways" Way0, Way1, Way2, and Way3. As shown, the non-volatile memory bit cell array 1006 can be coupled to Way0 to facilitate copying memory states between the non-volatile memory bit cells 1006 and Way0. In the example shown in FIG. Fig.11 In the alternative implementation shown in , only a portion of the nonvolatile memory bitcell array 1106 is coupled to the nonvolatile memory bitcell array 1104 .

[0073] exist Fig.10 and Fig.11 In certain embodiments, the addresses used to access volatile or nonvolatile memory bit cells may be subsets of each other. Fig.10 As shown in , the addresses used to access volatile memory 1004 (e.g., for portions of the cache) may be a subset of the addresses used to access portions of non-volatile memory 1006. Similarly, as shown in Fig.11 As shown in , the addresses used to access non-volatile memory 1106 may be a subset of the addresses used to access portions of volatile memory 1104 .

[0074] Combination of the above Figure 5 and Figure 6 The discussed embodiments involve copying memory states between non-volatile memory bit cells and volatile memory bit cells by asserting a single access signal on a word line to access both the non-volatile memory bit cells and the volatile memory bit cells (e.g., by asserting a single access signal on a word line 608 connected to a selected bit cell in the volatile memory array 604 and a selected bit cell in the non-volatile memory array 606, or by asserting a single access signal on a word line 504 to access a selected bit cell of the volatile memory array 504 such that the corresponding word line 508 is asserted through a buffer 514 to access the selected bit cell of the non-volatile memory array 506). Fig.12 1 is a schematic diagram illustrating the interleaving of volatile memory bit cells and non-volatile memory bit cells according to an embodiment. In particular, non-volatile memory bit cell array 1206 is coupled to two volatile memory bit cell arrays 1204 and 1224. According to an embodiment, memory states can be copied (or corresponding stored values ​​transferred) between non-volatile memory bit cell array 1206 and volatile memory bit cell array 1204 or volatile memory bit cell array 1224.

[0075] In one implementation, the bit cells in the volatile memory bit cell array 1204 may be accessed for a read operation in response to a voltage signal applied to a first word line 1220 causing a memory state to be detected by the I / O circuit 1216. The bit cells in the non-volatile memory bit cell array 1206 may be subsequently accessed for a write operation in response to a second word line signal applied to the word line 1208 to copy the memory state detected by the I / O circuit 1216. For example, the word line 1208 may be coupled to the word line 1220 through at least a tri-state buffer 1230 and possibly optional flip-flop circuit 1234. Similar actions may be performed to copy a memory state detected in a bit cell of volatile memory bit cell array 1224 to a bit cell in nonvolatile memory bit cell array 1206, facilitated by tri-state buffer 1232 and possibly flip-flop circuit 1236, or to copy a memory state detected in a bit cell of nonvolatile memory bit cell array 1206 to a bit cell of volatile memory bit cell array 1204 or 1224.

[0076] According to an embodiment, tri-state buffer 1230 may isolate volatile memory bit cell array 1204 while copying memory states between non-volatile memory bit cell array 1206 and volatile memory bit cell array 1224. For example, to copy memory states between bit cells in non-volatile memory bit cell array 1206 and volatile memory bit cell array 1224 during read and write operations, tri-state buffer 1230 may disconnect word line 1220 from word line 1208 and tri-state buffer 1232 may connect selected word lines 1208 and 1222. Similarly, tri-state buffer 1232 may isolate volatile memory bit cell array 1224 while copying memory states between non-volatile memory bit cell array 1206 and volatile memory bit cell array 1204. For example, in order to copy memory states between bit cells in non-volatile memory bit cell array 1206 and volatile memory bit cell array 1204 during read and write operations, tri-state buffer 1230 may disconnect word line 1220 from word line 1208 and tri-state buffer 1232 may connect selected word lines 1208 and 1222.

[0077] As pointed out above, Fig.12Embodiments of the present invention may include optional flip-flop circuits 1234 and 1236 to enable pipelining of operations to copy memory states between bit cells in the nonvolatile memory bit cell array 1206 and bit cells in the volatile memory bit cell array 1204 or 1224. For example, during a read operation, to detect the memory state of a selected bit cell in the bit cell array 1204, a rising edge voltage on the word line 1220 may couple the selected bit cell to the I / O circuit 1216. In response to the rising edge voltage on the word line 1220, the flip-flop circuit 1234 may apply a voltage signal on the word line 1208 to decouple the corresponding bit cell in the nonvolatile memory bit cell array 1206 from the I / O circuit 1218. After the read operation is completed, the falling edge of the voltage on the word line 1220 applied to the flip-flop circuit 1234 may change the voltage on the word line 1208 to couple the corresponding bit cell in the nonvolatile memory bit cell array 1206 from the I / O circuit 1218 for a write operation.

[0078] According to an embodiment, tri-state buffers 1230 and 1232 may additionally include level shifters to enable different word line voltages to be applied to access volatile memory (on word line 1220 or 1222) and to access non-volatile memory on word line 1208. In addition, the non-volatile memory bit cell of 1206 may be accessed by a signal originating from decoder 1210 or decoder 1212. For example, if a value is transferred between volatile memory bit cell 1204 and non-volatile memory bit cell 1206, the voltage on word line 1208 may be affected to access non-volatile memory bit cell 1206 in response to decoder circuit 1210. Likewise, if a value is transferred between volatile memory bit cell 1224 and non-volatile memory bit cell 1206, the voltage on word line 1208 may be affected to access non-volatile memory bit cell 1206 in response to decoder circuit 1212, for example.

[0079] Fig.13A 1 is a schematic diagram of a bit cell circuit 1300 including a volatile memory element and a nonvolatile memory element according to an embodiment. The bit cell circuit 1300 includes two nonvolatile memory elements NV 1 and NV 2 and two volatile memory elements partially formed by PFETs P1 and P2. Nonvolatile memory element NV 1 and NV 2The bit cell circuit 1300 may be formed using any of a number of different types of non-volatile memory devices including, for example, flash memory devices, correlated electronic memory devices, phase change memory (PCM) devices, magnetic memory devices, to provide just a few examples. Depending on whether the bit cell circuit 1300 is operating in a volatile memory mode or a non-volatile memory device mode, a voltage source 1306 is coupled to the non-volatile memory element NV 1 and NV 2 and first terminals of PFETs P1 and P2. Multiplexers 1302 and 1304 may select nonvolatile memory elements NV in response to the state of a select signal SEL. 1 and NV 2 The second terminal of or the second terminal of PFET P1 and P2 is selectively coupled to the latch nodes T and C. In one example, the selection signal SEL may include a single signal input. In another example, the selection signal SEL may include multiple signal inputs on a bus. In a particular embodiment, in order to enable an appropriate transition to a volatile storage mode, multiplexers 1302 and 1304 may enable the connection of latch nodes T and C from non-volatile memory element NV 1 and NV 2 to the second terminals of PFETs P1 and P2, so that the latch nodes T and C are connected to the nonvolatile memory element NV 1 and NV 2 Prior to disconnection, latch nodes T and C may be connected to second terminals of PFETs P1 and P2.

[0080] In one embodiment, the bit cell circuit 1300 may be employed in a device that transitions between a powered-on state and a powered-down state. Here, it may be desirable to save a particular memory state of a volatile memory element as the device transitions to the powered-down state so that the particular memory state can be restored at a future time when the device transitions back to the powered-down state. For example, as the device transitions to the powered-down state, it may be desirable to copy the current state of a nonvolatile memory element to a nonvolatile memory element NV 1 and NV 2 This may include, for example, a read operation for detecting the state of a non-volatile memory element, followed by a read operation for writing the detected state to the non-volatile memory element NV. 1 and NV 2 operation.

[0081] In another example, as the device transitions from a powered-down state to a powered-up state, it may be desirable to store the nonvolatile memory element NV 1 and NV 2The current state of the nonvolatile memory element NV may be copied or transferred to the nonvolatile memory element. This may include, for example, detecting the current state of the nonvolatile memory element NV 1 and NV 2 A read operation to detect the state of a nonvolatile memory element is followed by a write operation to store the detected state to the nonvolatile memory element.

[0082] According to an embodiment, the bit cell circuit 1300 may store the data stored in the nonvolatile memory element NV during a power-on operation (eg, a power-on reset). 1 and NV 2 The value in is transferred or copied to the nonvolatile memory element. For example, this may occur while FETs N1 and N2 are turned off (e.g., when signal WL is low), causing the bit cell circuit to be disconnected from bit lines BL and BL'. In a particular embodiment, the nonvolatile memory element NV 1 and NV 2 A value, parameter, condition or symbol can be stored as a complementary resistance state or impedance state. In other words, the non-volatile memory element NV 1 and NV 2 You can use NV 1 is in a high impedance / resistance state and enables NV 2 is in a low impedance / resistance state to store a first value, parameter, condition, or symbol, and can be 1 is in a low impedance / resistance state and enables NV 2 is in a high impedance / resistance state to store a second value, parameter, condition, or symbol. When power is applied at voltage source 1306, multiplexer 1302 can store NV 1 The terminal of NV is connected to the node T (while disconnecting PFET P1 from the node T) and the multiplexer 1304 can connect NV 2 The terminal of NV is connected to node C (while disconnecting PFET P2 from node C). Multiplexer 1302 can then connect PFET P1 to node T and multiplexer 1304 can connect PFET P2 to node C. If NV 1 is in high impedance / resistance state and NV 2 In a low impedance / resistance state (e.g., to store a first value, parameter, condition, or symbol), node C may be placed at a voltage higher than node T, thereby opening PFET P1 and closing PFET P2 as part of a latch circuit including PFETs P1 and P2 and NFETs N3 and N4. Conversely, if NV 1 is in low impedance / resistance state and NV 2In order to place the nonvolatile memory element NV in a high impedance / resistance state (e.g., to store a second value, parameter, condition, or symbol), the node C may be placed at a voltage lower than the node T, thereby closing the PFET P1 and opening the PFET P2. After connecting the PFETs P1 and P2 to the nodes T and C, respectively, the multiplexer 1302 may cause the nonvolatile memory element NV 1 The node T is disconnected and the multiplexer 1304 can make NV 2 The bit cell circuit 1300 is disconnected from node C, thereby allowing the bit cell circuit 1300 to operate as a volatile memory bit cell. In a particular embodiment, when the bit cell circuit is operating in a non-volatile mode (when the non-volatile memory element NV 1 and NV 2 When connected to nodes T and C respectively), the voltages at nodes T and C can represent the voltages stored in the non-volatile memory element NV 1 and NV 2 Similarly, when the bit cell circuit is operating in volatile mode (when the nonvolatile memory element NV 1 and NV 2 When the voltage at the nodes T and C is disconnected from the nodes T and C, respectively, the voltage at the nodes T and C may represent the value stored in the volatile memory element. Further, the voltage at the nodes T and C may represent the value stored in the non-volatile memory element NV 1 and NV 2 The value transferred between the volatile memory element formed in part by PFETs P1 and P2.

[0083] Fig. 13B is a schematic diagram of a specific implementation of a bit cell circuit 1300 shown as a bit cell circuit 1310, wherein multiplexer 1302 is implemented as NFET N5 and PFET P3, and multiplexer 1304 is implemented as NFET N6 and PFET P4. Here, the bit cell circuit 1310 may be switchable between operating in a volatile storage mode by lowering the voltage of a signal power on reset (POR) and operating in a non-volatile storage mode by raising the voltage of the signal POR.

[0084] exist Fig. 13C In an alternative embodiment, the bit cell circuit 1312 can be Fig. 13B The bit cell circuit 1310 of a specific embodiment removes PFETs P3 and P4 while maintaining NFETs N5 and N6 in response to a signal POR signal applied to the gates of NFETs N5 and N6 to store nonvolatile memory element NV 1 is selectively connected to node T and the nonvolatile memory element NV 21306. The PFET 1304 is selectively connected to the node C. However, the PFET 1304 remains connected to the node T and the PFET 1305 remains connected to the node C. Here, the voltage of the signal POR may be raised immediately after the power is applied at the voltage source 1306. In an embodiment, the voltage of the signal POR may be raised quickly enough and maintained at the raised voltage for a sufficient duration to enable the nonvolatile memory element NV to be charged. 1 and NV 2 The stored value is reliably transferred to initialize the memory state of the bitcell circuit 1312 to operate as a volatile memory bitcell circuit.

[0085] According to an embodiment, Fig.13D As shown in the bit cell circuit 1315 shown in the schematic diagram of the bit cell circuit 1312, the bit cell circuit 1312 can be further modified to store the memory state in a single non-volatile memory element NV 1 Here, the value, symbol, or condition expressed by the memory state maintained by PFETs P1 and P2 and NFETs N3 and N4 can be transferred to the nonvolatile memory element NV 1 (e.g., in a power-off event) to be expressed or represented as a high impedance or insulating state or a low impedance or conductive state. Similarly, the nonvolatile memory element NV 1 The value, sign, or condition expressed by the maintained memory state (expressed or represented as a high impedance or insulating state or a low impedance or conductive state) is transferred to the volatile memory state maintained by PFETs P1 and P2 and NFETs N3 and N4 (e.g., during a power-up event). Specific implementations of bit cell circuit 1315 do not include NFET N6 and NV 2 This enables the bit cell circuit volatile memory elements and non-volatile memory elements to be implemented using fewer components.

[0086] In one embodiment, the bit cell circuit 1315 can initialize the state of the latch circuit formed by the PFETs P1 and P2 and the NFETs N3 and N4 to 1 The nonvolatile memory state maintained is transferred to the volatile memory state maintained by the nonvolatile memory elements PFETs P1 and P2 and NFETs N3 and N4. When power is applied at the voltage source 1306, NFET N5 may be turned on in response to the signal POR to close, thereby transferring the NV 1 The terminal of NV is connected to the node T (while disconnecting PFETP1 from the node T). 1If NV is in a high impedance / resistance state (e.g., to store a first value, parameter, condition, or symbol), node C may be placed at a voltage higher than node T, thereby opening PFET P1 and closing PFET P2 as part of a latch circuit including PFETs P1 and P2 and NFETs N3 and N4. Conversely, if NV 1 In order to place the NFET N5 in a low impedance / resistance state (e.g., to store a second value, parameter, condition, or symbol), the node C may be placed at a voltage lower than the node T, thereby closing the PFET P1 and opening the PFET P2. After connecting the PFETs P1 and P2 to the nodes T and C, respectively, the signal POR may be lowered to open the NFET N5 and disconnect the nonvolatile memory element NV. 1 , thereby allowing the bit cell circuit 1315 to operate as a volatile memory bit cell.

[0087] As noted above, some scenarios in which it may be desirable for a computing platform to transfer values ​​stored in a volatile memory state to a non-volatile memory state. After transferring the values ​​stored in the volatile memory state to the non-volatile memory state, for example, the computing platform may be shut down. Transferring values ​​stored in a volatile memory state to a non-volatile memory state and then returning from the non-volatile memory state to a volatile memory state may allow the computing platform to quickly recover from a state that occurred prior to a power loss event (e.g., without having to load a memory state from an external non-volatile memory device, which may be a much slower operation).

[0088] According to an embodiment, a value stored in a volatile memory state at the bit cell circuit 1300 may be transferred to a nonvolatile memory element NV 1 and NV 2 Similarly, the value stored in the volatile memory state at the bit cell circuit 1315 can be transferred to be stored in the nonvolatile memory element NV 1 In one embodiment, a read-modify-write procedure may be used to transfer the value stored in the volatile memory state at the bit cell circuit 1300 to the nonvolatile memory element NV 1 and NV 2 at (or only at the non-volatile memory element NV 1 Likewise, a read-modify-write procedure may be similarly used to transfer a value stored in a volatile memory state at the bit cell circuit 1315 to a nonvolatile memory element NV 1 The state is stored in non-volatile memory.

[0089] In a particular embodiment of the bit cell circuit 1300, the volatile memory state may be detected in a read operation that includes, for example, connecting the bit cell circuit 1300 to the bit lines BL and BL'. For example, the bit cell circuit 1300 may be accessed for a read operation by raising the voltage of the word line signal WL to close or enable the NFETs N1 and N2. The bit lines BL and BL' may be coupled to a read circuit (not shown) that is capable of detecting the volatile memory state maintained at the bit cell circuit 1300. The detected volatile memory state of the bit cell circuit 1300 may be temporarily stored and then written to the non-volatile memory element NV in a subsequent write operation. 1 and NV 2 . For example, in a subsequent write operation, a write driver circuit (not shown) coupled to the bit lines BL and BL' can apply a programming signal based on the temporarily stored value. For example, the bit cell circuit 1300 can be accessed again for a write operation by raising the voltage of the word line signal WL to close or enable NFETs N1 and N2 to connect the bit lines BL and BL' to the bit cell circuit 1300. For example, when NFETs N1 and N2 are closed, the write driver circuit (not shown) can apply a programming signal to the non-volatile memory element NV 1 and NV 2 The non-volatile memory element NV 1 and NV 2 is placed in a memory state representing the value detected in the previous read operation. 1 and NV 2 In a specific embodiment including a CES element, such a programming signal may be applied to a non-volatile memory element NV 1 and NV 2 Apply a suitable voltage and current density (e.g., as in the following combination Fig.14A and Fig. 14B As described above) to convert the non-volatile memory element NV 1 and NV 2 The volatile memory state stored in the bit cell circuit 1315 can be similarly transferred to the non-volatile memory element NV 1 The non-volatile memory status in.

[0090] In such Fig.13E In an alternative embodiment shown in FIG. 1 , the volatile memory state maintained by PFETs P1 and P2 and NFETs N3 and N4 can be transferred to the nonvolatile memory element NV without accessing the bit cell circuit 1320 by enabling word line designs N1 and N2. 1 and NV 2In other words, the volatile memory state maintained by PFETs P1 and P2 and NFETs N3 and N4 in the bit cell circuit 1320 can be transferred to the non-volatile memory element NV without connecting the bit lines BL and BL' to the bit cell circuit 1320 (e.g., by raising the voltage of the word line signal WL to close NFETs N1 and N2). 1 and NV 2 The complementary write driver circuits 1326 and 1328 may then apply appropriate programming signals to the nonvolatile memory element NV 1 and NV 2 As can be observed, the transmission gate T2 of the write driver circuit 1328 can be controlled based on the voltage at the node T, while the transmission gate T1 of the write driver circuit 1326 can be controlled based on the voltage at the node C. 1 and NV 2 In a specific embodiment formed as related electronic switches, for example, write driver circuits 1326 and 1328 can apply current I set The voltage V set The programming signal V is used to place a specific non-volatile memory element in a low impedance or conductive state. Specific elements indicated by 1322 and 1324 as part of write drivers 1326 and 1328 can be coupled with T1 and T2 to create a combined voltage and current source and enable the programming signal V through multiplexers 1302 and 1304. set / I set (e.g., for placing a nonvolatile memory element in a low impedance or conductive state) or a programming signal V reset / I reset (For example, to place the non-volatile memory element in a high impedance or isolation state). These elements may include supplying appropriate voltages and currents through T1 and T2 to enable the non-volatile memory element NV 1 and NV 2 The voltage sources of T and C are written to the appropriate states to maintain when power 1306 is removed and no power is applied to bitcell 1320 .

[0091] As noted above, the above-described nonvolatile memory bit cells may include bit cells comprising correlated electronic switch (CES) elements incorporating correlated electronic materials (CEMs). In this context, the CES elements may exhibit abrupt conductor / insulator transitions resulting from electron correlations rather than solid-state structural phase transitions (e.g., crystalline / amorphous in phase change memory (PCM) devices as discussed above or filament formation and conduction in resistive RAM devices). In contrast to melting / solidification or filament formation, the abrupt conductor / insulator transitions in the CES elements may be in response to quantum mechanical phenomena. Such quantum mechanical transitions between conductive and insulating states in CEM memory devices may be understood in any of several aspects.

[0092] The quantum mechanical transition of a CES element between an insulating state and a conducting state can be understood as a Mott transition. In a Mott transition, a material can switch from an insulating state to a conducting state if the Mott transition conditions occur. This can be understood by the condition (n C ) 1 / 3 a=0.26 to define the criterion, where n C is the concentration of electrons and "a" is the Bohr radius. If a critical carrier concentration is reached such that the Mott criterion is satisfied, a Mott transition may occur and the state may change from high resistance / capacitance to low resistance / capacitance.

[0093] The Mott transition can be controlled by localization of electrons. With carriers localized, the strong Coulomb interactions between electrons split the energy bands of the material to create an insulator. If the electrons are no longer localized, the weak Coulomb interactions may dominate the band splitting, leaving behind the metal (conducting) bands. This is sometimes described as the "crowded elevator" phenomenon. Although the elevator has only a few people in it, people can move around easily, which is similar to the on-state. When the elevator reaches a certain crowd concentration, on the other hand, passengers can no longer move, which is similar to the insulating state. However, it should be understood that this classical description provided for illustration purposes, like all classical descriptions of quantum phenomena, is only an incomplete analogy, and the subject matter claimed is not limited in this respect.

[0094] In a specific embodiment of various aspects of the present disclosure, a resistive switching integrated circuit memory may include: a resistive switching memory cell, the resistive switching memory cell including a CES element; a write circuit for placing the resistive switching memory cell in a first resistance state or a second resistance state depending on a signal provided to the memory cell, wherein the impedance of the CES element is higher in the second impedance state than in the first impedance state; and a read circuit for sensing the state of the memory cell and providing an electrical signal corresponding to the sensed state of the memory cell. In a specific embodiment, the CES element may switch the resistance state in response to a Mott transition in a majority of the volume of the CES element. The CES element may include a material selected from the group including aluminum, cadmium, chromium, cobalt, copper, gold, iron, manganese, mercury, molybdenum, nickel, palladium, rhenium, ruthenium, silver, tin, titanium, vanadium and zinc (which may be linked to cations such as oxygen or other types of ligands) or combinations thereof.

[0095] In a particular embodiment, the CES element may be formed as a "CEM Random Access Memory (CeRAM)" device. In this context, a CeRAM device includes a material that can utilize a quantum mechanical Mott transition to transition between or among a plurality of predetermined detectable memory states based at least in part on a transition of at least a portion of the material between a conductive state and an insulating state. In this context, "memory state" means a detectable state of a memory device that indicates a value, symbol, parameter, or condition, to provide just a few examples. In one particular embodiment, as described below, the memory state of the memory device may be detected based at least in part on a signal detected at a terminal of the memory device in a read operation. In another particular embodiment, as described below, the memory device may be placed in a particular memory state to represent or store a particular value, symbol, or parameter by applying one or more signals across the terminals of the memory device in a write operation.

[0096] In certain embodiments, the CES element may include a material sandwiched between conductive terminals. The material may be switched between the aforementioned conductive state and the insulating memory state by applying a specific voltage and current between the terminals. As discussed below in certain example embodiments, the material may be switched between the conductive state and the insulating memory state by applying a specific voltage and current between the terminals. reset and current I reset The first programming signal is applied to the terminals of the CES element to place the material of the CES element sandwiched between the conductive terminals in an insulating or high-impedance memory state, or by applying a first programming signal across a voltage V set and current I setA second programming signal is applied to the terminals of the CES element to place the material of the CES element sandwiched between the conductive terminals in a conductive or low impedance memory state. In this context, it should be understood that terms such as "conductive or low impedance" memory state and "insulating or high impedance" memory state are relative terms and are not specific to any particular amount or value of impedance or conductance. For example, when the memory device is in a first memory state, referred to as an insulating or high impedance memory state, the memory device is less conductive (or more insulating) than when the memory device is in a second memory state, referred to as a conductive or low impedance memory state. In addition, as discussed below with respect to specific embodiments, the CES element can be placed in any of two or more different and distinguishable low impedance or conductive states.

[0097] In a particular embodiment, a CeRAM memory cell may include a metal / CEM / metal (M / CEM / M) stack formed on a semiconductor. For example, such an M / CEM / M stack may be formed on a diode. In an example, such a diode may be selected from the group consisting of a junction diode and a Schottky diode. In this context, it should be understood that "metal" means a conductor, i.e., any material that acts like a metal, including, for example, polysilicon or a doped semiconductor.

[0098] Fig.14A A graph showing the relationship between the current density of a CES element and the voltage across the terminals (not shown) according to an embodiment. Based at least in part on the voltage applied to the terminals of the CES element (e.g., in a write operation), the CES element can be placed in a conductive state or an insulating state. For example, applying a voltage V set and current density J set The CES element can be placed in a conductive memory state while applying a voltage V reset and current density J reset The CES element can be placed in an insulating memory state. After the CES element is placed in an insulating state or a conductive state, the CES element can be placed in an insulating memory state or a conductive state by applying a voltage V read (eg, in a read operation) and detecting the current or current density at the terminals of the CeRAM device to detect a particular state of the CES element.

[0099] According to an embodiment, Fig.14AThe CES devices may include any TMOs such as, for example, perovskites, Mott insulators, charge exchange insulators, and Anderson disordered insulators. In particular embodiments, the CES devices may be formed from switching materials such as nickel oxide, cobalt oxide, iron oxide, yttrium oxide, and perovskites such as Cr-doped strontium titanate, lanthanum titanate, and a family of manganates including praseodymium calcium manganate, praseodymium lanthanum manganate, to provide just a few examples. In particular, oxides incorporating elements with incomplete d and f orbital shells may exhibit sufficient resistive switching properties to be used in CES devices. In embodiments, CES devices may be prepared without electroforming. Other embodiments may employ other transition metal compounds without departing from the claimed subject matter. For example, {M(chxn) may be used without departing from the claimed subject matter. 2 Br}Br 2 and other such metal complexes, wherein M may include Pt, Pd, or Ni, and chxn includes 1R,2R-cyclohexanediamine.

[0100] Fig.14A The CES device may include materials that are TMO metal oxide variable resistor materials, but it should be understood that these are exemplary only and are not intended to limit the subject matter claimed. Specific embodiments may also employ other variable resistor materials. Nickel oxide NiO is disclosed as a specific TMO. The NiO materials discussed herein may be doped with extrinsic ligands that may stabilize the variable resistor characteristics. In particular, the NiO variable resistor materials disclosed herein may include NiO (C x ) is indicated by the carbon-containing ligand. Here, one skilled in the art can simply determine the value of x for any specific carbon-containing ligand and any specific combination of carbon-containing ligand and NiO by balancing the valence. In another specific example, NiO doped with an extrinsic ligand can be expressed as NiO(L x ), where L x is the ligand element or compound and x indicates the number of units of the ligand used for one unit of NiO. One skilled in the art can determine the value of x for any specific ligand and any specific combination of ligand and NiO or any other transition metal simply by balancing the valences.

[0101] If sufficient bias is applied (e.g., exceeding the band splitting potential) and the aforementioned Mott condition is met (injected electron holes = electrons in the switching region), the CES element can rapidly switch from the conducting state to the insulating state via the Mott transition. This can occur in Fig.14AAt point 1408 of the curve graph in . At this point, the electrons are no longer shielded and become localized. This correlation can produce a strong electron-electron interaction potential that splits the energy band to form an insulator. When the CES element is still in an insulating state, current can be generated by the transmission of electron holes. If enough bias is applied across the terminals of the CES element, electrons can be injected into the metal-insulator-metal (MIM) diode over the potential barrier of the MIM device. If enough electrons have been injected and enough potential is applied across the terminals to place the CES element in a specific low impedance or conductive state, the increase in electrons can shield the electrons and remove the localization of the electrons, which may collapse the energy band splitting potential that forms the metal.

[0102] According to an embodiment, the current in the CES element may be controlled by an externally applied "consistency" condition determined at least in part based on an external current limited to placing the CES element in a conductive or low impedance state during a write operation. This externally applied consistency current may also set the current density condition for placing the CES element in a high impedance or insulating state for a subsequent reset operation. Fig.14A In the specific embodiment shown in FIG. 14 , the current density J applied during the write operation at point 1416 to place the CES element in a conductive or low impedance state comp A consistent condition can be determined for placing the CES device in a high impedance or isolation state in a subsequent write operation. As shown, the voltage V reset The applied current density J reset ≥J comp to subsequently place the CES device in an isolation or high impedance state.

[0103] The uniformity can thus set a number of electrons in the CES element that will be "trapped" by holes to undergo the Mott transition. In other words, the current applied during a write operation to place the CES element in a conductive memory state can determine the number of holes to be injected into the CES element in order to subsequently transition the CES element to an insulating memory state.

[0104] As noted above, a reset condition may occur in response to a Mott transition at point 1408. As noted above, such a Mott transition may occur under the condition that the electron concentration n in the CES element is equal to the electron hole concentration p. This condition may be modeled according to expression (1) as follows:

[0105]

[0106] in:

[0107] λ TF is the Thomas Fermi screening length; and

[0108] C is a constant.

[0109] According to an embodiment, Fig.14A The current or current density in region 1404 of the graph shown in FIG. 1404 may exist in response to the injection of holes from a voltage signal applied across the terminals of the CES element. Here, as a threshold voltage V is applied across the terminals of the CES element, MI , the hole injection can satisfy the Mott current I MI The transition criterion for the transition from the conducting state to the insulating state is given by equation (2):

[0110]

[0111] Where Q(V MI ) is the injected charge (holes or electrons) and is a function of the applied voltage.

[0112] The injection of electron holes enables the Mott transition between energy bands and in response to the critical voltage V MI and critical current I MI By making the electron concentration n equal to the charge concentration according to expression (1) and by following the formula (2) MI The injected holes induce the Mott transition, and the critical voltage V can be expressed as MI The Thomas Fermi screening length λ TF The dependency of is modeled as follows:

[0113]

[0114] in:

[0115] A CeRam is the cross-sectional area of ​​the CES element; and

[0116] J reset (V MI ) is the threshold voltage V MI The current density applied to the CES element to place the CES element in an insulating state.

[0117] According to an embodiment, a CES element may be placed in a conducting memory state (eg, by transitioning from an insulating memory state) by injecting a sufficient number of electrons to satisfy the Mott transition criterion.

[0118] When the CES is transitioned to the conductive memory state, sufficient electrons have been injected and the potential across the terminals of the CES device overcomes the critical switching potential (e.g., V set), so the injected electrons begin to screen and delocalize the doubly occupied electrons to reverse the disproportionation reaction and close the band gap. The threshold voltage V for enabling the transition to the conductive memory state can be expressed according to expression (4): IM The current density J that makes CES transition to the conductive memory state Set (V IM ) is expressed as follows:

[0119]

[0120] Q(V IM )=qn(V IM )

[0121]

[0122]

[0123] in:

[0124] a B is the Bohr radius.

[0125] According to an embodiment, a “read window” 1402 for detecting the impedance state of a CES element in a read operation may be stated as when the CES element is in an insulating state. Fig.14A The portion 106 of the graph is compared with the voltage V when the CES element is at the reading read When in conductive state Fig.14A In certain embodiments, the read window 1402 can be used to determine the Thomas Fermi screening length λ of the material making up the CES element. TF For example, at voltage V reset Under this condition, the current density J can be obtained according to expression (5): reset and J set Related as follows:

[0126]

[0127] In another embodiment, the “write window” 1410 for placing the CES element in the insulating or conductive memory state during a write operation can be set to V reset (In J reset ) and V set (In J set Establish |V set |>|V reset |Enables switching between conductive and insulating states. V reset can be approximated to be at the band splitting potential generated by the correlation and V setIt may be approximately twice the band splitting potential. In certain embodiments, the size of the write window 1410 may be determined based at least in part on the material and doping of the CES element.

[0128] The transition from high resistance / capacitance to low resistance / capacitance in a CES element can be represented by the singular impedance of the CES element. Fig. 14B A schematic diagram depicting an equivalent circuit for an example variable impedance device, such as a CES element, such as variable impedance device 1424. As mentioned, variable impedance device 1424 may include characteristics of both variable resistance and variable capacitance. For example, in an embodiment, an equivalent circuit for a variable impedance device may include a variable resistor, such as variable resistor 1426, in parallel with a variable capacitor, such as variable capacitor 1428. Of course, while variable resistor 1426 and variable capacitor 1428 may be used in a variety of embodiments, the equivalent circuit may include a variable resistor, such as variable resistor 1426, in parallel with a variable capacitor, such as variable capacitor 1428. Fig. 14B 1400 as including discrete components, but a variable impedance device such as variable impedance device 1424 may include substantially isomorphic OEM elements, where the CEM elements include characteristics of variable capacitance and variable resistance. Table 1 below depicts an example truth table for an example variable impedance device such as variable impedance device 1400.

[0129] resistance capacitance impedance <![CDATA[R high (In applied )]]> <![CDATA[C high (In applied )]]> <![CDATA[Z high (In applied )]]> <![CDATA[R low (In applied )]]> <![CDATA[C low (In applied )~0]]> <![CDATA[Z low (In applied )]]>

[0130] Table 1

[0131] exist Fig.14A In a particular embodiment of a CES element, the CES element can be placed in either of two different impedance states: a low impedance or conductive state in response to a set operation and a high impedance or insulating state in response to a reset operation.

[0132] According to various embodiments, the structure of volatile memory bit cells and non-volatile memory bit cells may be formed in a single integrated circuit (IC) or integrated across multiple ICs (according to various embodiments discussed above). Fig.15 A structure that integrates volatile memory bit cells and non-volatile memory bit cells across multiple ICs in a three-dimensional (3D) IC 1500 as shown in FIG. For example, a layer 1502 in a 3D IC 1500 may include a monolithic IC coupled by metal interconnects, wherein one particular layer 1502 includes one or more memory arrays that include exclusively volatile memory bit cells, while another particular layer 1502 includes one or more memory arrays that include exclusively non-volatile memory bit cells. In other implementations, a structure that integrates volatile memory bit cells and non-volatile memory bit cells may be formed and integrated in a single layer 1502 according to the embodiments discussed above.

[0133] A write operation performed in conjunction with certain embodiments described herein is a specific process of placing a memory device, such as a CES element, in a specific memory state among a plurality of predetermined memory states by applying a "programming signal" to the terminals of the memory device. A specific memory state among the predetermined memory states may correspond to a specific voltage level (e.g., V set and V reset ). Similarly, a particular one of the predetermined memory states may correspond to a particular current level (eg, I set and I reset ). Thus, in certain embodiments, a programming signal used to place a CES device in a certain memory state in a write operation may be controlled to have certain voltage levels and current levels corresponding to the certain memory state.

[0134] As described below in a specific embodiment, a voltage signal having a voltage level of a programming signal for placing a memory device in a predetermined memory state can be selected at least in part at a signal selection circuit based on a data signal. A conductive element connected to the signal selection circuit can selectively connect the voltage signal to the memory device or disconnect the voltage signal from the memory device at a current level corresponding to a predetermined memory state based at least in part on the data signal. In this context, a "conductive element" includes a circuit element that can allow a current to pass between two nodes. In a specific embodiment, a conductive element can change the current allowed to pass between nodes based at least in part on a specific condition. The following specific embodiment uses a FET as a conductive element to allow a current to pass between a source terminal and a drain terminal based at least in part on a voltage applied to a gate terminal. However, it should be understood that other types of devices such as bipolar transistors, diodes, variable resistors, etc. can also be used as conductive elements, and the subject matter claimed for protection is not limited in this regard. In this context, a conductive element having a first terminal and a second terminal can "connect" the first terminal and the second terminal by providing a conductive path between the first terminal and the second terminal with a very small or negligible impedance for a specific signal. In a specific example embodiment, the conductive element can change the impedance between the first terminal and the second terminal based at least in part on a signal provided to a third terminal of the conductive element (e.g., based on a voltage or current applied to the third terminal). The conductive element can be "closed" in response to a signal provided on the third terminal to thereby connect the first terminal and the second terminal. Similarly, the conductive element can be "opened" in response to a different signal provided on the third terminal to thereby disconnect the first terminal and the second terminal. The conductive element in the open state can isolate the first part of the circuit from the second part of the circuit by removing or destroying the conductive path between the first part and the second part of the circuit. The conductive element can change the impedance between the first terminal and the second terminal between the open state and the closed state based on the signal provided to the third terminal.

[0135] Particular embodiments described herein relate to a device comprising: a first memory array comprising a plurality of volatile memory bit cells; a second memory array comprising a plurality of non-volatile memory bit cells; and a decoder circuit coupled to the first memory array and the second memory array to access the non-volatile memory bit cells and the volatile memory bit cells, wherein the decoder circuit is configured to apply access signals to a plurality of word lines for accessing the volatile memory bit cells and the non-volatile memory bit cells. In a particular embodiment, the decoder circuit is further configured to apply access signals to at least one of the plurality of word lines to access at least one of the plurality of volatile memory bit cells and at least one of the plurality of non-volatile memory bit cells. For example, the decoder circuit may include: a first portion dedicated to applying access signals to the first plurality of word lines to access at least one of the plurality of volatile memory bit cells; and a second portion dedicated to applying access signals to the second plurality of word lines to access at least one of the plurality of volatile memory bit cells. In another embodiment, one or more first word lines of the plurality of word lines are adapted to maintain a first access signal (e.g., at a first word line voltage) to access at least one of the plurality of volatile memory bit cells, and one or more second word lines of the plurality of word lines are adapted to maintain a second access signal (e.g., at a second word line voltage) different from the first access signal to access at least one of the plurality of non-volatile memory bit cells. For example, a level shifter circuit coupled between at least one of the one or more first word lines and at least one of the one or more second word lines can be used to maintain at least one of the one or more first word lines and at least one of the one or more second word lines at different voltage levels. In another specific embodiment, the decoder circuit is further configured to: apply a first access signal to a selected word line for a read operation to detect a state of at least one of the volatile memory bit cells; and apply a second access signal to the selected word line for a write operation to write to at least one of the non-volatile memory bit cells based on at least one of the detected states. In another example, the decoder circuit can be further configured to: apply a first access signal to the selected word line for a read operation to detect a state of at least one of the non-volatile memory bit cells; and apply a second access signal to the selected word line for a write operation to write to at least one of the volatile memory bit cells based on at least one of the detected states.In addition, the device may further include: at least one sense amplifier for detecting the state of at least one of the volatile memory bit cells; at least one write driver circuit for generating a programming signal to place at least one of the non-volatile memory bit cells in the detected state; and a bus circuit for transmitting a signal representing the detected state between the at least one sense amplifier and the at least one write driver circuit. In addition, the first access signal and the second access signal may appear within a single memory clock cycle. In another specific embodiment, the decoder is further configured to: apply an access signal to a selected word line to access one or more of the volatile memory bit cells and one or more of the non-volatile memory bit cells; and while the access signal is continuously maintained at a specific voltage to access one or more volatile memory bit cells and one or more non-volatile memory bit cells, perform a read operation to detect the state of at least one of the volatile memory bit cells and perform a write operation on at least one of the non-volatile memory bit cells based on at least one of the detected states. For example, the device may further include: at least one sense amplifier for detecting the state of at least one of the volatile memory bit cells; at least one write driver circuit for generating a programming signal to place at least one of the non-volatile memory bit cells in the detected state; and a bus circuit for sending a signal representing the detected state between the at least one sense amplifier and the at least one write driver circuit. In addition, the device may further include one or more first bit lines coupled to one or more of the volatile memory bit cells and one or more second bit lines coupled to one or more of the non-volatile memory bit cells, and wherein the voltage of the one or more first bit lines and the one or more second voltages are restored in a single memory cycle after a write operation. In addition, a read operation and a write operation occur in a single memory cycle. In another specific embodiment, the decoder circuit is further configured to apply access signals to a plurality of word lines, wherein at least one of the plurality of word lines is connected to at least one of the volatile memory bit cells and at least one of the non-volatile memory bit cells. In another specific embodiment, the decoder circuit is further configured to access volatile memory bit cells or non-volatile memory bit cells or a combination thereof, and the decoder is further configured to initiate access to volatile memory bit cells in a first portion of the address space and initiate access to non-volatile memory bit cells in a second portion of the address space. For example, at least a portion of the first address space may overlap with a portion of the second address space.In addition, the decoder can be further configured to: decode the encoded address to obtain an address in the address space; and access a portion of the volatile memory bit cells or a portion of the non-volatile memory bit cells or a combination thereof based on the decoded address. In another specific embodiment, the decoder circuit can be further configured to: apply a first access signal to the selected word line for a read operation to detect the state of at least one of the non-volatile memory bit cells; and apply a second access signal to the selected word line for a write operation to write to at least one of the volatile memory bit cells based on at least one of the detected states.

[0136] Specific embodiments described herein relate to a method comprising: applying a first access signal to at least a first word line of a plurality of word lines to enable access to one or more volatile memory bit cells and one or more non-volatile memory bit cells. In one specific embodiment, applying the first access signal to the first word line enables access to at least one volatile memory bit cell of the one or more volatile memory bit cells, and the method further comprises applying a second access signal to at least a second word line of the plurality of word lines to enable access to at least one of the one or more non-volatile memory bit cells. In another specific embodiment, the first word line is connected to the one or more volatile memory bit cells, and wherein applying the first signal to the first word line initiates applying the second signal to the second word line connected to the one or more non-volatile memory bit cells. For example, the first word line may be connected to the second word line via a level shifter to apply the second signal to the first word line in response to applying the first signal to the first word line. In another specific embodiment, the method may further include: transferring stored values ​​between the one or more volatile memory bit cells and the one or more non-volatile memory bit cells while accessing the one or more volatile memory bit cells and the one or more non-volatile memory bit cells in response to applying a signal to a first word line. In another specific embodiment, the one or more volatile memory bit cells are formed in a first volatile memory array and the non-volatile memory bit cells are formed in a non-volatile memory array, the non-volatile memory bit cells in the non-volatile memory array being coupled to the volatile memory bit cells in a second volatile memory array, the method further includes: disconnecting the volatile memory bit cells from the non-volatile memory bit cells in the second volatile memory array while applying a first access signal to a first word line to access the one or more volatile memory bit cells and the one or more non-volatile memory bit cells in the first volatile memory array. For example, disconnecting the volatile memory bit cells in the second volatile memory array from the nonvolatile memory bit cells can further include applying one or more signals to one or more flip-flop circuits coupled between word lines coupled to the volatile memory bit cells in the second volatile memory array of the nonvolatile memory bit cells.

[0137] Certain embodiments described herein are also directed to a device comprising: at least a first nonvolatile memory array including a plurality of nonvolatile memory bit cells; at least a first volatile memory array including a first plurality of volatile memory bit cells; at least a second volatile memory array including a second plurality of volatile memory bit cells; and a decoder circuit configured to assert a first access signal on at least a first word line coupled to at least a portion of the nonvolatile memory bit cells, a portion of the first plurality of volatile memory bit cells, and a portion of the second plurality of volatile memory bit cells, and wherein a portion of the nonvolatile memory bit cells is coupled to the first word line between a portion of the first plurality of volatile memory bit cells and a second portion of the volatile memory bit cells. In a particular embodiment, the device further includes: a first access device coupled to a first word line between a first plurality of volatile memory bit cells and a portion of non-volatile memory bit cells to disconnect the first plurality of volatile memory bit cells while accessing a second plurality of volatile memory bit cells; and a second access device coupled between the second plurality of volatile memory bit cells and a portion of non-volatile memory bit cells via the first word line to disconnect the second plurality of volatile memory bit cells while accessing the first plurality of volatile memory bit cells. For example, the decoder circuit can be further configured to transfer data between the first portion of volatile memory bit cells and the portion of non-volatile memory bit cells while the second portion of volatile memory bit cells is disconnected from the portion of non-volatile memory. Additionally, the first access device can include at least a first tri-state buffer, and wherein the second access device includes at least a second tri-state buffer. For example, a first word line may be connected to a portion of a first volatile memory bit cell, wherein the device further includes a second word line connected to a portion of a non-volatile memory bit cell and a third word line connected to a portion of a second volatile memory bit cell, and wherein: the first tri-state buffer further includes a first level shifter circuit for applying a second access signal to the second word line in response to the first access signal, the first access signal and the second access signal having different voltages; and the second tri-state buffer further includes a second level shifter circuit for applying a third access signal to the second word line in response to a fourth access signal applied to the third word line, the third access signal and the fourth access signal having different voltages.

[0138] Certain embodiments described herein are also directed to a device formed in an integrated circuit device, the device comprising: at least a first nonvolatile memory array, the at least first nonvolatile memory array comprising a plurality of addressable nonvolatile memory bit cells; at least a second volatile memory array, the at least second volatile memory array comprising a plurality of addressable volatile memory bit cells; at least one nonvolatile memory data bus, the at least one nonvolatile memory data bus configurable to transfer data from the nonvolatile bit cells in a read operation and to transfer data to the nonvolatile bit cells in a write operation; at least one volatile memory data bus configurable to transfer data from the nonvolatile bit cells in a read operation and to transfer data to the nonvolatile bit cells in a write operation; The first nonvolatile memory array and the first volatile memory array are configured to transfer data values ​​to the nonvolatile bit cells or the volatile memory bit cells or a combination thereof in a write operation in response to a first signal applied to an external terminal of the integrated circuit device, the data values ​​being based at least in part on the first signal, and to apply a second signal to the external terminal of the integrated circuit device in response to data transferred on the at least one nonvolatile memory data bus or the at least one volatile memory data bus in response to a read operation of the nonvolatile bit cells or the volatile memory bit cells. In a particular embodiment, the first nonvolatile memory array and the first volatile memory array are formed substantially adjacent to each other such that the first nonvolatile memory array and the first volatile memory array share a single combined hierarchical memory register transfer logic (RTL) boundary. In another specific embodiment, at least one non-volatile memory data bus and at least one volatile memory data bus may be configured to copy memory states between non-volatile memory bit cells and volatile memory bit cells independently of the data port. In another specific embodiment, at least one non-volatile memory data bus and at least one volatile memory data bus may be configured to copy memory states between non-volatile memory bit cells and volatile memory bit cells in a single memory cycle. In another specific embodiment, at least one non-volatile memory data bus and at least one volatile memory data bus have different data widths, and wherein at least one non-volatile memory data bus and at least one volatile memory data bus may be configured to copy memory states between non-volatile memory bit cells and volatile memory bit cells in multiple memory cycles.

[0139] Specific embodiments described herein also relate to a method comprising: applying a first signal to an external terminal of an integrated circuit device to access a volatile memory bit cell in a volatile memory array formed in the integrated circuit device; and applying a second signal to an external terminal of the integrated circuit device to access a non-volatile memory bit cell in a non-volatile memory array formed in the integrated circuit device, wherein the external terminal is coupled to a data port of the integrated circuit device, the data port being configured to access the volatile memory array or the non-volatile memory array or a combination thereof. In one specific embodiment, the method further comprises applying one or more signals to at least one of the external terminals to initiate a transfer of a value stored in the volatile memory array to the non-volatile memory array. In another specific embodiment, the method further comprises applying one or more signals to at least one of the external terminals to initiate a transfer of a value stored in the non-volatile memory array to the volatile memory array.

[0140] Certain embodiments described herein also relate to a device comprising: a plurality of first columns of volatile memory bit cells, the volatile memory bit cells in each of the first columns of volatile memory bit cells being accessible via one or more common first bit lines; a plurality of second columns of nonvolatile memory bit cells, the nonvolatile memory bit cells in each of the second columns of nonvolatile memory bit cells being accessible via one or more common second bit lines, wherein the first columns and the second columns are arranged in an array such that the first columns of volatile memory bit cells alternate and interleave with the second columns of nonvolatile memory bit cells; and circuitry for transferring values ​​between selected volatile memory bit cells in a selected column of the first columns of volatile memory bit cells and selected nonvolatile memory bit cells in a selected column of the second columns of nonvolatile memory bit cells, the selected column of the first columns of volatile memory bit cells and the selected column of the second columns of nonvolatile memory bit cells being adjacent. In certain embodiments, the selected volatile memory bit cells and the selected nonvolatile memory bit cells can be selected in response to a word line signal. In another specific embodiment, the circuit for transferring a value between a selected volatile memory bit cell and a selected non-volatile memory bit cell further comprises: a read circuit connected to one or more common first bit lines to detect a value stored in the selected volatile memory bit cell; and a write driver circuit connected to one or more common second bit lines to generate a programming signal to store the detected value in the selected non-volatile memory bit cell. In another specific embodiment, the circuit for transferring a stored value between a selected volatile memory bit cell and a selected non-volatile memory bit cell further comprises: a read circuit connected to one or more common second bit lines to detect a value stored in the selected volatile memory bit cell; and a write driver circuit connected to one or more common first bit lines to generate a programming signal to store the detected value in the selected volatile memory bit cell. In another specific embodiment, the selected column of the first column of volatile memory bit cells is two or more bit cells wide, and wherein the device further includes a multiplexer configurable to connect a bit line connected to the selected volatile memory bit cell from among the bit lines connected to the bit cells in the selected column of the first column of nonvolatile memory bit cells to a circuit to transfer a value between the selected volatile memory bit cell and the selected nonvolatile memory bit cell. In another specific embodiment, the selected column of the first column of nonvolatile memory bit cells is two or more bit cells wide, and wherein the device further includes a multiplexer configurable to connect a bit line connected to the selected nonvolatile memory bit cell from among the bit cells in the selected column of the first column of nonvolatile memory bit cells to a circuit to transfer a value between the selected volatile memory bit cell and the selected nonvolatile memory bit cell.

[0141] Certain embodiments described herein are also directed to a method comprising: transferring values ​​between selected volatile memory bit cells in a first selected column of first columns of volatile memory bit cells and selected non-volatile memory bit cells in a second selected column of second columns of non-volatile memory bit cells, the first selected column of the first columns of volatile memory bit cells and the selected column of the second columns of non-volatile memory bit cells being adjacent, wherein: the volatile memory bit cells in each first column of volatile memory bit cells are accessible via one or more common first bit lines; the non-volatile memory bit cells in each second column of non-volatile memory are accessible via one or more common second bit lines; and the first column and the second column are configured in an array such that the first column of volatile memory bit cells alternates and interleaves with the second column of non-volatile memory bit cells. In certain embodiments, transferring values ​​between the selected volatile memory bit cells and the selected non-volatile memory bit cells further comprises applying a signal to a word line coupled to the selected volatile memory bit cells and the selected non-volatile memory bit cells. In another specific embodiment, transferring a value between a selected volatile memory bit cell and a selected non-volatile memory bit cell further comprises: detecting a value stored in the selected volatile memory bit cell at a read circuit connected to one or more common first bit lines in a read operation; and generating a programming signal to store the detected value in the selected non-volatile memory bit cell. In another specific embodiment, transferring a value between a selected volatile memory bit cell and a selected non-volatile memory bit cell further comprises: detecting a value stored in the selected non-volatile memory bit cell at a read circuit connected to one or more common bit lines in a read operation; and generating a programming signal to store the detected value in the selected volatile memory bit cell. In another specific embodiment, wherein the selected column in the first column of volatile memory bit cells is two or more bit cells wide, transferring a value between the selected volatile memory bit cell and the selected non-volatile memory bit cell further comprises connecting a bit line connected to the selected volatile memory bit cell from among the bit lines connected to the bit cells in the selected column in the first column of volatile memory bit cells to a circuit to transfer the value between the selected volatile memory bit cell and the selected non-volatile memory bit cell. In another specific embodiment, the selected column in the first column of non-volatile memory bit cells is two or more bit cells wide, and wherein transferring a value between the selected volatile memory bit cell and the selected non-volatile memory bit cell further comprises connecting a bit line connected to the selected non-volatile memory bit cell from among the bit lines connected to the bit cells in the selected column in the first column of non-volatile memory bit cells to a circuit to transfer the value between the selected volatile memory bit cell and the selected non-volatile memory bit cell.

[0142] References throughout this specification to an embodiment, an embodiment, an example, an embodiment, and / or the like mean that the specific features, structures, and / or characteristics described in conjunction with the specific embodiment and / or embodiment are included in at least one embodiment and / or embodiment of the claimed subject matter. Therefore, the appearance of such phrases, for example, in various places throughout this specification is not necessarily intended to refer to the same embodiment or to any one of the specific embodiments described. In addition, it should be understood that the specific features, structures, and / or characteristics described can be combined in various ways in one or more embodiments, and therefore, for example, within the scope of the predetermined claims. In general, of course, these and other issues vary with the context. Therefore, the specific context of the description and / or usage provides helpful guidance on the reasoning to be drawn.

[0143] Although what are presently considered to be example features have been illustrated and described, it will be understood by those skilled in the art that various other modifications may be made, and equivalents may be substituted, without departing from the claimed subject matter. Additionally, many modifications may be made to adapt a particular situation to the teachings of the claimed subject matter without departing from the central concept described herein. Therefore, it is intended that the claimed subject matter not be limited to the specific examples disclosed, but that such claimed subject matter may also include all aspects falling within the scope of the appended claims and their equivalents.

Claims

1. An integrated circuit device, include: at least a first nonvolatile memory array comprising a first plurality of nonvolatile memory bit cells; at least a first volatile memory array comprising a first plurality of volatile memory bit cells; a shared decoder circuit coupled to the first nonvolatile memory array and the first volatile memory array to access one or more of the first plurality of nonvolatile memory bit cells and one or more of the first plurality of volatile memory bit cells via assertion of one or more access signals; a shared bus structure, the shared bus structure comprising at least one nonvolatile memory bus, the at least one nonvolatile memory bus being configurable to transfer memory states from the first plurality of nonvolatile bit cells in a read operation and to transfer memory states to the first plurality of nonvolatile bit cells in a write operation, and the shared bus structure further comprising at least one volatile memory bus, the at least one volatile memory bus being configurable to transfer memory states to the first plurality of volatile bit cells in a write operation and to transfer memory states from the first plurality of volatile bit cells in a read operation; as well as at least one external port for transferring input values ​​and output values ​​between the shared bus structure and one or more external terminals of the integrated circuit device, The shared bus structure is coupled to the at least one non-volatile memory bus and the at least one volatile memory bus to enable memory state transfer between the at least one external port and the at least one non-volatile memory bus or the at least one volatile memory bus or a combination thereof, and / or to enable memory state transfer between the at least one non-volatile memory bus and the at least one volatile memory bus.

2. The integrated circuit device according to claim 1, in, The first plurality of non-volatile memory bit cells comprises associated electronic switch CES elements adaptable to be placed in a first impedance state comprising a conductive state and / or a low impedance state and a second impedance state comprising an insulating and / or a high impedance state.

3. An integrated circuit device according to claim 1 or claim 2, in, The shared decoder circuit is configured to apply the one or more access signals to at least one of the plurality of word lines to access one or more of the first plurality of volatile memory bit cells and one or more of the first plurality of non-volatile memory bit cells.

4. The integrated circuit device according to claim 3, in, The shared decoder circuit is further configured to apply the one or more access signals to the plurality of word lines to access one or more of the first plurality of volatile memory bit cells and one or more of the first plurality of non-volatile memory bit cells.

5. The integrated circuit device according to claim 3, in, The shared decoder circuit is configured to assert the one or more access signals to a first word line, wherein a first portion of the first word line is used to access at least one of the first plurality of volatile memory bit cells; and wherein a second portion of the first word line is used to access at least one of the first plurality of non-volatile memory bit cells.

6. The integrated circuit device according to claim 3, in, One or more first word lines of the plurality of word lines are used to maintain a first access signal at a first voltage to access one or more of the first plurality of volatile memory bit cells, and wherein one or more second word lines of the plurality of word lines are used to maintain a second access signal at a second voltage different from the first voltage to access one or more of the first plurality of non-volatile memory bit cells.

7. The integrated circuit device according to claim 3, in, The shared decoder circuit is further configured to: applying a first access signal of the one or more access signals to a selected word line to perform a read operation to detect one or more states of one or more of the first plurality of volatile memory bit cells; and Based at least in part on at least one of the one or more detected states, a second one of the one or more access signals is applied to the selected word line for a write operation to write to one or more of the first plurality of nonvolatile memory bit cells.

8. The integrated circuit device according to claim 7, in, The shared decoder circuit is further configured to: applying the first access signal to a selected word line to perform a read operation to detect one or more memory states of one or more of the first plurality of nonvolatile memory bit cells; and Based on at least one of the one or more detected memory states, a second access signal is applied to the selected word line for a write operation to write to one or more of the first plurality of volatile memory bit cells.

9. The integrated circuit device according to claim 3, in, The shared decoder circuit is further configured to: applying a first access signal of the one or more access signals to a selected word line of the plurality of word lines to access one or more of the first plurality of volatile memory bit cells and one or more of the first plurality of non-volatile memory bit cells; and When the first access signal is maintained at a voltage to access one or more of the first plurality of volatile memory bit cells and one or more of the first plurality of non-volatile memory bit cells, a read operation is performed to detect one or more memory states of one or more of the first plurality of volatile memory bit cells and a write operation is performed on one or more of the first plurality of non-volatile memory bit cells based at least in part on at least one of the one or more detected memory states.

10. The integrated circuit device according to claim 9, further comprising: include: at least one sense amplifier for detecting one or more memory states of one or more of the first plurality of volatile memory bit cells; at least one write driver circuit for generating a programming signal to place one or more of the first plurality of nonvolatile memory bit cells into one or more detected memory states; as well as A bus circuit is provided for transmitting signals and / or states indicative of one or more detected memory states between the at least one sense amplifier and the at least one write driver circuit.

11. The integrated circuit device according to claim 3, in, The shared decoder circuit is further configured to apply the one or more access signals to a plurality of word lines, wherein at least one of the plurality of word lines is connected to one or more of the first plurality of volatile memory bit cells and one or more of the first plurality of non-volatile memory bit cells.

12. The integrated circuit device of claim 1 or claim 2, further comprising at least a second volatile memory array comprising a second plurality of volatile memory bit cells; and in, The shared decoder circuit is configured to assert a first access signal on at least a first word line coupled to at least a portion of the first plurality of non-volatile memory bit cells, a portion of the first plurality of volatile memory bit cells, and a portion of the second plurality of volatile memory bit cells, and wherein a portion of the first plurality of non-volatile memory bit cells is coupled to the first word line between a portion of the first plurality of volatile memory bit cells and a portion of the second plurality of volatile memory bit cells.

13. The integrated circuit device according to claim 12, further comprising: include: a first access device coupled to the first word line between a portion of the first plurality of non-volatile memory bit cells and the first plurality of volatile memory bit cells to disconnect the first plurality of volatile memory bit cells when accessing the second plurality of volatile memory bit cells; as well as A second access device is coupled between a portion of the first plurality of non-volatile memory bit cells and the second plurality of volatile memory bit cells through the first word line to disconnect the second plurality of volatile memory bit cells when the first plurality of volatile memory bit cells are accessed.

14. An integrated circuit device according to claim 1 or claim 2, in: The first volatile memory array includes at least a first column of volatile memory bit cells, wherein the volatile memory bit cells in the first column of volatile memory bit cells are accessible via one or more common first bit lines; The non-volatile memory array comprises at least a second column of non-volatile memory bit cells, wherein the non-volatile memory bit cells in the second column of non-volatile memory bit cells are accessible via one or more common second bit lines; and the integrated circuit device further comprises: Circuitry for transferring signals representing one or more values ​​between one or more selected volatile memory bit cells in the first column of volatile memory bit cells and one or more selected non-volatile bit cells in the second column of non-volatile memory bit cells, the first column of volatile memory bit cells and the second column of non-volatile memory bit cells being adjacent.

15. The integrated circuit device according to claim 14, in, The one or more selected volatile memory bit cells and the one or more selected non-volatile memory bit cells are selectable in response to a word line signal.

16. The integrated circuit device according to claim 14, in, The circuit for transferring signals representing one or more values ​​between the one or more selected volatile memory bit cells and the one or more selected non-volatile memory bit cells further comprises: a read circuit connected to the one or more common first bit lines to detect the signal representing the value stored in at least one of the one or more selected volatile memory bit cells; and a write driver circuit connected to the one or more common second bit lines to generate a programming signal to store the detected signal representing the value stored in at least one of the one or more selected volatile memory bit cells in at least one of the one or more selected non-volatile memory bit cells, and Wherein, the circuit for transferring signals representing one or more values ​​between the one or more selected volatile memory bit cells and the one or more selected non-volatile memory bit cells further comprises: a read circuit connected to the one or more common second bit lines to detect the signal representing the value stored in at least one of the one or more selected non-volatile memory bit cells; and a write driver circuit connected to the one or more common first bit lines to generate a programming signal to store the detected signal representing the value stored in at least one of the one or more selected non-volatile memory bit cells in at least one of the one or more selected volatile memory bit cells.

17. A method for integrating a memory, include: applying a first access signal to at least a first word line of a plurality of word lines via a shared decoder circuit to access one or more volatile memory bit cells in a first volatile memory array formed in the integrated circuit device and one or more nonvolatile memory bit cells in a nonvolatile memory array formed in the integrated circuit device; applying one or more first signals to one or more external terminals of the integrated circuit device to access at least some volatile memory bit cells in the first volatile memory array; as well as applying one or more second signals to one or more external terminals of the integrated circuit device to access at least some of the one or more nonvolatile memory bit cells in the nonvolatile memory array, The one or more external terminals are coupled to a port of the integrated circuit device, and the port is used to access the first volatile memory array or the non-volatile memory array or a combination thereof.

18. The method according to claim 17, in, The first access signal is applied to the first word line to access the one or more volatile memory bit cells and the one or more non-volatile memory bit cells in the non-volatile memory array, the method also including applying the first access signal to at least a first portion of the first word line to access the one or more volatile memory bit cells, and applying the first access signal to at least a second portion of the first word line to access the one or more non-volatile memory bit cells in the non-volatile memory array.

19. The method according to claim 17 or claim 18, further comprising: include: One or more stored signals representing one or more values ​​are transferred between the one or more volatile memory bit cells and the one or more nonvolatile memory bit cells while accessing the one or more volatile memory bit cells and the one or more nonvolatile memory bit cells in response to applying the first access signal to at least the first word line of the plurality of word lines.

20. The method according to claim 17 or claim 18, in, The one or more nonvolatile memory bit cells in the nonvolatile memory array are coupled to one or more volatile memory bit cells in a second volatile memory array formed in the integrated circuit device, the method further comprising: The one or more volatile memory bit cells in the second volatile memory array are disconnected from the one or more nonvolatile memory bit cells in the nonvolatile memory array while the first access signal is applied to at least the first of the plurality of word lines to access one or more volatile memory bit cells in the first volatile memory array and one or more nonvolatile memory bit cells in the nonvolatile memory array.

21. A method for integrating a memory, include: applying one or more first signals to one or more external terminals of an integrated circuit device to access one or more volatile memory bit cells in a volatile memory array formed in the integrated circuit device; as well as applying one or more second signals to one or more external terminals of the integrated circuit device to access one or more nonvolatile memory bit cells in a nonvolatile memory array formed in the integrated circuit device, wherein at least one of the one or more volatile memory bit cells and at least one of the one or more non-volatile memory bit cells are connected to a word line of a plurality of word lines formed in the integrated circuit device, wherein a shared decoder circuit is coupled to the nonvolatile memory array and the volatile memory array to access the one or more nonvolatile memory bit cells and the one or more volatile memory bit cells by applying access signals to word lines of the plurality of word lines, The one or more external terminals are coupled to a port of the integrated circuit device, and the port is used to access the volatile memory array or the non-volatile memory array or a combination thereof.

22. The method according to claim 21, further comprising: include: transferring one or more signals representing values ​​between selected volatile memory bit cells in a first selected column of a first plurality of columns of the volatile memory array and selected non-volatile memory bit cells in a second selected column of a second plurality of columns of the non-volatile memory array, the first selected column of the first plurality of columns of the volatile memory array and the selected column of the second plurality of columns of the non-volatile memory array being adjacent, wherein: One or more volatile memory bit cells in a first column of the volatile memory array are accessible via one or more common first bit lines; One or more nonvolatile memory bit cells in a second column of the nonvolatile memory array are accessible via one or more common second bit lines; and The first and second columns are arranged in an array such that the first plurality of columns of the volatile memory array alternate and interleave with the second plurality of columns of the non-volatile memory array.

23. The method according to claim 21 or claim 22, further comprising: include: One or more signals representing one or more values ​​are transferred between the volatile memory array and the non-volatile memory array using a shared data bus structure.

24. The method of claim 23, further comprising transferring one or more signals representing one or more values ​​between the shared data bus structure and the one or more external terminals, in, The shared decoder circuit is further used to assert one or more second access signals on a plurality of word lines, wherein at least one of the plurality of word lines is connected to at least one of the volatile memory bit cells and at least one of the non-volatile memory bit cells.

Citation Information

Patent Citations

  • Page-buffer and non-volatile semiconductor memory including page buffer

    CN1779859A

  • Nonvolatile storage device

    JP2007012140A

  • Memory provided with associated volatile and non-volatile memory cells

    US20160329100A1

  • Direct interface between SRAM and non-volatile memory

    US20170220491A1

  • Hybrid nonvolatile shadowed DRAM with an overlapping region between a volatile storage die and a nonvolatile storage die

    US8837236B2