WO2010077414A1 - Non-volatile memory device for concurrent and pipelined memory operations - Google Patents
Non-volatile memory device for concurrent and pipelined memory operations Download PDFInfo
- Publication number
- WO2010077414A1 WO2010077414A1 PCT/US2009/060911 US2009060911W WO2010077414A1 WO 2010077414 A1 WO2010077414 A1 WO 2010077414A1 US 2009060911 W US2009060911 W US 2009060911W WO 2010077414 A1 WO2010077414 A1 WO 2010077414A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- data
- buffers
- memory
- buffer
- read
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C16/00—Erasable programmable read-only memories
- G11C16/02—Erasable programmable read-only memories electrically programmable
- G11C16/06—Auxiliary circuits, e.g. for writing into memory
- G11C16/26—Sensing or reading circuits; Data output circuits
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C16/00—Erasable programmable read-only memories
- G11C16/02—Erasable programmable read-only memories electrically programmable
- G11C16/06—Auxiliary circuits, e.g. for writing into memory
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C16/00—Erasable programmable read-only memories
- G11C16/02—Erasable programmable read-only memories electrically programmable
- G11C16/06—Auxiliary circuits, e.g. for writing into memory
- G11C16/10—Programming or data input circuits
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C2216/00—Indexing scheme relating to G11C16/00 and subgroups, for features not directly covered by these groups
- G11C2216/12—Reading and writing aspects of erasable programmable read-only memories
- G11C2216/22—Nonvolatile memory in which reading can be carried out from one memory bank or array whilst a word or sector in another bank or array is being erased or programmed simultaneously
Definitions
- non-volatile memory devices are designed so that portions of memory space must be written or erased together as a single storage unit, typically as a "page" or “block;” owing to variability in each memory cell within these designs, operations that change the cell data state (e.g. program or erase) in turn are often based on iterative cycles. These iterative cycles are called “program- verify” cycles and, despite use of "program” within this term, it should be understood that the term refers to the same basic operation whether writing or erasing of a block is performed.
- the memory device or a controller (i) attempts to change the state of the storage unit using a specific voltage, and (ii) checks to see which cells have not changed state correctly; cycles are then repeated as necessary for those cells which have not yet correctly changed state using a slightly higher voltage each iteration until the cells either have the correct state or until an error is generated (the unit would then typically be marked as "bad").
- FIG. 1 is a functional diagram of a memory device that has internal routing that can support interleaved use of a sense amplifier unit 1 17 for both reads to a non-volatile storage array 103 and a multiple cycle state change operation that involves sense amplifier unit feedback to a write data register.
- FIG. 2 is a functional diagram of a second memory device. Unlike the device seen in FIG. 1 , the embodiment presented by FIG. 2 uses an array 21 1 of buffers (213, 215 and 217) and a crossbar 225 to (1 ) connect any one of the buffers to the array so that each buffer may serve as a temporary destination for read data (e.g., as a cache to later serve pages of read data to a controller) and (2) connect any one of the buffers to write circuitry 223 (e.g., so that the buffers may be used to queue upcoming erase and write transactions until an operation already in-progress within the write circuitry is completed).
- an array 21 1 of buffers (213, 215 and 217) and a crossbar 225 to (1 ) connect any one of the buffers to the array so that each buffer may serve as a temporary destination for read data (e.g., as a cache to later serve pages of read data to a controller) and (2) connect any one of the buffers to write circuitry 223 (e.g.,
- FIG. 3 is a functional diagram showing yet another memory device as part of a memory system 301 , with the memory system also including a flash controller 303.
- FIG. 3 indicates that the system effectively provides two crossbars, one between the buffers and a flash storage array 307 for each memory device, and one between the buffers and the flash controller.
- a vertical, dashed line 315 denotes that separate control logic may be used for each crossbar, such that input/output (“IO") control for the device may be managed independently of memory array control. That is to say, control logic for the IO can manage flows of data between the controller and the various buffers 313 independently of flows of data between the buffers and the flash storage array 307.
- FIG. 4 is a flow diagram of a method 401 of tracking buffer usage in a multiple buffer embodiment. As indicated by dashed-line functional blocks, the method 401 provides for the execution of one memory operation while data for another memory operation is concurrently being transferred into or out of a buffer.
- FIG. 5 is a flow diagram of a method 501 used by IO control logic to process newly received memory requests from a controller.
- the method 501 calls for managing a status register 517 that tracks the usage and status of each of a plurality of buffers; three buffers, labeled "B1 ,” “B2,” and “B3" are tracked by the register.
- the register permits assignment of buffers to new memory requests, but it also permits the read of data that is currently in the midst of a state change operation.
- the buffers may collectively be used as a data cache while the device sense array and write data register are busy.
- the status register may be configured to support a multiple master implementation.
- FIG. 6 is a flowchart that shows a method 601 of managing read requests as they are received from a controller.
- the method 601 may be implemented by IO control logic that manages communication flow primarily on "one side" of buffers, between the buffers and an external data bus.
- FIG. 7 is a flowchart of a method 701 that shows general memory array management tasks associated with memory array control logic (i.e., that may be used to manage communication flow on the "other side" of the buffers, between the buffers and the memory array). As will be explained below (and as was the case with the functions of FIG. 6), these functions may be managed either by a memory controller or by hardware or instructional logic resident on each memory device.
- memory array control logic i.e., that may be used to manage communication flow on the "other side" of the buffers, between the buffers and the memory array.
- FIG. 8 is a flowchart of a method 801 associated with programming tasks (i.e., tasks that involve write circuitry for the memory array).
- data may be copied from an assigned buffer into a write data register ("WDR"), permitting the data to be retained in the assigned buffer and read during state change operations, notwithstanding that write data register (“WDR") contents (and contents of the pertinent memory space within the array) are themselves in transition during programming.
- FIG. 9 is a flowchart of a method 901 associated with each program-verify cycle of a program or erase command. The cycles are repeated as necessary until (1 ) the associated space in memory has correctly changed state (in which case the assigned buffer used to cache program data is released) or (2) a maximum number of cycles is reached (represented as a cap upon the variable "/c" seen in FIG. 9).
- FIG. 10 is a flowchart of a method 1001 used by the memory array control logic to process a read operation (i.e., the step of reading a page of data from the memory array and transferring that data to the buffers).
- the method includes configuring the crossbar to route data from a sense amplifier unit to an assigned buffer, and changing buffer status register flags so that the data is not read out of the buffer while in the process of being loaded.
- This disclosure provides several embodiments of a non-volatile memory device that variously (a) permits the performance of one memory operation while another memory operation is queued, (b) permits (within a single "bank” of memory as well as multiple banks) the generally parallel read of one storage unit while another is the subject of a state change operation (e.g., while it is being written), (c) permits generally parallel state changing of a storage unit and concurrent read of the new (i.e., write or erase data being placed into that storage unit), and (d) provides methods of controlling the memory device, in a system or otherwise, to achieve one of these functions.
- non-volatile memory e.g., within a flash memory device
- the embodiments discussed herein provide for a significant improvement in the usability of certain non-volatile devices. That is to say, the teachings provided by this disclosure should result in increased use of non-volatile memory, especially flash memory, to a much broader range of applications, potentially extending the capabilities of those applications and enabling new applications.
- a first memory device embodiment provides a non-volatile storage device, labeled 101 in FIG. 1 .
- This non-volatile storage device permits reads of a specific storage array 103 notwithstanding that a state change operation may be contemporaneously in progress.
- the array is divided into a number of storage units 105 that each represents the minimum grouping of memory cells that must be programmed together or erased together.
- certain conventional flash devices reflect a design philosophy of high density and very low cost, and so sacrifice access granularity in favor of cost and density; certain designs are such that their structures must be programmed in "pages” and erased in "blocks," e.g., NAND flash -
- a "page” may consist of about 4,000 bytes of data, and a "block” may consist of thousands of pages.
- the term "page” will be used to refer to a memory unit that is at least as large as the minimum structure or unit that can be programmed or erased as a single unit.
- each individual memory cell (either a single level cell, i.e., "SLC,” or multi-level cell, i.e., "MLC,” depending on design) can have different levels of defects as an artifact of manufacturing or developed through use, each cell may require a different voltage before it can be changed to the desired logic state. Accordingly, following application of the voltage (i.e., following the "programming phase") the device 101 then performs a verify operation using a sense array (or "sense amplifier unit") 1 17, to compare the contents of the storage unit 1 15 with the contents of the write data register 1 1 1 using the write circuitry 1 13 (i.e., the "verify phase").
- a sense array or "sense amplifier unit"
- the embodiment seen in FIG. 1 permits reads of a specific storage array 103 during a state change operation using novel internal routing and what effectively is a switching mechanism, depicted by numeral 121 in FIG. 1 .
- the routing includes three effective paths, including a first path 123 that effectively couples the write data register 1 1 1 with the sense amplifier unit 1 17, a second path 125 that effectively couples the IO interface 107 with the write data register 1 1 1 and a third path 127 that couples the sense amplifier unit with the IO interface; the third path is used during a data read to load data from the memory space (as sensed by the sense amplifier unit) into a buffer 129, for subsequent output to a memory controller.
- FIG. 1 The routing includes three effective paths, including a first path 123 that effectively couples the write data register 1 1 1 with the sense amplifier unit 1 17, a second path 125 that effectively couples the IO interface 107 with the write data register 1 1 1 and a third path 127 that couples the sense amplifier unit with the IO interface; the third path is used during
- concurrent writes and reads may occur because the data output from the sense amplifier unit can be rerouted during a multiple cycle state change operation between either the write data register (via the first path 123) or to the buffer (via the third path 127). This operation is also performed under the auspices of the control logic 1 19.
- FIG. 1 represents a different approach to the latency problem than mere use of multiple banks.
- multiple bank structures it is possible to read one bank (or a different device) at the same time that a state change operation is in-progress in a second bank (or device), because each bank or device has its own write data register; by contrast, the embodiment presented in FIG. 1 permits the same memory storage space, served by that write data register, to be read during a multiple cycle state change operation that ties up that register.
- some embodiments presented by this disclosure permit simultaneous read of data in the same nonvolatile storage array, but it should be understood that the embodiment of FIG. 1 is not required to have this capability.
- FIG. 2 illustrates a second embodiment that may be configured to provide this current read capability for data being that will programmed or otherwise written in to memory (“program data").
- FIG. 2 shows a second memory device 201 having some similar elements to those discussed above in connection with FIG. 1 .
- the memory device includes a non-volatile storage array 203, with the storage space being divided into a number of storage units 205, an IO interface 207, and pins (not shown) for connecting the IO interface with an external bus 209.
- the memory device 201 of FIG. 2 also includes an array or set 21 1 of data buffers, each buffer sized to at least correspond to a storage unit size (e.g., page size, for a NAND flash memory device).
- This second embodiment is seen to have at least three buffers in the array, respectively numbered at 213, 215 and 217, each coupled to the IO interface via an internal bus 219.
- the buffers may each be used to store a page of inbound write data, such that concurrent operations may be processed in parallel.
- processing of multiple read operations e.g., one buffer (such as buffer 213) can be filled with data transferred from the sense amplifier unit 221 while another buffer (such as buffer 215) is feeding its contents to the external bus;
- processing of multiple state change operations e.g., one buffer (such as buffer 215) can be filled with data from the external bus 209 while another buffer (such as buffer 213) is feeding its contents to the write circuitry 223 (and to the write data register, not separately depicted in FIG. 2);
- one buffer (such as buffer 217) can be filled with data transferred from the sense amplifier unit 221 either while another buffer (such as buffer 213) is being filled with data from the IO interface 207 or while the sense amplifier unit and write circuitry are performing a multiple cycle write or erase operation; and
- data to be written into the storage array 203 may be stored as a copy in one of the buffers (213, 215 or 217) while the operation is ongoing (i.e., with the sense amplifier unit feeding back data to the write circuitry) such that during an operation, data may be read from the buffer array 21 1 instead of from the storage array 203.
- the buffer array 21 1 is coupled to the sense amplifier unit 221 , on the one hand, and to the write circuitry 223, on the other hand, by a crossbar unit (denoted by the label "XBAR" and numeral 225 in FIG. 2).
- this switching is provided on at least a page-wide basis, e.g., if page size if four kilobytes, then the crossbar would provide 4kB of simultaneous switching for each point of origin in order to affect the transfers alluded to above.
- control logic may be provided (on an on-board basis if desired) in order to effect this switching within the internal routing, e.g., to control the crossbar in the embodiment of FIG. 2.
- control logic may be provided (on an on-board basis if desired) in order to effect this switching within the internal routing, e.g., to control the crossbar in the embodiment of FIG. 2.
- a dashed line that connects boxes 221 and 223 it is typically desired to feed data back from the sense arrays to the write circuitry for use in state change operations; this operation may be implemented by a direct connection between these two elements or, alternatively, by providing the crossbar 225 with capability to perform this routing.
- the write data register (not separately identified in FIG. 2) is typically a "working buffer" that is modified with each cycle to store change data, that is, to identify which cells have not been correctly changed thus far, and to indicate state changes still needed with each cycle. It would also be possible to configure the write data register as a conventional buffer in alternate implementations.
- FIG. 2 shows a dashed line box 227, which represents an optional addressing and control function to support external addressing into the memory device.
- a dashed line box 227 represents an optional addressing and control function to support external addressing into the memory device.
- the buffers 213, 215 and 217 may be implemented such that an external device may write program data directly to these buffers or retrieve read data from these buffers (once a corresponding page has been relayed into the buffer array).
- this addressing and control function may permit a controller to initiate operations and then pull read data whenever the memory device has sensed the data, loaded into a buffer, and is primed to provide the data.
- FIG. 3 represents yet another implementation of principles provided by this disclosure.
- FIG. 3 shows a system 301 that includes a flash controller 303 and one or more flash devices - only one exemplary flash device 305 is depicted in FIG. 3.
- Each flash device includes a flash storage array 307 and means for connecting to an external data bus 309, as well as write circuitry 31 1 (including a write data register), a sense amplifier unit 313 and buffer array 314, as was discussed above in connection with the embodiment of FIG. 2.
- FIG. 3 further includes, however, a vertical line 315 used to distinguish memory array control from IO control, respectively denoted by reference numerals 317 and 319.
- FIG. 3 shows a vertical line 315 used to distinguish memory array control from IO control, respectively denoted by reference numerals 317 and 319.
- IO control logic governs the transfer of data between the buffer array 314 and the external bus 309
- memory array control logic governs communications between the buffer array 314 and the flash storage array.
- Each of the memory array control logic and IO control logic functions may fill buffers with new data and may unload (i.e., transfer) a copy of data from one of the buffers elsewhere.
- these buffers can be cleared (i.e., released) once data has been transferred out, or they may be caused to retain data until occurrence of a specific event (such as completion of a specific programming or state change step).
- a specific event such as completion of a specific programming or state change step
- a switching mechanism such as a crossbar
- a memory sense mechanism e.g. a sense array unit
- the write data register for use in a multiple cycle state change operation, to modify data in that register and continue the operation as necessary via additional cycles
- the external bus in connection with a read of a memory page
- this operation may be implemented, perhaps the simplest of which is to provide a remote memory controller (e.g., a flash controller) with granularity into memory device operation, down to each state change step or cycle within memory device(s) in question.
- a remote memory controller e.g., a flash controller
- the controller can issue sequential, individual commands to: (a) move contents from a buffer into the write data register; (b) perform one or more erase cycles to reset contents of a storage unit (e.g., memory page) or group of storage units (e.g., memory block) at a specific address where an operation is to be performed; (c) copy contents from one storage unit into a buffer; (d) attempt to program or erase a memory unit (having a specific address) with contents of the write data register using a first voltage " VV" (e) compare the storage unit contents with contents of the write data register, with atomic update of differences into the contents of the register; (f) direct a subsequent data read to any storage unit location (other than the one subject to state change, or to a buffer holding data that is being currently written into a storage unit location); or (g) direct further state change and verification/comparison iterations at higher voltages.
- VV first voltage
- memory devices themselves (e.g., flash devices) to have internal structure and logic to support many of these operations. Considerations pertinent to each of these designs and their various permutations will be discussed below, together with design considerations pertinent to design of a memory device adapted to interact with several different masters (e.g., several different controllers or CPUs).
- masters e.g., several different controllers or CPUs.
- FIG. 4 presents a flowchart of a control method 401 that may be used to manage multiple, pipelined memory operations for a flash device that has several buffers, for example, as exemplified by the structures of FIGS. 2 and 3.
- the control method will generally track the usage of each buffer to determine whether the buffer is free or is in use, and to dynamically allocate unused buffers to new memory transactions.
- some memory transactions e.g., reads mixed with state change operations
- an earlier state change operation may still be ongoing, even after a later-initiated read of a memory page has been completed; in such a situation, a buffer later-assigned to the read operation may be released and available for reuse at an earlier time than a buffer used for the state change operation.
- the first available buffer may be assigned to a new operation "/,” with buffer assignment occurring on a round robin or absolute ordering basis.
- a status register may be used for control purposes with an entry for each available buffer and flags "F/' to indicate whether the particular buffer is free, is used, or is in the process of being loaded.
- the status may be stored locally to the controller, with the controller "knowing" whether one or more buffers are available; in embodiments where the memory device manages this process, the memory device may perform dynamic buffer assignment and may raise an error signal to the controller if all buffers are in use and the operation "/" cannot be queued.
- the memory device when the memory device is ready, it performs the memory operation "/,” be it a state change operation (such as a write or erase operation), a read, or some other transaction.
- FIG. 4 Two dashed-line function blocks 409 and 41 1 are illustrated in FIG. 4, to illustrate concurrent processing operations when two operations are presented at the same time. In this regard, it should be assumed for purposes of this discussion that a second memory operation "i+1" is received prior to the completion of the first memory operation "/.”
- controller may itself manage this process in implementations where the memory device supports machine commands or, alternatively, control logic in the memory device may break down requests from the controller (e.g., a read command, program command, or erase command) into individual steps, and control their administration.
- control logic in the memory device may break down requests from the controller (e.g., a read command, program command, or erase command) into individual steps, and control their administration.
- steps that typically need to be effectuated include charging the sense circuitry, controlling the read of a memory page and controlling the transfer of contents.
- the controller may need to be informed (in implementations where it does not manage each step in the process) that the page has been read and is available for upload and, accordingly, the memory device may need to signal the controller that the read data is available, such as by setting a status register that will be read by the controller, or raising a signal on a free/busy pin.
- Steps associated with an erase request and write request may be very similar to one another, in that multiple voltage pulses may be applied in iterations ("cycles"), with each cycle concluding with a comparison or "verify” step to determine the extent to which the cycle has successfully programmed or erased the memory cells in question.
- Each cycle may involve commands to charge circuitry to have a specific voltage, to gate voltage into specific circuitry (e.g., to control transistors that selectively apply voltage to column lines or to each memory cell gate), to sense cell contents, to perform an "exclusive- OR" operation, and so forth.
- Each program command may follow an erase of the memory storage unit that is to be the subject of the command (this is especially true for flash memory), in order to free up one or more blocks that will be used for programming.
- Each step associated with each operation type may be performed by sequential commands or hardware logic, and the structures and the related teachings provided by this disclosure permit the controller or device control logic to interlace multiple steps.
- multiple memory operations may be effectuated simultaneously for the disclosed structures by interleaving state change and read operations.
- reads of multiple pages may be pipelined, with contents for each page being stored in a separate buffer (if multiple buffers are available).
- the sensing mechanism e.g., the sense amplifier unit
- different circuitry, voltages and voltage paths may be used for read, write and erase commands, and it may be desired to close out on an iteration or cycle for each of these operations (e.g., for an entire page) before moving on to the next read.
- the write data register and its tie-in to the sensing mechanism typically provides a restriction on concurrent operations, meaning that operations are most easily interrupted between iterative cycles that collectively make up the state change operation. It is also possible to reapply the sensing mechanism in the array in between the programming and verify phases of each cycle but, typically, the same column and supporting circuitry will be used for each phase, so the most practical operation may be to interleave other operations between individual cycles that make up the state change operation.
- concurrent operations may be generally performed independently for each side, in a manner that will be further discussed below.
- communications between the IO interface and an array of buffers may be controlled independently of memory array control functions, and may be effectuated by a memory controller or by on-board control logic within the device.
- functions associated with this IO control logic include monitoring buffer availability, loading buffers with data from the external bus, updating a status register used for buffer control, fractioning an erase, program or read command into a number of constituent steps and queuing of those steps for the appropriate control logic, notifying a controller (if required for an embodiment) that read data is ready, or otherwise feeding read data out onto a bus for upstream transmission.
- FIG. 5 illustrates a method 501 for IO control logic.
- the method begins when a new memory operation is received, as indicated by a process start block 503; this box is labeled with the acronym "R/Pg(w/e)" to indicate that the operation may include a read command for a particular address or a state change operation, such as a write or erase command. Initially, it must be determined whether the memory device has capacity to initiate the transaction, as represented by a decision block 505. If there is at least one buffer available, the memory device may immediately receive the command and, if supported by the embodiment, it may still be possible for the device to receive and queue commands if immediate buffer assignment is not required.
- the memory device can be designed to receive and queue any number of read commands and associated memory addresses (up to cache limits) and can simply assign a buffer as it has bandwidth to support a new read operation. For state change operations, it may be desired (again if the device is charged with buffer assignment) for the device to raise an error flag (as indicated by process termination block 507) if no buffers are instantaneously available to process a desired operation.
- the memory device may be configured to toggle a logic level of an error detection pin, which would inform the controller that it should queue the additional transaction until memory is ready to process it.
- the memory device may be configured simply to set an internal free/busy status register flag, which the controller could be caused to read prior to commanding any new operation.
- the controller commands individual steps within the memory device, the controller would itself track buffer allocation and, so, would "know" when a new memory operation could be initiated. 1 . Handling Of New Transaction Requests.
- the method 501 then proceeds to determine whether a requested operation is a programming operation or a read command - read requests may be, depending on implementation, queued for processing without immediate buffer assignment.
- a programming request will typically counsel immediate buffer assignment and loading of data from an external bus.
- the controller may begin transmitting a new page of data to be programmed into memory with, or immediately subsequent to a program command, and so, a buffer may need to be quickly assigned in order for the device to receive this data.
- the method 501 may add the read request and associated memory address to a read request stack to be processed on a first-in, first-out basis, and the method may then terminate (as indicated by reference numerals 51 1 and 513).
- the IO control logic may perform steps of assigning a buffer, transferring program data into the assigned buffer, updating of a status register, and adding (once the buffer is loaded) suitable commands into a program operation stack, all as represented by reference numeral 515.
- FIG. 5 shows an exemplary status register (or state machine) 517 controlled in connection with buffer assignment.
- the register has a number of fields, including (1 ) an identity field for each buffer (labeled "B1," "B2,” and “B3, respectively, with a continuation arrow to indicate that more buffers may be provided for if desired), (2) a set of flags ("Fr) to indicate whether the particular register is (a) free or in-use for a transaction, or (b) busy being loaded or is stable, (3) an address of memory storage space to be associated with buffer contents and (4) for implementations where that support multiple masters, an identity associated with the particular master.
- the process may terminate until a new command is received from an external device (e.g., a memory controller), as indicated by numeral 519.
- this status register or state machine may be stored resident in each memory device, while in other embodiments, a flash controller (e.g., such as the controller 303 from FIG. 3) may maintain a status register for each memory device.
- individual commands issued from a controller would include a buffer identity or opcode field, as also represented by the buffer identity associated with reference numeral 517.
- a storage unit in the non-volatile memory may be effectively read notwithstanding that the unit (or data in the write data register) is in a state of transition. This may be effectively accomplished by leaving a copy of data (program data) in a buffer after that data is copied to the write data register, and using the status register (as discussed in connection with FIG. 5 above, for example) as a reference for incoming memory reads. Such an operation is illustrated for example in FIG. 6, which presents a control method 601 for reading data that is in the midst of being programmed.
- the IO control logic when the IO control logic detects an inbound read request, it compares the address represented by the request with the contents of the status register entry (e.g., the register 517 of FIG. 5). Logic can effectively compare entries for buffers in the order of most recently assigned buffer, and can stop with the first match (to ensure that it does not treat a prior erase request as a cache for reads to the associated memory space). Other options also exist.
- the read address matches the entry for any buffer, the match indicates that a copy of the data sought may already be stored in one of the buffers, and the flags of the associated entry in the status register are then examined to ensure that the associated data is not still currently being loaded into the buffer (i.e., that the data as represented in the buffer is accurate).
- Facilitating this functionality is one reason why a designer might wish to immediately assign a buffer for a new transaction instead of postponing buffer assignment to be contemporaneous with sensing of page contents from memory (as was alluded to above in connection with FIG. 5).
- immediate buffer assignment upon receipt of a controller request ensures that a subsequent read for the same page (instead of queuing a command with deferred buffer assignment) will necessarily trigger a match using the method of FIG. 6.
- the IO control logic may wait until the loading is finished but, once data loading is complete, the controller can be alerted that data is ready to be read.
- there are a number of mechanisms that may be used to perform this alert including toggling a "free/busy" pin, setting a status register that is periodically checked by the controller, sending a command to the controller, or some other mechanism.
- the controller and the memory device can trigger transfer of the data out of the buffer and onto the external bus using, for example, another controller-issued read command (or by different commands, e.g., triggering first a load of data into a buffer, followed by a read command directed to the buffer in question).
- these functions are variously represented by numerals 607, 609, 61 1 , and 613 in FIG. 6.
- the IO control logic determines that the requested memory page is not to be found in a local buffer, and it adds the command in question (and associated storage unit address) into a read stack for the memory array control logic as indicated by numeral 615 of FIG. 6.
- the method may then be terminated, as indicated by numeral 617.
- it may in connection with commanding this operation be desired to immediately assign an open buffer to a new read command (such that any subsequent read for the same page will trigger a "hit" in the local buffers and so be directed to read a copy of that buffer); this method of facilitating cache preparation, denoted by numeral 619, can be expected to be both faster than a read to the storage array and potentially more accurate, i.e., as the associated storage unit may be in the midst of loading when a subsequent memory request is received.
- buffer usage providing, for example, indicators of whether each buffer is free or in-use, and whether each buffer is also in the process of being loaded.
- status flags in a register may be used to indicate the following buffer states:
- buffer is in use and is currently stable, i.e., is not busy being loaded; whether or not a programming operation (i.e., write or erase) is ongoing, current contents associated with the particular memory address may be read directly from the buffer by the IO control logic without even involving the memory array control logic or the memory storage space; and
- buffer is in use and is currently being loaded; in this state, buffer contents may be inaccurate (e.g., partially valid, partially invalid data) and thus a memory operation (whether controlled as a step or command managed by the memory array control logic or an IO operation managed by the IO control logic should be queued and effectively "wait” until the associated flag is changed to indicate that loading is complete.
- this status data e.g., the data represented by register flags
- buffer identification and the page address corresponding to buffer contents By coupling this status data (e.g., the data represented by register flags) with buffer identification and the page address corresponding to buffer contents, most memory operations contemplated by this disclosure may be performed. For example, parallel operation processing may be effected during a programming operation by serving data directly from the buffers, essentially using these buffers as a form of local cache within the memory device.
- Certain steps performed by the IO control logic as well as the memory array control logic can be used to release a buffer for new uses as each memory operation is completed. For embodiments that support retaining a copy of program data in a buffer, so that that data can be read even while a multiple cycle program command is ongoing, an associated buffer can be released once the programming is verified as completely accurate (e.g., comparison between the write data register and the storage unit undergoing programming indicate that nothing remains to be programmed).
- the memory array control logic may simply change a register status bit (e.g., the first "Ff field for the pertinent buffer) to indicate that the IO control logic may now apply the buffer in question to a new transaction (i.e., because the data just programmed may now be reliably obtained from the memory storage space).
- buffers can similarly be recycled once a read operation is complete (e.g., by a controller, in implementations where the controller must pull read data, or by the IO control logic in implementations where the memory must affirmatively transfer the data).
- the buffer may also be used to indefinitely hold data if a programming operation does not complete correctly; for example, if a predetermined number of programming cycles (e.g., 20) do not "reset” or program a specific unit (e.g., page or block) correctly, an error signal can be raised to the controller, and the buffer holding the data can be used as temporary storage until the controller marks the storage unit in question as "bad” and reassigns the data to other memory space (either in the same device or a different non-volatile memory device).
- a predetermined number of programming cycles e.g. 20
- a specific unit e.g., page or block
- the memory device can poll its status registers to select and assign the first available buffer. By changing buffer status to "free" as soon as corresponding request is fulfilled, the control logic helps ensure that the device can immediately reapply freed buffers and accept new memory requests as soon as presented.
- the controller manage each individual step and operation within the non-volatile memory device(s), for example, the moving of data into and out of each buffer, and the operations involved in each individual state change cycle.
- management functions described above may, in a controller-managed environment, be implemented relatively easily via controller instructional logic, and the support for simplified commands is relatively "cheap" in terms of required device hardware and is straightforward to implement.
- a status register or state machine describing buffer usage and contents may be retained in the controller for each memory device; this structure if use renders it relatively easy for the controller, for example, to build commands associated with moving data into and out of each buffer, and of course, to monitor status.
- a controller designed according to these principles may rely upon command generation logic to generate commands to: transfer data from a sense amplifier unit to a specific buffer; transfer data from a specific buffer to the system data bus; write program data into a specific buffer; read data that is the subject of a program command from a specific buffer; copy data from a specific buffer into a write data register; or non-destructively copy data form a specific buffer into the write data register.
- Commands generated by the controller would in this event, as appropriate, generally specify a specific target buffer using an op-code as alluded to above.
- Other commands are of course possible, as will occur to those having skill in digital systems design.
- memory devices which implement the management logic and structures called for above may be designed to be compatible with existing memory controllers and operating systems. Memory devices designed in this manner may also be relatively easily adapted to a multiple master environment (e.g., as a component in a network or system that has multiple microprocessors); as a corollary, in this event, individual commands issued by a controller or other "master" may include a master identification field, or opcode, as further represented by numeral 517 in FIG. 5. Further, each controller will generally have a number of tasks that it is called upon to manage and, for at least some applications and to the extent that functions can be offloaded to each memory device without impacting overall speed, overall system efficiency can be improved.
- the memory array control logic is called upon to manage the flow of information between a set of buffers and the memory storage space, including the filling the write data register with data used to program memory pages, the transferring of verification data from the sense array (the sense amplifier unit) to the write data cache, and the moving of read data from the sense array to an assigned buffer.
- data associated with a programming operation may be retained in a buffer, and data for a subsequent program operation may be stored in another buffer and queued until earlier multiple cycle state change operations are completed; as a new page of memory is to be operated upon, the memory array control logic functions transfer data from the pertinent buffer into the write data register.
- the memory array control logic may be designed to effectively queue program and read operations in two different stacks, separating control of state change operation steps and a read operation steps. This differentiation, as will be seen below, represents one possible way of structuring device operations so that reads and state change operations may be automatically interleaved without significant impact on memory operations.
- FIG. 7 presents a logic flow diagram that illustrates a method 701 by which state change and read operations may be interleaved.
- read and state change steps may be separated into separate functional stacks or queues.
- the memory array control logic function operates in a wait mode while no unexecuted memory operations are required by either queue.
- the method 701 proceeds to execute a state change step, first, followed by a read step, second, as indicated by functional references 703, 705 and 707. Following this order of execution, the method returns to the starting block 703 and again operates in the same cycle, i.e., executing any queued state change step followed by any queued read operation step, to the extent that each are represented in the respective queues.
- each step may result in the effective insertion of additional steps in each functional queue, as well as the update of the status register for each affected buffer. For example, as previously mentioned, once a state change operation is finished (e.g., a storage unit is correctly programmed), the associated buffer may be released because complete, valid data may be read from non-volatile memory and the buffer is no longer needed. Put another way, each step may represent a single "cycle" of a multiple cycle state change operation (i.e., program or erase) or, in the case of reads, a transfer of a single memory page to a buffer. As part of the verify phase of the each program-verify cycle, if it is determined that data has correctly changed state, the operation can terminate and the buffer can be released.
- a state change operation i.e., program or erase
- the verify phase can insert a new command (i.e., a new cycle), effectively at the top of the programming queue.
- a new command i.e., a new cycle
- the method 701 of FIG. 7 is effectively looped back to perform another cycle.
- state change and read steps may be automatically interleaved, with a read (or several read operations if desired) occurring between each cycle of a multiple cycle state change operation.
- FIG. 8 provides a method 801 used to represent this logic flow just described, focusing on steps associated with a programming operation.
- the IO control logic queues a series of programming steps that are to program a specific memory address (e.g., a block or a page of memory, for example).
- a specific memory address e.g., a block or a page of memory, for example.
- an erase of the unit in question typically first occurs, corresponding to a reset of a block of data used to free up memory cells.
- An erase operation is performed (as indicated by numeral 803) for as many cycles as needed to ensure the "reset" of all memory cells associated with the block in question.
- each cycle of a multiple cycle programming operation may be viewed as a single process or step.
- the entire multiple step process may be managed so as to have any pending reads inserted in between erase cycles, in interleaved fashion.
- FIG. 9 provides a detailed flowchart of operations associated with each cycle of a state change operation.
- FIG. 9 depicts a method 901 that effectively begins when the program stack logic signals a pending program operation for a specific page, with an address fed to the write circuitry and associated data (i.e., specific data in the case of a program operation), as collectively represented by numeral 903 in FIG. 9.
- the method then invokes a program phase, identified by numeral 905, during which the write circuitry charges the memory storage space as necessary and attempts to program each memory cell corresponding to a page value that is a "zero" (i.e., this value is used to turn "on" the control gate associated with an affected memory cell).
- the first such program operation will use a default voltage value that is set by the write circuitry and represents a minimum programming voltage.
- the particular way in which a cell is programmed may differ if the cell is to be set (i.e., "programmed") or reset (i.e., "erased”).
- an erase may be applied for a block, depending on design, by charging a substrate to essentially remove any quantum tunneling effects created in the substrate, whereas a "set" typically may be effected by applying turning on a transistor, to thereby apply voltage to a control gate that will charge an associated floating gate for a memory cell.
- "set” and “reset” operations may involve different voltages.
- the particular way in which charges are scaled and routed is defined by the write circuitry in dependence upon settings established when data is initially transferred from the buffers to the write circuitry and the write data register.
- For erase operations proper erasure may be performed by detecting whether any memory cell in a block has not been correctly reset.
- the erase may be applied again, i.e., using what are essentially PV cycles to repeat the operation until all bits have been reset to a logic "1 .”
- the method then invokes a verify phase, which performs an exclusive-OR between the storage unit (i.e., the memory space being programmed) and the write data register and inverts the results. If there are no "zeros" represented in these results, then the page has been correctly programmed (or erased) and the programming operation may be terminated, with any associated buffer being released - these operations are collectively represented by numerals 907, 909, 91 1 and 913 in FIG. 9.
- the complete results are used to overwrite the write data register; the write data register will then have a "zero" for each memory page location (i.e., for each cell in the storage space) where prior state change attempts were unsuccessful, and the particular cycle (program and verify phases) is completed.
- the step ends with increasing " V” (i.e., by increasing the default voltage used by the write circuitry, such that the next state change attempt will use a higher voltage) and by adding another state change operation for the same memory page to the top of the processing queue.
- the operation also increases a value "k,” representing the number of iterative PV cycles already performed for the storage unit in question; this value, as alluded to earlier, is used to detect a "bad" memory block, i.e., should some predetermined limit (such as "20 cycles") be reached, the memory device may in this case signal an error to the controller and await further instruction for the data in question.
- k representing the number of iterative PV cycles already performed for the storage unit in question
- FIG. 10 illustrates one method 1001 by which a read operation may be accomplished in association with the methods, structures and logic discussed above.
- each read may begin with a configuration of the crossbar (if an embodiment that relies on a crossbar is used, e.g., such as the embodiment seen in FIGS. 2 and 3).
- this configuration function serves to route data from the sense amplifier unit to an assigned buffer that will serve as a temporary destination for page contents.
- the buffer status flag for an assigned buffer may changed to indicate that the assigned buffer is "busy" and that if read at this time, its data may be in a state of transition and may not be accurate.
- Circuitry associated with the storage space is also provided with address data for the particular unit (e.g., the particular page of data) that is to be read, and read data is then transferred to the assigned buffer, as indicated by numerals 1007 and 1009.
- the status register is again accessed (per numeral 101 1 ) in order to again change the buffer status flag to indicate that the buffer is no longer "busy” (i.e., that it is no longer loading data and consequently that its contents represent a valid page of data). The method then terminates, as indicated by numeral 1013.
- Read operations are generally simpler than the state change operations described earlier. While these operations may also involve multiple steps, the charging and voltage applications that occur for sensing operations are typically different than for state change operations just described. Generally, because of the need to switch different particular voltages onto potentially different signal paths for read, program and erase operations, it generally is preferred to complete each cycle (e.g., each write or erase cycle, or each read of an entire page) prior to interleaving another command "within” a read operation, so as to not change memory array control settings but, again, there may be countervailing motivations dependent upon the desired application.
- each cycle e.g., each write or erase cycle, or each read of an entire page
- the structures proposed by this disclosure facilitate faster non-volatile memory operation, because multiple operations can be sent to a memory device and commenced at least in part, without requiring completion of a prior operation.
- This concurrency can be applied to the same memory space, that is, to enable concurrent access by multiple memory operations to the cells served in common by a single write data register (or parallel group of such registers) and by a single sense array.
- the IO control logic may stack multiple operations without awaiting completion of a first operation.
- the program control logic may transfer data between the sense array (sense amplifier unit) and one buffer while the IO control logic transfers a second page of memory from a different buffer to an external data bus, for transmission to a controller.
- the use of a crossbar permits the program control logic to switchably-connect any buffer to the sense array, so that any buffer may be dynamically assigned as a temporary destination for read data.
- the filling of buffers associated with read requests may consequently be pipelined for quick execution. Multiple pages of memory may then be read out of a flash device in quick succession, without having to await the latency normally associated with per request activation and transfer of data from the sense array for each requested page of memory.
- the structures described above also permit the pipelining of multiple write or erase requests in much the same manner as just described.
- the IO control logic may initiate state change operations and assign buffers to write (program) requests, such that multiple pages of data to be written into nonvolatile memory can be loaded in quick succession into different buffers; while the memory device is busy with one state change operation, a page of data for a subsequent write operation can be queued up, notwithstanding the busy status of write circuitry and sense array associated with the memory space. That is to say, the memory devices introduced above generally do not need to wait for full completion of a state change operation (and associated representation of "busy" status to a controller), but can proceed to accept new state change requests for additional pages or blocks of memory even while a multiple cycle operation is ongoing.
- the operations just described may also be extended to simultaneous read operations and write (program) or erase operations. Specifically, while one page of memory is being programmed, instead of maintaining the device as "busy,” one or more pages of memory can be read by streaming data (already loaded into data buffers) out over the external bus to a controller, notwithstanding that write circuitry and sense circuitry are "busy.” Because these operations can take multiple cycles, the logic described above permits the interleave of read operations, such that a multiple cycle write or erase operation can be interrupted, to allow program control logic to quickly transfer a page of data into an assigned buffer and then resume state change operations, all without disturbing write circuitry settings or write data register contents.
- the IO control logic then transfers that page of memory out over the external bus to satisfy a controller read request.
- the memory device may be configured for use as a slave of a single master (e.g., a single controller) or for multiple devices (i.e., by providing structure to permit each memory device to identify the master requesting each read and notifying that master when requested data is ready). Any amount of concurrency can be supported, depending on the number of buffers designed into a given memory implementation.
- This disclosure has presented embodiments of a non-volatile device and related systems and methods of operation. More particularly, using a device that separates paths associated with loading a write data register, outputs read data from a sense amplifier unit, and feeds data from the sense amplifier unit to update the write data registers, one may in-parallel perform concurrent processing within the same memory organization (e.g., within the same memory bank). In some embodiments, multiple buffers may be each dynamically assigned to a memory page this is to be read or programmed, providing for parallel processing of transactions and permitting one to read data that is currently amidst a program operation. Other embodiments, applications and permutations of these structures and methods exist.
- these structures potentially enable a much more widespread usage of non-volatile devices; by reducing latency for these devices, the embodiments discussed above potentially permit these devices to be faster, making them a more practical choice for applications where speed is at issue. Because some applications may be constrained by power as well as program and access speed considerations, these teachings may enable applications previously thought impractical, e.g., by permitting faster non-volatile read, programming and erasing in portable and mobile devices.
- certain functions may be tracked by a memory system controller, such as a flash or other non-volatile controller.
- the controller may be vested with responsibility for waiting for read data, or for "pinging" a status register and associated buffer to retrieve data corresponding to a queue read request.
- some of these functions may be instead made resident in a memory device; for example, especially for embodiments where one memory device can interact with multiple processors, it may be desired to have the memory device dynamically assign specific buffers to specific transactions (e.g., a read operation) and inform the associated controller of the assigned buffer.
- the buffer assignment can be made transparent to the controller, with cross reference between individual memory requests and associated buffers being made purely within the memory device. Other variations of these principles may also occur to those having skill in the art.
Abstract
Description
Claims
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2009801492759A CN102246240A (en) | 2008-12-09 | 2009-10-15 | Non-volatile memory device for concurrent and pipelined memory operations |
EP09836571.1A EP2377129A4 (en) | 2008-12-09 | 2009-10-15 | Non-volatile memory device for concurrent and pipelined memory operations |
US13/126,726 US8645617B2 (en) | 2008-12-09 | 2009-10-15 | Memory device for concurrent and pipelined memory operations |
JP2011539539A JP2012511789A (en) | 2008-12-09 | 2009-10-15 | Non-volatile memory device for parallel and pipelined memory operation |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US12108308P | 2008-12-09 | 2008-12-09 | |
US61/121,083 | 2008-12-09 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2010077414A1 true WO2010077414A1 (en) | 2010-07-08 |
Family
ID=42310089
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/US2009/060911 WO2010077414A1 (en) | 2008-12-09 | 2009-10-15 | Non-volatile memory device for concurrent and pipelined memory operations |
Country Status (6)
Country | Link |
---|---|
US (1) | US8645617B2 (en) |
EP (1) | EP2377129A4 (en) |
JP (1) | JP2012511789A (en) |
KR (1) | KR20110110106A (en) |
CN (1) | CN102246240A (en) |
WO (1) | WO2010077414A1 (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012119018A (en) * | 2010-11-30 | 2012-06-21 | Hitachi Ltd | Semiconductor device and information processing system |
CN113741799A (en) * | 2020-05-27 | 2021-12-03 | 旺宏电子股份有限公司 | Memory device |
Families Citing this family (82)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100178038A1 (en) * | 2009-01-12 | 2010-07-15 | Mediatek Inc. | Video player |
US20110320863A1 (en) * | 2010-06-24 | 2011-12-29 | International Business Machines Corporation | Dynamic re-allocation of cache buffer slots |
US9104583B2 (en) | 2010-06-24 | 2015-08-11 | International Business Machines Corporation | On demand allocation of cache buffer slots |
US8635407B2 (en) | 2011-09-30 | 2014-01-21 | International Business Machines Corporation | Direct memory address for solid-state drives |
US8625386B2 (en) * | 2012-03-22 | 2014-01-07 | Seiko Epson Corporation | Non-volatile memory device, circuit board, printing material container and printer |
CN103426470B (en) * | 2012-05-21 | 2016-04-20 | 北京兆易创新科技股份有限公司 | Improve the method for the page program speed of flash memory |
US20130318285A1 (en) * | 2012-05-23 | 2013-11-28 | Violin Memory Inc | Flash memory controller |
KR101919903B1 (en) * | 2012-09-14 | 2018-11-19 | 삼성전자 주식회사 | EMBEDDED MULTIMEDIA CARD(eMMC), HOST FOR CONTROLLING THE eMMC, AND METHODS FOR OPERATING THE eMMC AND THE HOST |
KR101988260B1 (en) * | 2012-09-14 | 2019-06-12 | 삼성전자주식회사 | EMBEDDED MULTIMEDIA CARD(eMMC), AND METHOD FOR OPERATING THE eMMC |
KR101932920B1 (en) | 2012-09-14 | 2019-03-18 | 삼성전자 주식회사 | Host for controlling non-volatile memory crad, system including the same and operating method there-of |
KR20140076128A (en) * | 2012-12-12 | 2014-06-20 | 에스케이하이닉스 주식회사 | Non-Volatile Memory Apparatus and Operating Method Thereof, and Data Processing System Having the Same |
WO2014098839A1 (en) * | 2012-12-19 | 2014-06-26 | Hewlett-Packard Development Company | Nvram path selection |
US9652376B2 (en) * | 2013-01-28 | 2017-05-16 | Radian Memory Systems, Inc. | Cooperative flash memory control |
JP5853973B2 (en) * | 2013-03-07 | 2016-02-09 | ソニー株式会社 | Storage control device, storage device, information processing system, and storage control method |
KR20140142960A (en) * | 2013-06-05 | 2014-12-15 | 중소기업은행 | SPI NAND Flash memory with parallel pipelined double latch |
US9923762B1 (en) * | 2013-08-13 | 2018-03-20 | Ca, Inc. | Upgrading an engine when a scenario is running |
US20150095551A1 (en) * | 2013-09-30 | 2015-04-02 | Micron Technology, Inc. | Volatile memory architecutre in non-volatile memory devices and related controllers |
KR101455682B1 (en) * | 2014-02-18 | 2014-11-04 | 연세대학교 산학협력단 | Hash partitioning apparatus and method |
US9348518B2 (en) | 2014-07-02 | 2016-05-24 | International Business Machines Corporation | Buffered automated flash controller connected directly to processor memory bus |
US9542284B2 (en) | 2014-08-06 | 2017-01-10 | International Business Machines Corporation | Buffered automated flash controller connected directly to processor memory bus |
EP3268864B1 (en) * | 2015-03-11 | 2021-01-27 | Rambus Inc. | High performance non-volatile memory module |
JP5889462B2 (en) * | 2015-03-26 | 2016-03-22 | 株式会社日立製作所 | Method for controlling semiconductor device and nonvolatile memory device |
CN106155573B (en) * | 2015-04-17 | 2019-12-17 | 伊姆西公司 | method and device for expanding storage device and expanded storage device |
KR102356071B1 (en) * | 2015-05-06 | 2022-01-27 | 에스케이하이닉스 주식회사 | Storage device and operating method thereof |
US10121553B2 (en) | 2015-09-30 | 2018-11-06 | Sunrise Memory Corporation | Capacitive-coupled non-volatile thin-film transistor NOR strings in three-dimensional arrays |
US9892800B2 (en) | 2015-09-30 | 2018-02-13 | Sunrise Memory Corporation | Multi-gate NOR flash thin-film transistor strings arranged in stacked horizontal active strips with vertical control gates |
US9842651B2 (en) | 2015-11-25 | 2017-12-12 | Sunrise Memory Corporation | Three-dimensional vertical NOR flash thin film transistor strings |
US11120884B2 (en) | 2015-09-30 | 2021-09-14 | Sunrise Memory Corporation | Implementing logic function and generating analog signals using NOR memory strings |
US10467157B2 (en) | 2015-12-16 | 2019-11-05 | Rambus Inc. | Deterministic operation of storage class memory |
US10254967B2 (en) | 2016-01-13 | 2019-04-09 | Sandisk Technologies Llc | Data path control for non-volatile memory |
US10579290B2 (en) * | 2016-03-23 | 2020-03-03 | Winbond Electronics Corp. | Option code providing circuit and providing method thereof |
WO2018080783A1 (en) * | 2016-10-31 | 2018-05-03 | Rambus Inc. | Hybrid memory module |
US10528286B2 (en) | 2016-11-11 | 2020-01-07 | Sandisk Technologies Llc | Interface for non-volatile memory |
US10528267B2 (en) | 2016-11-11 | 2020-01-07 | Sandisk Technologies Llc | Command queue for storage operations |
US10528255B2 (en) | 2016-11-11 | 2020-01-07 | Sandisk Technologies Llc | Interface for non-volatile memory |
US10114589B2 (en) * | 2016-11-16 | 2018-10-30 | Sandisk Technologies Llc | Command control for multi-core non-volatile memory |
CN108074617A (en) * | 2016-11-18 | 2018-05-25 | 中芯国际集成电路制造(上海)有限公司 | A kind of nonvolatile memory |
KR20180063419A (en) * | 2016-12-01 | 2018-06-12 | 삼성전자주식회사 | Storage device configured to perform two-way communication with host and operating method thereof |
US10452278B2 (en) * | 2017-03-24 | 2019-10-22 | Western Digital Technologies, Inc. | System and method for adaptive early completion posting using controller memory buffer |
US10509569B2 (en) | 2017-03-24 | 2019-12-17 | Western Digital Technologies, Inc. | System and method for adaptive command fetch aggregation |
DE112018000230T5 (en) * | 2017-03-24 | 2019-09-05 | Western Digital Technologies, Inc | System and method for speculative instruction execution using the control memory buffer |
JP7203054B2 (en) | 2017-06-20 | 2023-01-12 | サンライズ メモリー コーポレイション | Three-dimensional NOR-type memory array architecture and method of manufacturing the same |
US11180861B2 (en) | 2017-06-20 | 2021-11-23 | Sunrise Memory Corporation | 3-dimensional NOR string arrays in segmented stacks |
US10692874B2 (en) | 2017-06-20 | 2020-06-23 | Sunrise Memory Corporation | 3-dimensional NOR string arrays in segmented stacks |
US10608008B2 (en) | 2017-06-20 | 2020-03-31 | Sunrise Memory Corporation | 3-dimensional nor strings with segmented shared source regions |
WO2019074743A1 (en) | 2017-10-12 | 2019-04-18 | Rambus Inc. | Nonvolatile physical memory with dram cache |
US10896916B2 (en) | 2017-11-17 | 2021-01-19 | Sunrise Memory Corporation | Reverse memory cell |
JP6967959B2 (en) * | 2017-12-08 | 2021-11-17 | キオクシア株式会社 | Memory system and control method |
CN107978337B (en) * | 2017-12-21 | 2020-12-04 | 上海华力微电子有限公司 | High-speed circuit structure suitable for automatic test of random access memory and test method thereof |
US10402121B2 (en) | 2017-12-21 | 2019-09-03 | Apple Inc. | Systems and methods for reducing performance state change latency |
US10475812B2 (en) | 2018-02-02 | 2019-11-12 | Sunrise Memory Corporation | Three-dimensional vertical NOR flash thin-film transistor strings |
KR102052816B1 (en) * | 2018-03-15 | 2019-12-05 | 한양대학교 에리카산학협력단 | Method for recovering error in non-volatile main memory unit and system therefor |
US10942854B2 (en) | 2018-05-09 | 2021-03-09 | Micron Technology, Inc. | Prefetch management for memory |
US10714159B2 (en) | 2018-05-09 | 2020-07-14 | Micron Technology, Inc. | Indication in memory system or sub-system of latency associated with performing an access command |
US11010092B2 (en) * | 2018-05-09 | 2021-05-18 | Micron Technology, Inc. | Prefetch signaling in memory system or sub-system |
US10754578B2 (en) | 2018-05-09 | 2020-08-25 | Micron Technology, Inc. | Memory buffer management and bypass |
CN110554971A (en) * | 2018-06-01 | 2019-12-10 | 北京兆易创新科技股份有限公司 | Method, device and storage medium for responding to read operation during programming |
US11069696B2 (en) | 2018-07-12 | 2021-07-20 | Sunrise Memory Corporation | Device structure for a 3-dimensional NOR memory array and methods for improved erase operations applied thereto |
US11751391B2 (en) | 2018-07-12 | 2023-09-05 | Sunrise Memory Corporation | Methods for fabricating a 3-dimensional memory structure of nor memory strings |
WO2020014655A1 (en) | 2018-07-12 | 2020-01-16 | Sunrise Memory Corporation | Fabrication method for a 3-dimensional nor memory array |
TWI713195B (en) | 2018-09-24 | 2020-12-11 | 美商森恩萊斯記憶體公司 | Wafer bonding in fabrication of 3-dimensional nor memory circuits and integrated circuit formed therefrom |
KR20200048318A (en) * | 2018-10-29 | 2020-05-08 | 에스케이하이닉스 주식회사 | Storage device and operating method thereof |
WO2020118301A1 (en) | 2018-12-07 | 2020-06-11 | Sunrise Memory Corporation | Methods for forming multi-layer vertical nor-type memory string arrays |
US11670620B2 (en) | 2019-01-30 | 2023-06-06 | Sunrise Memory Corporation | Device with embedded high-bandwidth, high-capacity memory using wafer bonding |
EP3925004A4 (en) | 2019-02-11 | 2023-03-08 | Sunrise Memory Corporation | Vertical thin-film transistor and application as bit-line connector for 3-dimensional memory arrays |
TWI747369B (en) | 2019-07-09 | 2021-11-21 | 美商森恩萊斯記憶體公司 | Process for a 3-dimensional array of horizontal nor-type memory strings |
US11917821B2 (en) | 2019-07-09 | 2024-02-27 | Sunrise Memory Corporation | Process for a 3-dimensional array of horizontal nor-type memory strings |
US20210150311A1 (en) * | 2019-11-19 | 2021-05-20 | Alibaba Group Holding Limited | Data layout conscious processing in memory architecture for executing neural network model |
WO2021127218A1 (en) | 2019-12-19 | 2021-06-24 | Sunrise Memory Corporation | Process for preparing a channel region of a thin-film transistor |
US11138102B2 (en) * | 2019-12-24 | 2021-10-05 | Intel Corporation | Read quality of service for non-volatile memory |
WO2021159028A1 (en) | 2020-02-07 | 2021-08-12 | Sunrise Memory Corporation | High capacity memory circuit with low effective latency |
EP4100839A4 (en) | 2020-02-07 | 2024-03-13 | Sunrise Memory Corp | Quasi-volatile system-level memory |
US11508693B2 (en) | 2020-02-24 | 2022-11-22 | Sunrise Memory Corporation | High capacity memory module including wafer-section memory circuit |
WO2021173572A1 (en) | 2020-02-24 | 2021-09-02 | Sunrise Memory Corporation | Channel controller for shared memory access |
US11507301B2 (en) | 2020-02-24 | 2022-11-22 | Sunrise Memory Corporation | Memory module implementing memory centric architecture |
US11705496B2 (en) | 2020-04-08 | 2023-07-18 | Sunrise Memory Corporation | Charge-trapping layer with optimized number of charge-trapping sites for fast program and erase of a memory cell in a 3-dimensional NOR memory string array |
US11937424B2 (en) | 2020-08-31 | 2024-03-19 | Sunrise Memory Corporation | Thin-film storage transistors in a 3-dimensional array of nor memory strings and process for fabricating the same |
WO2022108848A1 (en) | 2020-11-17 | 2022-05-27 | Sunrise Memory Corporation | Methods for reducing disturb errors by refreshing data alongside programming or erase operations |
US11848056B2 (en) | 2020-12-08 | 2023-12-19 | Sunrise Memory Corporation | Quasi-volatile memory with enhanced sense amplifier operation |
US11662942B2 (en) * | 2021-03-16 | 2023-05-30 | Kioxia Corporation | Explicit buffer control |
TW202310429A (en) | 2021-07-16 | 2023-03-01 | 美商日升存儲公司 | 3-dimensional memory string array of thin-film ferroelectric transistors |
US11604732B1 (en) * | 2021-09-02 | 2023-03-14 | Micron Technology, Inc. | Memory performance during program suspend protocol |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020031007A1 (en) * | 1996-08-06 | 2002-03-14 | Kenichi Osada | Semiconductor memory apparatus, semiconductor apparatus, data processing apparatus and computer system |
US20030076719A1 (en) * | 2001-10-24 | 2003-04-24 | Samsung Electronics Co., Ltd. | Non-volatile semiconductor memory device with cache function and program, read, and page copy-back operations thereof |
US20040160829A1 (en) * | 1998-02-16 | 2004-08-19 | Tetsuya Tsujikawa | Semiconductor, memory card, and data processing system |
US20040165473A1 (en) * | 2000-06-30 | 2004-08-26 | Micron Technology, Inc. | Zero latency-zero bus turnaround synchronous flash memory |
US20050141298A1 (en) * | 2002-07-05 | 2005-06-30 | Aplus Flash Technology, Inc. | Combination nonvolatile memory using unified technology with byte, page and block write and simultaneous read and write operations |
Family Cites Families (61)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US3562554A (en) * | 1968-01-15 | 1971-02-09 | Ibm | Bipolar sense amplifier with noise rejection |
US4713797A (en) * | 1985-11-25 | 1987-12-15 | Motorola Inc. | Current mirror sense amplifier for a non-volatile memory |
US8027194B2 (en) * | 1988-06-13 | 2011-09-27 | Samsung Electronics Co., Ltd. | Memory system and method of accessing a semiconductor memory device |
US5282271A (en) * | 1991-10-30 | 1994-01-25 | I-Cube Design Systems, Inc. | I/O buffering system to a programmable switching apparatus |
JPH05274860A (en) * | 1992-03-26 | 1993-10-22 | Nec Corp | Semiconductor memory |
JPH05290585A (en) * | 1992-04-13 | 1993-11-05 | Hitachi Ltd | Nonvolatile memory writable electrically |
US5509134A (en) * | 1993-06-30 | 1996-04-16 | Intel Corporation | Method and apparatus for execution of operations in a flash memory array |
US5519847A (en) * | 1993-06-30 | 1996-05-21 | Intel Corporation | Method of pipelining sequential writes in a flash memory |
JP2922116B2 (en) * | 1993-09-02 | 1999-07-19 | 株式会社東芝 | Semiconductor storage device |
JPH07160569A (en) * | 1993-12-09 | 1995-06-23 | Matsushita Electric Ind Co Ltd | Memory controller |
US5570040A (en) * | 1995-03-22 | 1996-10-29 | Altera Corporation | Programmable logic array integrated circuit incorporating a first-in first-out memory |
US5717871A (en) * | 1995-08-17 | 1998-02-10 | I-Cube, Inc. | Crossbar switch with input/output buffers having multiplexed control inputs |
US5855343A (en) * | 1997-03-07 | 1999-01-05 | Irez Research, Corporation | Camera clip |
EP1981033B1 (en) * | 1997-10-10 | 2011-08-24 | Rambus Inc. | Apparatus and method for pipelined memory operations with write mask |
JPH11176174A (en) * | 1997-12-05 | 1999-07-02 | Sharp Corp | Writing circuit of non-volatile memory and non-volatile memory |
US5982700A (en) * | 1998-05-21 | 1999-11-09 | Integrated Device Technology, Inc. | Buffer memory arrays having nonlinear columns for providing parallel data access capability and methods of operating same |
US6370619B1 (en) * | 1998-06-22 | 2002-04-09 | Oracle Corporation | Managing partitioned cache |
CA2288451A1 (en) * | 1998-11-09 | 2000-05-09 | Seiji Muramoto | Video camera |
US6377502B1 (en) * | 1999-05-10 | 2002-04-23 | Kabushiki Kaisha Toshiba | Semiconductor device that enables simultaneous read and write/erase operation |
JP4439530B2 (en) * | 1999-05-10 | 2010-03-24 | 株式会社東芝 | Semiconductor device |
JP3440032B2 (en) * | 1999-07-15 | 2003-08-25 | パナソニック コミュニケーションズ株式会社 | Memory control device, facsimile device and image forming device |
US6240040B1 (en) * | 2000-03-15 | 2001-05-29 | Advanced Micro Devices, Inc. | Multiple bank simultaneous operation for a flash memory |
US6615307B1 (en) * | 2000-05-10 | 2003-09-02 | Micron Technology, Inc. | Flash with consistent latency for read operations |
US6851026B1 (en) * | 2000-07-28 | 2005-02-01 | Micron Technology, Inc. | Synchronous flash memory with concurrent write and read operation |
JP2001344986A (en) * | 2000-06-05 | 2001-12-14 | Mitsubishi Electric Corp | Non-volatile semiconductor memory |
US6442076B1 (en) * | 2000-06-30 | 2002-08-27 | Micron Technology, Inc. | Flash memory with multiple status reading capability |
US6278654B1 (en) * | 2000-06-30 | 2001-08-21 | Micron Technology, Inc. | Active terminate command in synchronous flash memory |
US6883044B1 (en) * | 2000-07-28 | 2005-04-19 | Micron Technology, Inc. | Synchronous flash memory with simultaneous access to one or more banks |
US7046633B2 (en) * | 2000-09-21 | 2006-05-16 | Avici Systems, Inc. | Router implemented with a gamma graph interconnection network |
GB2369694B (en) * | 2000-11-29 | 2002-10-16 | Sun Microsystems Inc | Efficient memory modification tracking |
US6889304B2 (en) * | 2001-02-28 | 2005-05-03 | Rambus Inc. | Memory device supporting a dynamically configurable core organization |
US6882562B2 (en) * | 2001-11-01 | 2005-04-19 | Agilent Technologies, Inc. | Method and apparatus for providing pseudo 2-port RAM functionality using a 1-port memory cell |
US6751129B1 (en) * | 2002-05-21 | 2004-06-15 | Sandisk Corporation | Efficient read, write methods for multi-state memory |
US7031192B1 (en) * | 2002-11-08 | 2006-04-18 | Halo Lsi, Inc. | Non-volatile semiconductor memory and driving method |
JP4331966B2 (en) * | 2003-04-14 | 2009-09-16 | 株式会社ルネサステクノロジ | Semiconductor integrated circuit |
US7155562B2 (en) * | 2003-05-08 | 2006-12-26 | Micron Technology, Inc. | Method for reading while writing to a single partition flash memory |
JP2004348790A (en) * | 2003-05-20 | 2004-12-09 | Sharp Corp | Semiconductor storage device and portable electronic apparatus |
JP4325275B2 (en) * | 2003-05-28 | 2009-09-02 | 株式会社日立製作所 | Semiconductor device |
JP4012144B2 (en) * | 2003-12-25 | 2007-11-21 | 株式会社東芝 | Semiconductor memory device |
US7173863B2 (en) * | 2004-03-08 | 2007-02-06 | Sandisk Corporation | Flash controller cache architecture |
US20050251617A1 (en) * | 2004-05-07 | 2005-11-10 | Sinclair Alan W | Hybrid non-volatile memory system |
TWI273766B (en) * | 2004-03-31 | 2007-02-11 | Wistron Neweb Corp | Lnbf |
JP4253272B2 (en) * | 2004-05-27 | 2009-04-08 | 株式会社東芝 | Memory card, semiconductor device, and control method of semiconductor memory |
US7048454B2 (en) * | 2004-05-28 | 2006-05-23 | Creative Technology Ltd | Three way video camera base |
US7142477B1 (en) * | 2004-06-18 | 2006-11-28 | Cypress Semiconductor Corp. | Memory interface system and method for reducing cycle time of sequential read and write accesses using separate address and data buses |
DE112004002927B4 (en) * | 2004-07-30 | 2011-07-21 | Spansion LLC (n.d.Ges.d. Staates Delaware), Calif. | Semiconductor device and method for writing the same |
US7572073B2 (en) * | 2004-10-15 | 2009-08-11 | Lifesize Communications, Inc. | Camera support mechanism |
US7299313B2 (en) * | 2004-10-29 | 2007-11-20 | International Business Machines Corporation | System, method and storage medium for a memory subsystem command interface |
US7071731B1 (en) * | 2005-01-21 | 2006-07-04 | Altera Corporation | Programmable Logic with Pipelined Memory Operation |
US7206230B2 (en) * | 2005-04-01 | 2007-04-17 | Sandisk Corporation | Use of data latches in cache operations of non-volatile memories |
US7431253B2 (en) * | 2005-06-03 | 2008-10-07 | Kye Systems Corp. | Support device for computer peripheral equipment |
JP4989872B2 (en) * | 2005-10-13 | 2012-08-01 | ルネサスエレクトロニクス株式会社 | Semiconductor memory device and arithmetic processing unit |
US7793068B2 (en) * | 2005-12-21 | 2010-09-07 | Sandisk Corporation | Dual mode access for non-volatile storage devices |
JP2007179669A (en) * | 2005-12-28 | 2007-07-12 | Toshiba Corp | Memory system |
KR100706816B1 (en) * | 2006-03-10 | 2007-04-12 | 삼성전자주식회사 | Nonvolatile memory device and program method thereof capable of improving program speed |
US7931243B2 (en) * | 2006-04-07 | 2011-04-26 | Logitech Europe S.A. | Speaker with monitor mountable stand |
US7280398B1 (en) * | 2006-08-31 | 2007-10-09 | Micron Technology, Inc. | System and memory for sequential multi-plane page memory operations |
TW200820266A (en) * | 2006-10-19 | 2008-05-01 | Benq Corp | Method for reading and writing data in a flash memory in an embedded system |
TW200917005A (en) * | 2007-10-12 | 2009-04-16 | Kye Systems Corp | Structure of supporting base |
CN102027455A (en) * | 2008-05-13 | 2011-04-20 | 拉姆伯斯公司 | Fractional program commands for memory devices |
US7775486B2 (en) * | 2008-10-04 | 2010-08-17 | Hewlett-Packard Development Company, L.P. | Webcam assembly with clamp and stand forms |
-
2009
- 2009-10-15 CN CN2009801492759A patent/CN102246240A/en active Pending
- 2009-10-15 WO PCT/US2009/060911 patent/WO2010077414A1/en active Application Filing
- 2009-10-15 US US13/126,726 patent/US8645617B2/en active Active
- 2009-10-15 KR KR1020117013129A patent/KR20110110106A/en not_active Application Discontinuation
- 2009-10-15 EP EP09836571.1A patent/EP2377129A4/en not_active Withdrawn
- 2009-10-15 JP JP2011539539A patent/JP2012511789A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020031007A1 (en) * | 1996-08-06 | 2002-03-14 | Kenichi Osada | Semiconductor memory apparatus, semiconductor apparatus, data processing apparatus and computer system |
US20040160829A1 (en) * | 1998-02-16 | 2004-08-19 | Tetsuya Tsujikawa | Semiconductor, memory card, and data processing system |
US20040165473A1 (en) * | 2000-06-30 | 2004-08-26 | Micron Technology, Inc. | Zero latency-zero bus turnaround synchronous flash memory |
US20030076719A1 (en) * | 2001-10-24 | 2003-04-24 | Samsung Electronics Co., Ltd. | Non-volatile semiconductor memory device with cache function and program, read, and page copy-back operations thereof |
US20050141298A1 (en) * | 2002-07-05 | 2005-06-30 | Aplus Flash Technology, Inc. | Combination nonvolatile memory using unified technology with byte, page and block write and simultaneous read and write operations |
Non-Patent Citations (1)
Title |
---|
See also references of EP2377129A4 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012119018A (en) * | 2010-11-30 | 2012-06-21 | Hitachi Ltd | Semiconductor device and information processing system |
US8773919B2 (en) | 2010-11-30 | 2014-07-08 | Hitachi, Ltd. | Semiconductor device and data processing system |
CN113741799A (en) * | 2020-05-27 | 2021-12-03 | 旺宏电子股份有限公司 | Memory device |
Also Published As
Publication number | Publication date |
---|---|
US20110208905A1 (en) | 2011-08-25 |
US8645617B2 (en) | 2014-02-04 |
EP2377129A4 (en) | 2013-05-22 |
JP2012511789A (en) | 2012-05-24 |
EP2377129A1 (en) | 2011-10-19 |
KR20110110106A (en) | 2011-10-06 |
CN102246240A (en) | 2011-11-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8645617B2 (en) | Memory device for concurrent and pipelined memory operations | |
US20210174875A1 (en) | Fractional program commands for memory devices | |
US7243185B2 (en) | Flash memory system with a high-speed flash controller | |
USRE49921E1 (en) | Memory device and controlling method of the same | |
US8924627B2 (en) | Flash memory device comprising host interface for processing a multi-command descriptor block in order to exploit concurrency | |
US10079048B2 (en) | Adjusting access of non-volatile semiconductor memory based on access time | |
US8041882B2 (en) | Method and device for writing to a flash memory | |
KR102514772B1 (en) | Nonvolatile memory device capable of performing asynchronous operations, nonvolatile memory system including the same, and operation performing method of the same | |
US9495101B2 (en) | Methods for balancing write operations of SLC blocks in different memory areas and apparatus implementing the same | |
US20210374060A1 (en) | Timed Data Transfer between a Host System and a Memory Sub-System | |
US20200142819A1 (en) | Verify before program resume for memory devices | |
US10503438B1 (en) | Memory sub-system supporting non-deterministic commands | |
US11294820B2 (en) | Management of programming mode transitions to accommodate a constant size of data transfer between a host system and a memory sub-system | |
US11314669B2 (en) | Deterministic operation of storage class memory | |
US9971546B2 (en) | Methods for scheduling read and write commands and apparatuses using the same | |
US20230114146A1 (en) | Command sequencing for read operations by solid-state drives | |
KR102242957B1 (en) | High speed NAND memory system and high speed NAND memory package device | |
CN115237824A (en) | Cache read context switching in a memory subsystem |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 200980149275.9 Country of ref document: CN |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 09836571 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2011539539 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13126726 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2009836571 Country of ref document: EP |
|
ENP | Entry into the national phase |
Ref document number: 20117013129 Country of ref document: KR Kind code of ref document: A |
|
NENP | Non-entry into the national phase |
Ref country code: DE |