Abstract: Methods, systems, and devices are described herein for using codewords to detect or correct errors in data (e.g., data stored in a memory device). A host device may generate one or more codewords associated with data to be stored in the memory device. In some cases, the host device may generate one or more codewords for error detection and correction (e.g., corresponding to data transmitted by the host device to the memory device). In some cases, the host device may transmit the codewords and the associated data using an extended (e.g., adjustable) burst length such that the one or more codewords may be included in the burst along with the data. Additionally or alternatively, the host device may transmit one or more of the codewords over one or more channels different than the one or more channels used to transmit the data.

Abstract: Certain aspects of the present disclosure provide techniques for performing compute in memory (CIM) computations. A device comprises a CIM module configured to apply a plurality of analog weights to data using multiply-accumulate operations to generate an output. The device further comprises a digital weight storage unit configured to store digital weight references, wherein a digital weight reference corresponds to an analog weight of the plurality of analog weights. The device also comprises a device controller configured to program the plurality of analog weights to the CIM module based on the digital weight references and determine degradation of one or more analog weights. The digital weight references in the digital weight storage unit are populated with values from a host device. Degraded analog weights in the CIM module are replaced with corresponding digital weight references from the digital weight storage unit without reference to the host device.

Abstract: Provided is a computing system including a memory system in communication with a host, and for storing data therein and the memory system includes a memory having a plurality of memory components and a memory array and coupled to the controller via a memory interface. Each memory component includes a memory cyclic-redundancy-check (CRC) engine that performs a CRC check of data during read and write operations between the host and the memory array. The memory system also includes a controller that has a plurality of controller CRC engines and communicates with the memory components to control data transmission between the memory, the host and the memory array.

Abstract: Recurrent Neural Networks (RNNs) wherein a non-volatile memory (NVM) array provides a memory bank for the RNN. The RNN may include a Neural Turning Machine (NTM) and the memory bank may be an NTM matrix stored in the NVM array. In some examples, a data storage device (DSD) that controls the NVM array includes both a data storage controller and a separate NTM controller. The separate NTM controller accesses the NTM matrix of the NVM array directly while bypassing flash translation layer (FTL) components of the data storage controller. Additionally, various majority wins error detection and correction procedures are described, as well as various disparity count-based procedures.

Abstract: A memory device that performs internal ECC (error checking and correction) can treat an N-bit channel as two N/2-bit channels for application of ECC. The memory device includes a memory array to store data and prefetches data bits and error checking and correction (ECC) bits from the memory array for a memory access operation. The memory device includes internal ECC hardware to apply ECC, with a first group of a first half the data bits checked by a first half of the ECC bits in parallel with a second group of a second half of the data bits checked by a second half of the ECC bits.

Abstract: A method includes receiving a data retrieval request. A plurality of identifiers are determined in accordance with the data retrieval request. Integrity information is generated based on determining the plurality of identifiers by performing a cyclic redundancy check. Stored integrity information corresponding to the data retrieval request is compared with the integrity information, where the stored integrity information was previously generated by performing the cyclic redundancy check. When the stored integrity information compares unfavorably with the integrity information, corruption associated with the plurality of identifiers is determined.

Abstract: User data units are received at a memory controller to be written to a RAID strip in non-volatile memory. A first parity value is calculated for the user data units using a first parity calculation. A second parity value different from the first parity value is also calculated for the plurality of user data units using a second parity calculation. The first parity value is stored in a first parity data unit in the non-volatile memory and the second parity value is stored in a second parity data unit in the non-volatile memory. Recovery from a failure of up to two data units thus enabled by recalculating the value of the failed data units based on one or more of the first parity data unit, the second parity data unit, and the values of other user data units of the plurality of data units.

Abstract: An apparatus is described. The apparatus includes a rank of memory chips to couple to a memory channel. The memory channel is characterized as having eight transfers of eight bits of raw data per burst access. The rank of memory chips has first, second and third X4 memory chips. The X4 memory chips conform to a JEDEC dual data rate (DDR) memory interface specification. The first and second X4 memory chips are to couple to an eight bit raw data portion of the memory channel's data bus. The third X4 memory chip to couple to an error correction coding (ECC) information portion of the memory channel's data bus.

Abstract: Various embodiments include an on-die error correction code (ECC) system that preserves rectangular symbols of arbitrary size and shape, where the dimensions of the symbol are powers of two. Further, the on-die ECC system preserves symbols that include multiple rectangles of arbitrary size and shape, where the dimensions of each rectangle are powers of two, and where the vertical and horizontal offset between consecutive rectangles are also powers of two. If the on-die ECC system miscorrects a memory bit, then the miscorrection is constrained or restricted to the same symbol that includes the other error bits. Therefore, all error bits, including the miscorrected bit, are in the same symbol. As a result, a user ECC system, such as a symbol-based ECC system, can correct and detect any number of errors within a single symbol, even when the on-die ECC system miscorrects a memory bit.

Abstract: Apparatuses and methods for data storage error protection are described. One example apparatus for data storage error protection includes an array of memory cells arranged in a first dimension and a second dimension. A controller is configured to determine a set of symbols corresponding to data stored in the memory cells. The controller is configured to add subsets of the set of symbols obliquely oriented to the first dimension and the second dimension to determine a number of parity check symbols. The controller is configured to use a same number of parity check symbols for protection of a first subset of memory cells oriented parallel to the first dimension as used for protection of a second subset of memory cells oriented parallel to the second dimension.

Abstract: Disclosed are devices, systems and methods improving the convergence of a soft bit-flipping decoder in a non-volatile memory device. An example method includes receiving a noisy codeword, the codeword having been generated based on a parity check matrix of an LDPC code and provided to a communication channel prior to reception by the soft bit-flipping decoder, generating, based on the noisy codeword, one or more messages for passing between a plurality of variable nodes and a plurality of check nodes of the soft bit-flipping decoder, generating a reliability metric for each of the one or more messages, storing the reliability metric only for messages comprising magnitudes that are less than or equal to a predetermined threshold value; and performing, based on the one or more messages and the associated reliability metric for at least one of the one more messages, a single decoding iteration of the soft bit-flipping decoder.

Abstract: The memory system includes a memory device including a volatile storage area and a non-volatile storage area; and a controller including first and second interfaces for transferring data between the memory system and a host, and suitable for transferring data between the volatile storage area and the host through the first interface and transferring data between the non-volatile storage area and the host through the second interface, wherein the controller is further suitable for determining whether or not an error occurs in data read from the volatile storage area in a normal operation mode, and dumping a whole of the volatile storage area into a predetermined first location of the non-volatile storage area when an error is determined to occur in the data read from the volatile storage area.

Abstract: A method for performing error correction for a plurality of storage drives and a storage appliance comprising a plurality of storage devices is disclosed. In one embodiment, the method includes generating a first set of parity bits from a first set of data of at least one of the plurality of storage devices, the first set of parity bits capable of correcting a first number of error bits of the first set of data. The method further includes generating a second set of parity bits from a concatenated set of the first data and a second set of data from at least another of the plurality of storage devices, the second set of parity bits capable of correcting a second number of error bits of the first set of data, the second number being greater than the first number.

Abstract: A technique relates to operating a memory controller. The memory controller drives first memory devices and second memory devices of the memory controller in a dual channel mode. A first error correcting code (ECC) memory device and a second ECC memory device protect the first memory devices and the second memory devices. The memory controller drives the first memory devices and the second memory devices in a single channel mode such that the second ECC memory device is a spare memory device, and the first ECC memory device protects the first memory devices and the second memory devices. The memory controller is configured to switch between the dual channel mode and the single channel mode.

Abstract: A system for a backplane serializer/deserializer (SerDes) including first and second integrated circuits (IC). The first and second ICs include transmitters and receivers coupled to each other through first and second bidirectional links. A first receiver is configured to receive first data at a data rate on a first channel supported by both the first bidirectional link and the second bidirectional link. A second receiver is configured to receive second data at the data rate on a second channel supported by both the first bidirectional link and the second bidirectional link. The backplane SerDes is configured to transfer the first and second data in full duplex mode by employing two-bit pulse-amplitude modulation (PAM-4) to reduce signaling speed of the first and second bidirectional links without reducing throughput of a lane pair including the first and second channels.

Abstract: According to one embodiment, a memory system includes a memory, and a processor. The memory converts an amount of charge held by a memory cell into a value. The processor executes a first process of reading first data from the memory. The processor executes a second process of reading the first data by making the memory use a first determination potential different in a case where error correction of the first data read through the first process is failed. The processor executes a third process of reading second data from the memory by making the memory use a third determination potential in a case where error correction of the first data read through the second process is succeeded. The third determination potential is the first determination potential used by the memory in a case where error correction of the first data read through the second process is succeeded.

Abstract: A data storage system stores a set of codewords in memory. The set of codewords are encoded in accordance with a joint nesting matrix specifying multiple layers of integrated interleaved codes, including first, second and third layers of integrated interleaved codes, and the set of codewords stored in the memory include first, second and third layers of parity information corresponding to the first, second and third layers of integrated interleaved codes. When decoding a first codeword and a first subgroup containing the first codeword fail, the system decodes a group of codewords that include two more subgroups of codewords, including the first subgroup of codewords, using the third layer parity information for the group of codewords. The second and third layers of integrated interleaved codes are configured to enable decoding of two codewords, in a subgroup of codewords, having errors beyond the correction capability of the first layer codes.

Abstract: A writing circuit for writing write data into a memory comprises an evaluator configured for providing an error handling code on the basis of the write data. A modifier reversibly modifies extended write data comprising both the write data and the error handling code in dependence on address information related to a writing address in order to provide modified extended write data. A writer writes the modified extended write data in a position of the memory defined by a writing address. A reading circuit for reading extended read data from a memory comprises a reader configured for reading the extended read data from a position of the memory defined by a reading address. A de-modifier modifies the extended read data in dependence on address information related to a reading address in order to provide extracted read data and an extracted error handling code. An error-detector detects based on the extracted error handling code whether the extracted read data comprises an error.

Abstract: A non-volatile storage system is provided that includes a mechanism to restore data that has been corrupted beyond the limits of traditional error correction. The system creates first level parity information for each subset of data to form multiple sets of programmable data, with each set of programmable data including a subset of data and corresponding first level parity. Separate second level parity is created for each set of programmable data. The system creates combined second level parity information based on a function of separate second level parity information for the multiple sets of programmable data. If a set of programmable data is found to be corrupt, the corrupt subset of data is recovered using the corrupt subset of data read from the non-volatile storage system, the corresponding first level parity read from the non-volatile storage system and the combined second level parity information.

Abstract: A method for performing error correction for a plurality of storage drives and a storage appliance comprising a plurality of storage devices is disclosed. In one embodiment, the method includes generating a first set of parity bits from a first set of data of at least one of the plurality of storage devices, the first set of parity bits capable of correcting a first number of error bits of the first set of data. The method further includes generating a second set of parity bits from a concatenated set of the first data and a second set of data from at least another of the plurality of storage devices, the second set of parity bits capable of correcting a second number of error bits of the first set of data, the second number being greater than the first number.

Abstract: A computer-implemented method includes receiving a first circuit design comprising a system of XOR gates, iteratively generating a plurality of candidate physical implementations of the system in adaptive logic modules included in logic array blocks of an integrated circuit, determining an overall metric for each of the plurality of candidate physical implementations using an objective function. The overall metric indicates at least an amount of the system that is implemented by each of the candidate physical implementations in a number of moves. The method also includes implementing an enhanced circuit design including a candidate physical implementation based at least in part on the overall metric of the candidate physical implementation on the integrated circuit.

Abstract: A system for a backplane serializer/deserializer (SerDes) including first and second integrated circuits (IC). The first and second ICs include transmitters and receivers coupled to each other through first and second bidirectional links. A first receiver is configured to receive first data at a data rate on a first channel supported by both the first bidirectional link and the second bidirectional link. A second receiver is configured to receive second data at the data rate on a second channel supported by both the first bidirectional link and the second bidirectional link. The backplane SerDes is configured to transfer the first and second data in full duplex mode by employing two-bit pulse-amplitude modulation (PAM-4) to reduce signaling speed of the first and second bidirectional links without reducing throughput of a lane pair including the first and second channels.

Abstract: According to one embodiment, a memory system includes a memory, and a processor. The memory converts an amount of charge held by a memory cell into a value. The processor executes a first process of reading first data from the memory. The processor executes a second process of reading the first data by making the memory use a first determination potential different in a case where error correction of the first data read through the first process is failed. The processor executes a third process of reading second data from the memory by making the memory use a third determination potential in a case where error correction of the first data read through the second process is succeeded. The third determination potential is the first determination potential used by the memory in a case where error correction of the first data read through the second process is succeeded.

Abstract: A data storage system includes a plurality of storage devices forming a storage array for storing data and associated error correction codes and a controller coupled to the plurality of storage devices. The controller is configured to, responsive to an error in a data element among the data, rebuild the data element from other data elements and an error correction code in the storage array and overwrite the error correction code with the rebuilt data element.

Abstract: A memory system according to the embodiment comprises a cell array of plural cells having three or more settable physical quantity levels and operative to store a code composed of symbols expressed by elements in a finite field Zp (p is a prime), wherein a set of two cells is defined as a pair cell and a combination of physical quantity levels of the two cells contained in the pair cell is defined as a pair cell level, wherein the pair cell uses a pair cell level of plural pair cell levels, which maximizes or minimizes a physical quantity level of one cell contained in the pair cell, to assign elements in the Zp to the pair cell levels, thereby storing symbols of the code.

Abstract: The disclosed embodiments relate to a Flash-based memory module having high-speed serial communication. The Flash-based memory module comprises, among other things, a plurality of I/O modules, each configured to communicate with an external device over one or more external communication links, a plurality of Flash-based memory cards, each comprising a plurality of Flash memory devices, and a plurality of crossbar switching elements, each being connected to a respective one of the Flash-based memory cards and configured to allow each one of the I/O modules to communicate with the respective one of the Flash-based memory cards. Each I/O module is connected to each crossbar switching element by a high-speed serial communication link, and each crossbar switching element is connected to the respective one of the Flash-based memory cards by a plurality of parallel communication links.

Abstract: Apparatus, systems, and methods for error correction in memory are described. In one embodiment, a memory controller comprises logic to receive a read request for data stored in a memory, retrieve the data and at least one associated error correction codeword, wherein the data and an associated error correction codeword is distributed across a plurality of memory devices in memory, apply a first error correction routine to decode the error correction codeword retrieved with the data and in response to an uncorrectable error in the error correction codeword, apply a second error correction routine to the plurality of devices in memory. Other embodiments are also disclosed and claimed.

Abstract: Provided are a method, system, and apparatus using reliability information from multiple storage units and a parity storage unit to recover data for a failed one of the storage units. A decoding operation of the codeword is performed in each of the storage units comprising the data storage units other than the target data storage unit and the parity storage unit to produce reliability information. In response to the decoding operation failing for at least one additional failed storage unit comprising the data and/or parity storage units other than the target data storage unit that failed to decode, reliability information is obtained for the data portion of the at least one additional failed storage unit. The reliability information obtained from the storage units other than the target data storage unit is used to produce corrected data for the data unit in the target data storage unit.

Abstract: A polynomial representation (bi(x)) in an AES finite field ( Z 2 ? [ x ] ( r ? ( x ) ) ) of input bytes (bi) of a state matrix (B) is obtained. A plurality (1) of irreducible polynomials (fi(y)) and a moving map (?i) are used to map each polynomial (bi(x)) of the polynomial representation into a respective field of polynomials ( Z 2 ? [ y ] ( f i ? ( y ) ) ) computed with respect to one of the irreducible polynomials (fi(y)), to obtain respective moved polynomials (?i(y)). The moved polynomials (?i(y)) are mapped into a polynomial (a(z)) of a polynomial ring ( Z 2 ? [ z ] ( p ? ( z ) ) ) , obtained by applying an isomorphism (?) between the fields of polynomials ( Z 2 ? [ y ] ( f i ? ( y ) ) ) and the polynomial ring ( Z 2 ? [ z ] ( p ? ( z ) ) ) based upon the Chinese remainder theorem (CRT). AES encryption is applied to the polynomial (a(z)).

Abstract: Described is an apparatus that comprises: a first sequential unit; a first queue coupled in parallel to the first sequential unit such that the first queue and first sequential unit receive a first input, the first sequential for double sampling the first input; a compare unit to receive an output from the first sequential unit; and a first selection unit controllable by a write pointer of a previous cycle, the first selection unit to receive outputs of each storage unit of the first queue, wherein the first selection unit to generate an output for comparison by the first compare unit.

Abstract: A memory controller provides error correcting code (ECC) capability for a memory. In some implementations, the controller is configured to identify an ECC protection level from a plurality of ECC protection levels for data that is to be stored in the memory device, generate ECC data for the data that is to be stored in the memory device using an ECC corresponding to the identified ECC protection level, store the generated ECC data in the cache, and store the data in the memory device.

Abstract: Embodiments of the present invention relate to an apparatus, method, and/or sequence for a distributed ECC that may be used in a storage system. In another embodiment of the invention, an apparatus for handling distributed error correction code (ECC) operations, includes: a plurality of ECC engines configured to perform ECC operations in parallel on multiple data parts; the plurality of ECC engines distributed in parallel to receive some of the multiple data parts that are read from storage media devices and to receive some of the other multiple data parts that are to be written to the storage media devices; and the plurality of ECC engines configured to use respective ECC bytes corresponding to respective ones of the multiple data parts.

Abstract: A basic property of flash memory is that: a 0-bit can be changed into a 1-bit, but not vice-versa, which severely limits the possibilities of reusing storage space with new data. A family of new coding methods is presented that enables double use of the memory, effectively expanding the combined amount of stored data. This can then be used as a compression booster, adding an additional layer to, and improving the compression of some rewriting methods that are not context sensitive.

Abstract: Receive a request to write a unit of data, having a first half of bits and a second half of bits, to an index of a ternary content addressable memory (TCAM). Generate a first error-correcting code (ECC) codeword for first bits of the first half of bits of the unit of data and first bits of the second half of bits of the unit of data. Generate a second error-correcting code (ECC) codeword for second bits of the first half of bits of the unit of data and second bits of the second half of bits of the unit of data. Store the first half of bits of the unit of data in the first row of the index. Store the second half of bits of the unit of data in the second row of the index.

Abstract: An address detection circuit comprises first to N-th address storage units suitable for storing an address, first to N-th calculation units each suitable for performing a counting operation when an address is stored in a corresponding address storage unit among the address storage units or the address stored in the corresponding address storage unit is inputted, a control unit suitable for sequentially storing an input address in the address storage units, and storing the input address in a selected address storage unit among the address storage units when of the address storage units each store an address, and a detection unit suitable for detecting an address, which is inputted a reference number of times or more, among the addresses stored in the address storage units, based on outputs of the calculation units.

Abstract: Methods and apparatus are provided for reading and writing data in q-level cells of solid-state memory, where q>2. Input data is encoded into codewords having N qary symbols, wherein the symbols of each codeword satisfy a single-parity-check condition. Each symbol is written in a respective cell of the solid state memory by setting the cell to a level dependent on the qary value of the symbol. Memory cells are read to obtain read signals corresponding to respective codewords. The codewords corresponding to respective read signals are detected by relating the read signals to a predetermined set of N-symbol vectors of one of which each possible codeword is a permutation.

Abstract: Examples of the present invention include an electronic-memory-system component. The electronic-memory-system component includes an array of data-storage elements and an encoder that receives input data, processes the input data as a two-dimensional array of bits by carrying out two passes, in one pass subjecting a portion of each row of the two-dimensional array of bits having more than a threshold weight to a first weight-reduction operation, and, in another pass, subjecting a portion of each considered column of the two-dimensional array of bits having more than a threshold weight to a second weight-reduction operation, one of the first and second weight-reduction operations employing an antipodal mapping and the other of the first and second weight-reduction operations employing bit inversion, generates a codeword corresponding to the input data, and stores the codeword in the array of data-storage elements.

Abstract: Techniques and mechanisms for handling data faults in a memory system which includes multiple integrated circuit (IC) dies, each die including a respective one of multiple memory arrays. In an embodiment, control logic monitors for a die failure of the multiple dies, and further monitors for a request to perform error correction for the multiple memory arrays. Each of the multiple memory arrays may store a respective vertical error correction code specific to data of that memory array. Another IC die may store a Bose, Ray-Chaudhuri, Hocquenghem (BCH) code of a horizontal codeword which spans the multiple memory arrays. In another embodiment, the BCH code is available to decode logic for data recovery operations in response to a die failure, where the BCH code is further available to the decode logic for error correction operations when all of the memory arrays are operative.

Abstract: A data processing device can perform error detection and correction in two stages: in the first stage, error detection is performed for the load data using the in-line error detection information. If a first type of error is detected in the data segment, the error is corrected using the in-line error detection information. If a second type of error is detected error correction is performed using the residual sum.

Abstract: A memory module (22, 122, 322, 522) including memory devices (24, 324) comprises a memory module buffer (26, 326, 526) having a spare state input (36) and a buffer memory (28). The memory module buffer (26, 326, 526) stores data in the buffer memory (28), the data being re-created from a portion of at least one of the memory devices (24, 324) determined to include an error.

Abstract: Embodiments of the present disclosure provide methods and apparatuses related to NVM devices with extended error correction protection. In some embodiments, a parity cache is used to store parity values of data values stored in a plurality of codewords of an NVM device. Other embodiments may be described and claimed.

Abstract: Provided is a semiconductor device configured to encode input data into a codeword including M different symbols, each of which includes Nm symbols. The semiconductor device including a first storage unit configured to store a first state value which is reset according to M and Nm; a second storage unit corresponding to any one of the M different symbols and configured to store M second state values determined through the corresponding symbol and the first state value; a third storage unit configured to store a third state value.

Abstract: A memory device is configured to correct errors in codewords written to a memory array. Errors, if any, in a first codeword are corrected and a codeword corrector output is generated including a corrected first codeword. A data buffer receives the codeword corrector output and a first user data associated with the addressed page and generates a data buffer output including the corrected first codeword, as modified by the first user data, defined as a first codeword output. A codeword encoder receives the data buffer output and encodes the first codeword output to generate an encoded first codeword output included in a codeword encoder output. A write buffer receives the codeword encoder output and saves the same for writing to the memory array. Writing to the memory array is performed while receiving a second user data, which has a second codeword associated therewith, and correcting the second codeword.

Abstract: An unequal bit-reliability information storage method for communication and storage systems at least includes one storage unit having a first memory and a second memory; the most significant information bits are stored in the first memory; and least significant information bits are stored in the second memory.

Abstract: Various embodiments include apparatus and methods to store data in a first semiconductor memory unit and to store error correction information in a second semiconductor memory unit to recover the data. The error correction information has a value equal to at least the value of the data store in the first memory unit.

Abstract: Methods and apparatus for encoding codewords which are particularly well suited for use with low density parity check (LDPC) codes and long codewords are described. The described methods allow encoding graph structures which are largely comprised of multiple identical copies of a much smaller graph. Copies of the smaller graph are subject to a controlled permutation operation to create the larger graph structure. The same controlled permutations are directly implemented to support bit passing between the replicated copies of the small graph. Bits corresponding to individual copies of the graph are stored in a memory and accessed in sets, one from each copy of the graph, using a SIMD read or write instruction. The graph permutation operation may be implemented by simply reordering bits, e.g., using a cyclic permutation operation, in each set of bits read out of a bit memory so that the bits are passed to processing circuits corresponding to different copies of the small graph.

Abstract: Embodiments of the present disclosure describe methods, apparatus, and system configurations for providing rank-specific cyclic redundancy checks in memory systems.

Abstract: A multimedia information receiving apparatus receives multimedia information which is transmitted by a broadcast system and receives multimedia information which is simultaneously transmitted by another transmission system such as IP communications, and generates one received information by selecting elements having a few errors from elements of demodulated broadcast system information and elements of demodulated other transmission system information and then arranging the selected elements.

Abstract: An information processing apparatus comprising: a reception unit adapted to receive a packet containing first data to be stored in a storage unit, a first address indicating an address of second data held in the storage unit, and a second address indicating an address at which the first data is to be written in the storage unit; an access unit adapted to read out the second data from the storage unit based on the first address, and write the first data in the storage unit based on the second address; and a transmission unit adapted to replace the first data of the packet received by the reception unit with the second data read out by the access unit, and transmit the packet.

Abstract: Memory devices and methods are described that include serially chained memory devices. In one or more of the configurations shown, a serial chain of memory devices includes a number of memory devices, and an error recovery device at an end of the chain. In one configuration shown, the serial chain of memory devices includes a chain of devices where each device is a stacked die memory device. Methods are described that show using the error recovery device in write operations and data recovery operations.