Abstract: Memory cell presetting for improved performance including a method for using a computer system to identify a region in a memory. The region includes a plurality of memory cells characterized by a write performance characteristic that has a first expected value when a write operation changes a current state of the memory cells to a desired state of the memory cells and a second expected value when the write operation changes a specified state of the memory cells to the desired state of the memory cells. The second expected value is closer than the first expected value to a desired value of the write performance characteristic. The plurality of memory cells in the region are set to the specified state, and the data is written into the plurality of memory cells responsive to the setting.

Abstract: Encoding data into constrained memory using a method for writing data that includes receiving write data to be encoded into a write word, receiving constraints on symbol values associated with the write word, encoding the write data into the write word, and writing the write word to a memory. The encoding includes: representing the write data and the constraints as a first linear system in a first field of a first size; embedding the first linear system into a second linear system in a second field of a second size, the second size larger than the first size; solving the second linear system in the second field resulting in a solution; and collapsing the solution into the first field resulting in the write word, the write word satisfying the constraints on symbol values associated with the write word.

Abstract: A system to improve memory failure management may include memory, and an error control decoder to determine failures in the memory. The system may also include an agent that may monitor failures in the memory. The system may further include a table where the error control decoder may record the failures, and where the agent can read and write to.

Abstract: Adaptive endurance coding including a method for storing data that includes receiving write data and a write address. A compression algorithm is applied to the write data to generate compressed data. An endurance code is applied to the compressed data to generate a codeword. The endurance code is selected and applied in response to the amount of space saved by applying the compression to the write data. The codeword is written to the write address.

Abstract: Providing increased capacity in heterogeneous storage elements including a method for storing data including a write process writing to a memory and a read process reading from the memory. Physical characteristics of memory cells in the memory support different sets of data levels. The write process takes into account the different sets of data levels when writing to the memory. The read process first obtains data in the memory and subsequently determines how to interpret the data.

Abstract: Memory cell presetting for improved performance including a system that includes a memory, a cache, and a memory controller. The memory includes memory lines made up of memory cells. The cache includes cache lines that correspond to a subset of the memory lines. The memory controller is in communication with the memory and the cache. The memory controller is configured to perform a method that includes scheduling a request to set memory cells of a memory line to a common specified state in response to a cache line attaining a dirty state.

Abstract: Adaptive endurance coding including a method for accessing memory that includes retrieving a codeword from a memory address. The codeword is multiplied by a metadata matrix to recover metadata for the codeword. The metadata includes a data location specification. The data in the codeword is identified in response to the metadata and the data is output as read data.

Abstract: Providing increased capacity in heterogeneous storage elements including a method for storing data in a heterogeneous memory that includes receiving a write message and a write address corresponding to a block of memory cells where at least two of the memory cells support different data levels, determining physical characteristics of the memory cells, and identifying virtual memories associated with the block of memory cells in response to the physical characteristics. The following is performed for each of the virtual memories: generating a constraint vector that describes the virtual cells in the virtual memory; and calculating a virtual write vector in response to the constraint vector and the write data, the calculating including writing the write data, bit by bit, in order, into the virtual memory, skipping locations known to be stuck to a particular value as indicated by the constraint vector.

Abstract: Providing increased capacity in heterogeneous storage elements including a method for reading from memory. The method includes receiving a read word from a block of memory cells, where physical characteristics of the memory cells support different sets of data levels. The read word is separated into two or more virtual read vectors. For each of the virtual read vectors, the codebook that was utilized to generate the virtual read vector is identified and a partial read data vector is generated. The generating includes multiplying the virtual read vector by a matrix that represents the codebook. The partial read data vectors are combined into a read message and the read message is output.

Abstract: A computer-implemented method for verifying a RAIM/ECC design using a hierarchical injection scheme that includes selecting marks for generating an error mask, selecting a fixed bit mask based on the selected marks, determining whether to inject errors into at least one of a marked channel and at least one marked chip of a channel; and randomly injecting errors into the at least one of the marked channel and the at least one marked chip when determined.

Abstract: A data protection method is provided. The method includes receiving data; generating compressed data based on the data; determining a degree of compressibility based on the compressed data; determining an amount of free space based on the degree of compressibility; and setting one or more error bits based on the amount of free space.

Abstract: A memory system that includes a plurality of memory arrays having memory cells characterized by a variable write time. The memory system also includes a memory bus configured to receive write commands, and a plurality of data buffers configured to communicate with the memory arrays. The memory system further includes an address buffer configured to communicate with the memory arrays to store the write addresses. A mechanism configured to receive a write command and to split a data line received with the write command into a number of parts is also included in the memory system. The parts of the data line are stored in different data buffers and the writing of the parts of the data line to memory arrays at the write address is initiated. The write command is completed when write completion signals specifying the write address have been received from all of the memory arrays.

Abstract: Performing wear-leveling and bad block management of limited lifetime memory devices. A method for performing wear-leveling in a memory includes receiving logical memory addresses and applying a randomizing function to the logical memory addresses to generate intermediate addresses within a range of intermediate addresses. The intermediate addresses are mapped into physical addresses of a memory using an algebraic mapping. The physical addresses are within a range of physical addresses that include at least one more location than the range of intermediate addresses. The physical addresses are output for use in accessing the memory. The mapping between the intermediate addresses and the physical addresses is periodically shifted. In addition, contents of bad blocks are replaced with redundantly encoded redirection addresses.

Abstract: Performing wear-leveling and bad block management of limited lifetime memory devices. A method for performing wear-leveling in a memory includes receiving logical memory addresses and applying a randomizing function to the logical memory addresses to generate intermediate addresses within a range of intermediate addresses. The intermediate addresses are mapped into physical addresses of a memory using an algebraic mapping. The physical addresses are within a range of physical addresses that include at least one more location than the range of intermediate addresses. The physical addresses are output for use in accessing the memory. The mapping between the intermediate addresses and the physical addresses is periodically shifted. In addition, contents of bad blocks are replaced with redundantly encoded redirection addresses.

Abstract: Correcting memory device (chip) and memory channel failures in the presence of known memory device failures. A memory channel failure is located and corrected, or alternatively up to c chip failures are corrected and up to d chip failures are detected in the presence of up to u chips that are marked as suspect. A first stage of decoding is performed that results in recovering an estimate of correctable errors affecting the data or in declaring an uncorrectable error state. When an uncorrectable error state is declared, a second stage of decoding is performed to attempt to correct u erasures and a channel error in M iterations where the channel location is changed in each iteration. A correctable error is declared in response to exactly one of the M iterations being successful.

Abstract: A phase-change memory (PCM) includes a matrix of storage cells, including at least a first group with at least one cell. Each cell includes a phase change material having at least a first resistance value and a second resistance value, such that the first group can have an identical message encoded therein in at least a first way and a second way. The memory also includes a controller configured to encode the identical message in the at least first group the first or second way, based on which way causes the least amount of writing cost, given current levels of the group. Another embodiment of memory includes a matrix of storage cells. Each of the storage cells has at least two levels, such that each of the storage cells can have an identical message encoded therein in at least a first way and a second way.

Abstract: Techniques for writing to memory using adaptive write techniques. An adaptive write technique includes receiving at a computer a message including a plurality of symbols. The message is written to a memory. The writing to memory includes performing for each symbol in the message: writing a data value to a memory location in the memory and reading contents of the memory location after the data value has been written. The data value is determined at the computer in response to the symbol and to the contents of any memory locations previously read as part of writing the message to the memory. It is determined at the computer if the contents of the memory locations reflect the message. The writing is restarted at the computer in response to determining that the contents of the memory locations do not reflect the message.

Abstract: Systems and methods are provided that schedule task requests within a computing system based upon the history of task requests. The history of task requests can be represented by a historical log that monitors the receipt of high priority task request submissions over time. This historical log in combination with other user defined scheduling rules is used to schedule the task requests. Task requests in the computer system are maintained in a list that can be divided into a hierarchy of queues differentiated by the level of priority associated with the task requests contained within that queue. The user-defined scheduling rules give scheduling priority to the higher priority task requests, and the historical log is used to predict subsequent submissions of high priority task requests so that lower priority task requests that would interfere with the higher priority task requests will be delayed or will not be scheduled for processing.

Abstract: Marking memory chips as faulty when a fault is detected in data from the memory chip. Upon detecting that a plurality of memory chips are faulty, determining which of a plurality of memory channels contains the faulty memory chips. Marking one of a plurality of memory channels as failing in response to determining that the number of failing memory chips has exceeded a threshold.

Abstract: A system and method for providing a cyclical redundancy code (CRC) for use in a high-speed serial link. The system includes a cascade interconnect memory system including a memory controller, a memory hub device and a downstream link. The downstream link is in communication with the memory controller and the memory hub device and includes at least thirteen signal lanes for transmitting a multiple transfer downstream frame from the memory controller to the memory hub device. A portion of the downstream frame includes downstream CRC bits to detect errors in the downstream frame. The downstream CRC bits capable of detecting any one of a lane failure, a transfer failure and up to five bit random errors.