Abstract: A method of correcting a memory error of a dynamic random-access memory module (DRAM) using a double data rate (DDR) interface, the method includes conducting a memory transaction including multiple bursts with a memory controller to send data from data chips of the DRAM to the memory controller, detecting one or more errors using an ECC chip of the DRAM, determining a number of the bursts having the errors using the ECC chip of the DRAM, determining whether the number of the bursts having the errors is greater than a threshold number, determining a type of the errors, and directing the memory controller based on the determined type of the errors, wherein the DRAM includes a single ECC chip per memory channel.

Abstract: A method of memory deduplication includes identifying hash tables each corresponding to a hash function, and each including physical buckets, each physical bucket including ways and being configured to store data, identifying virtual buckets each including some physical buckets, and each sharing a physical bucket with another virtual bucket, identifying each of the physical buckets having data stored thereon as being assigned to a single virtual bucket, hashing a data line according to a hash function to produce a hash value, determining whether a corresponding virtual bucket has available space for a block of data according to the hash value, sequentially moving data from the corresponding virtual bucket to an adjacent virtual bucket when the corresponding virtual bucket does not have available space until the corresponding virtual bucket has space for the block of data, and storing the block of data in the corresponding virtual bucket.

Abstract: A system and method for mitigating overhead for accessing metadata for a cache in a hybrid memory module are disclosed. The method includes: providing a hybrid memory module including a DRAM cache, a flash memory, and an SRAM for storing a metadata cache; obtaining a host address including a DRAM cache tag and a DRAM cache index; and obtaining a metadata address from the DRAM cache index, wherein the metadata address includes a metadata cache tag and a metadata cache index. The method further includes determining a metadata cache hit based on a presence of a matching metadata cache entry in the metadata cache stored in the SRAM; in a case of a metadata cache hit, obtaining a cached copy of data included in the DRAM cache and skipping access to metadata included in the DRAM cache; and returning the data obtained from the DRAM cache to a host computer. The SRAM may further store a Bloom filter, and a potential DRAM cache hit may be determined based on a result of a Bloom filter test.

Abstract: A method of correcting a memory error of a dynamic random-access memory module (DRAM) using a double data rate (DDR) interface, the method includes conducting a memory transaction including multiple bursts with a memory controller to send data from data chips of the DRAM to the memory controller, detecting one or more errors using an ECC chip of the DRAM, determining a number of the bursts having the errors using the ECC chip of the DRAM, determining whether the number of the bursts having the errors is greater than a threshold number, determining a type of the errors, and directing the memory controller based on the determined type of the errors, wherein the DRAM includes a single ECC chip per memory channel.

Abstract: A method of processing in-memory commands in a high-bandwidth memory (HBM) system includes sending a function-in-HBM instruction to the HBM by a HBM memory controller of a GPU. A logic component of the HBM receives the FIM instruction and coordinates the instructions execution using the controller, an ALU, and a SRAM located on the logic component.

Abstract: A dynamic random access memory (DRAM) processing unit (DPU) may include at least one computing cell array having a plurality of DRAM-based computing cells arranged in an array having at least one column in which the at least one column may include at least three rows of DRAM-based computing cells configured to provide a logic function that operates on a first and a second row of the at least three rows and configured to store a result of the logic function in a third row of the at least three rows; and a controller that may be coupled to the at least one computing cell array to configure the at least one computing cell array to perform a DPU operation.

Abstract: A hybrid memory includes a plurality of tiles including a plurality of rows including a first row having a first type of memory cells and a second row having a second type of memory cells; a pair of bitline select signals including a bitline select signal and a bitline select bar signal that is an inverse of the bitline select signal; a wordline driver that is configured to receive an input data; a sense amplifier that is configured to output an output data; a write bitline coupled to the first row and the second row; a read bitline coupled to the first row and the second row; a wordline coupled to each of the plurality of rows; and a bitline that is coupled to the write bitline and the read bitline based on set values of the pair of bitline select signals.

Abstract: A memory device includes an array of 2T1C DRAM cells and a memory controller. The DRAM cells are arranged as a plurality of rows and columns of DRAM cells. The memory controller is internal to the memory device and is coupled to the array of DRAM cells. The memory controller is capable of receiving commands input to the memory device and is responsive to the received commands to control row-major access and column-major access to the array of DRAM cells. In one embodiment, each transistor of a memory cell includes a terminal directly coupled to a storage node of the capacitor. In another embodiment, a first transistor of a memory cell includes a terminal directly coupled to a storage node of the capacitor, and a second transistor of the 2T1C memory cell includes a gate terminal directly coupled to the storage node of the capacitor.

Abstract: A method of coordinating memory commands in a high-bandwidth memory HBM+ system, the method including sending a host memory controller command from a host memory controller to a memory, receiving the host memory controller command at a coordinating memory controller, forwarding the host memory controller command from the coordinating memory controller to the memory, and scheduling, by the coordinating memory controller, a coordinating memory controller command based on the host memory controller command.

Abstract: A system includes a library, a compiler, a driver and at least one dynamic random access memory (DRAM) processing unit (DPU). The library may determine at least one DPU operation corresponding to a received command. The compiler may form at least one DPU instruction for the DPU operation. The driver may send the at least one DPU instruction to at least one DPU. The DPU may include at least one computing cell array that includes a plurality of DRAM-based computing cells arranged in an array having at least one column in which the at least one column may include at least three rows of DRAM-based computing cells configured to provide a logic function that operates on a first row and a second row of the at least three rows and configured to store a result of the logic function in a third row of the at least three rows.

Abstract: A memory module has a logic including a programming register, a deduplication ratio control logic, and a deduplication engine. The programming register stores a maximum deduplication ratio of the memory module. The control logic is configured to control a deduplication ratio of the memory module according to the maximum deduplication ratio. The deduplication ratio is programmable by the host computer.

Abstract: A memory module includes a memory controller including: a host layer; a media layer coupled to a non-volatile memory; and a logic core coupled to the host layer, the media layer, and a volatile memory, the logic core storing a first write group table including a plurality of rows, and the logic core being configured to: receive a persistent write command including a cache line address and a write group identifier; receive data associated with the persistent write command; write the data to the volatile memory at the cache line address; store the cache line address in a selected buffer of a plurality of buffers in a second write group table, the selected buffer corresponding to the write group identifier; and update a row of the first write group table to identify locations of the selected buffer containing valid entries, the row corresponding to the write group identifier.

Abstract: A hardware coding mechanism is described. The coding mechanism may include a first encoder to produce a first code using a base number of bits and a second encoder to produce a second code using a supplementary number of bits. The second code and the first code together may be stronger than the first code alone. A mode register stored in a storage may specify whether a switch to the second encoder is open or closed: the first coder is always used.

Abstract: A hybrid memory controller performs receiving first and second central processing unit (CPU) requests to write to/read from a hybrid memory group, identifying a volatile memory device and a non-volatile memory device as a first target and second target of the first and second CPU requests, respectively, by decoding and address mapping of the first and second CPU requests, queuing the first and second CPU requests in first and second buffers, respectively, generating, based on an arbitration policy, a first command corresponding to one of the first and second CPU requests to an associated one of the first and second targets, and generating a second command corresponding to another one of the first and second CPU requests to an associated another one of the first and second targets, and transmitting the first and second commands to respective ones of the volatile and non-volatile memory devices.

Abstract: A method of choosing a cache line of a plurality of cache lines of data for eviction from a frontend memory, the method including assigning a baseline replacement score to each way of a plurality of ways of a cache, the ways respectively storing the cache lines, assigning a validity score to each way based on a degree of validity of the cache line stored in each way, assigning an eviction decision score to each way based on a function of the baseline replacement score for the way and the validity score for the way, and choosing a cache line of the way having a highest eviction decision score as the cache line for eviction.

Abstract: A hybrid memory includes a plurality of tiles including a plurality of rows including a first row having a first type of memory cells and a second row having a second type of memory cells; a pair of bitline select signals including a bitline select signal and a bitline select bar signal that is an inverse of the bitline select signal; a wordline driver that is configured to receive an input data; a sense amplifier that is configured to output an output data; a write bitline coupled to the first row and the second row; a read bitline coupled to the first row and the second row; a wordline coupled to each of the plurality of rows; and a bitline that is coupled to the write bitline and the read bitline based on set values of the pair of bitline select signals.

Abstract: According to one embodiment, the method includes: providing a hybrid memory module including a DRAM cache, a flash memory, and an SRAM for storing a metadata cache; obtaining a host address by decoding a data access request received from a host computer, wherein the host address includes a DRAM cache tag and a DRAM cache index; obtaining a metadata address from the DRAM cache index, wherein the metadata address includes a metadata cache tag and a metadata cache index; determining a metadata cache hit based on a presence of a matching metadata cache entry in the metadata cache of the SRAM; in a case of the metadata cache hit, obtaining the data from the DRAM cache and skipping an access to the metadata of the DRAM cache; and returning the data obtained from the DRAM cache to the host computer.

Abstract: A method of wear leveling for a storage device or a memory device includes: receiving an inputted memory address; randomizing the inputted memory address to be a randomized memory address; and periodically reassigning the randomized memory address to be a different memory address.

Abstract: A method for replacing data on a volatile memory cache is provided. The volatile memory cache includes one or more memory banks and each of the memory banks includes a plurality of memory lines. The method includes: identifying a replacement ID for at least one of the memory lines to be replaced; identifying a refresh bank ID for one of the memory banks to be refreshed; determining whether or not a conflict exists between the replacement ID and the refresh bank ID; and selecting a new replacement ID if the conflict exists.

Abstract: A method of memory deduplication includes identifying hash tables each corresponding to a hash function, and each including physical buckets, each physical bucket including ways and being configured to store data, identifying virtual buckets each including some physical buckets, and each sharing a physical bucket with another virtual bucket, identifying each of the physical buckets having data stored thereon as being assigned to a single virtual bucket, hashing a data line according to a hash function to produce a hash value, determining whether a corresponding virtual bucket has available space for a block of data according to the hash value, sequentially moving data from the corresponding virtual bucket to an adjacent virtual bucket when the corresponding virtual bucket does not have available space until the corresponding virtual bucket has space for the block of data, and storing the block of data in the corresponding virtual bucket.