Abstract: According to various embodiments, an in-memory computation system is disclosed. The system includes a dynamic random access memory (DRAM) module. The system further includes a memory controller configured to violate a timing specification for the DRAM module and activate multiple rows of the DRAM module in rapid succession to enable bit-line charge sharing.

Abstract: A refresh processing method, apparatus, and system, and memory controllers are provided, to improve memory access efficiency. The refresh processing apparatus includes a plurality of memory controllers that are in one-to-one correspondence with a plurality of memory spaces. Any first memory controller in the plurality of memory controllers is configured to: receive N first indication signals and N second indication signals that are output by N memory controllers other than the first memory controller, where N is greater than or equal to 1; and determine a refresh policy of a first memory space based on at least one of the following information: the N first indication signals, the N second indication signals, and refresh indication information of the first memory space.

Abstract: A system for refresh operations in semiconductor memories, and an apparatus and method therefore, are described. The system includes, for example, memory cells in memory banks that are refreshed during a self-refresh operation or an auto refresh operation. The self-refresh operation includes a different number of refresh activations than the auto refresh operation. The system further includes a row control circuit configured to refresh the memory cells in the memory banks based on refresh control signals received from a refresh control circuit, the refresh control signals provided by the refresh control circuit based on internal control signals received by the refresh control circuit from a command control circuit. The auto refresh operation includes a per bank refresh operation configured to refresh a corresponding memory bank or an all-bank refresh operation configured to refresh all memory banks.

Abstract: Methods of operating a memory device are disclosed. A method may include determining an operating temperature of a memory bank of a memory device. The method may also include adjusting at least one refresh interval for the memory bank based on the operating temperature of the memory bank. Further, the method may include skipping at least one refresh of the memory bank based on at least one of the operation temperature of the memory bank and a number of active signals received at the memory bank. A memory device and an electronic system are also described.

Abstract: A semiconductor memory device includes a memory cell array, an error correction code (ECC) engine, a refresh control circuit, a scrubbing control circuit and a control logic circuit. The refresh control circuit generates refresh row addresses for refreshing a memory region on memory cell rows in response to a first command received from a memory controller. The scrubbing control circuit counts the refresh row addresses and generates a scrubbing address for performing a scrubbing operation on a first memory cell row of the memory cell rows whenever the scrubbing control circuit counts N refresh row addresses of the refresh row addresses. The ECC engine reads first data corresponding to a first codeword, from at least one sub-page in the first memory cell row, corrects at least one error bit in the first codeword and writes back the corrected first codeword in a corresponding memory location.

Abstract: A nonvolatile memory device is provided. A nonvolatile memory device comprises a word line, a bit line, a memory cell array including a first memory cell at an intersection region between the word line and the bit line, a word line voltage generating circuitry configured to generate a program voltage, the program voltage to be provided to the word line, a row decoder circuitry configured to receive the program voltage from the word line voltage generating circuitry and configured to provide the program voltage to the word line, a verification circuitry configured to generate a verification signal in response to verifying a success or a failure of programming of the first memory cell, and a control circuitry configured to apply the program voltage to the first memory cell in response to the verification signal, and configured to cut off the program voltage in response to the verification signal.

Abstract: The memory controller controls a memory device. The controller is configured to determine to perform a target operation on a first memory block and determine an activation voltage level transferred to a block word line based on block state information of a second memory block.

Abstract: A memory device may include a cell array and a cycle calculating circuit. The cycle calculating circuit may calculate an operating cycle of a refresh operation to be performed at the cell array, based on an operating temperature of the memory device. In response to the operating temperature being lower than a first temperature, the cycle calculating circuit may be configured to calculate the operating cycle by integrating one or more slope values of a second slope value to an nth slope value that are arranged from a highest temperature to a lowest temperature. The second slope value may correspond to a second temperature, the nth slope value may correspond to an nth temperature, n may be a natural number of 2 or more, and a number of the one or more slope values may be based on the operating temperature.

Abstract: The present disclosure includes apparatuses and methods to determine timing of operations. An example method includes performing a first operation type that uses a shared resource in a memory device. The method includes applying a scheduling policy for timing of continued performance of the first operation type based upon receipt of a request to the memory device for performance of a second operation type that uses the shared resource.

Abstract: Apparatuses and methods for pure-time, self-adopt sampling for RHR refresh. An example apparatus includes a memory bank comprising a plurality of rows each associated with a respective row address, and a sampling timing generator circuit configured to provide a timing signal having a plurality of pulses. Each of the plurality of pulses is configured to initiate sampling of a respective row address associated with a row of the plurality of rows to detect a row hammer attack. The sampling timing generator includes first circuitry configured to provide a first subset of pulses of the plurality of pulses during a first time period and includes second circuitry configured to initiate provision of a second subset of pulses of the plurality of pulses during a second time period after the first time period.

Abstract: An information handling system organic light emitting diode (OLED) display adapts presentation of visual information to manage a display thermal state, such as to maintain a uniform thermal condition. In one example embodiment, a first display zone of plural pixels presents visual images with a first refresh rate, such as the nominal refresh rate established by the display settings, and a second display zone presents visual images with a partial pixel refresh rate that reduces power dissipation and, thus, thermal energy release so that a more uniform thermal state may be achieved at the display.

Abstract: A system for refresh operations including multiple refresh activations, and a method and an apparatus therefore, are described. The system includes, for example, a memory array; a command address input circuit configured to provide a command for a per bank refresh operation or an all-bank refresh operation, a command control circuit configured to receive the command, and provide first and second internal control signals; a refresh control circuit configured to provide a first refresh control signal; and a row control circuit configured to provide a second refresh control signal. The provided first internal control signal is based on the provided command. For the per bank refresh operation, the provided second internal control signal is based on the second refresh control signal, and, for the all-bank refresh operation, the provided second internal control signal is based on the first internal control signal delayed by the command control circuit.

Abstract: Provided herein are memory devices, systems including memory devices, and methods of operating memory devices in which multiple counters are provided to permit memory refresh commands greater freedom in targeting subsets of the memory device for data refresh operations. In one embodiment, a memory device is provided, comprising a plurality of memory banks, and circuitry configured to (i) store a plurality of values, each of the plurality of values corresponding to one of the plurality of memory banks; (ii) refresh first data stored in a first one of the plurality of memory banks; and (iii) update a first one of the plurality of values corresponding to the first one of the plurality of memory banks based at least in part on refreshing the first data.

Abstract: A method for controlling the refresh of data in reprogrammable nonvolatile memories includes a plurality of memory pages for storing data.

Abstract: A phase change memory cell includes a first electrode, a second electrode located over the first electrode, a vertical pillar structure located between the first and second electrodes, the pillar structure containing a first phase change memory (PCM) material portion, a second PCM material portion and an intermediate electrode located between the first PCM material portion and the second PCM material portion, and a resistive liner containing a first segment electrically connected in parallel to the first PCM material portion between the first electrode and the intermediate electrode, and a second segment electrically connected in parallel to the second PCM material portion between the intermediate electrode and the second electrode. The first PCM material portion has a different electrical resistance than the second PCM material portion, and the first segment of the resistive liner has a different electrical resistance than the second segment of the resistive liner.

Abstract: An integrated circuit may include a first standard cell including first and second active regions extending in a first horizontal direction and a first gate line extending in a second horizontal direction orthogonal to the first horizontal direction; and a second standard cell including third and fourth active regions extending in the first horizontal direction and a second gate line aligned in parallel to the first gate in the second horizontal direction and being adjacent to the first standard cell. A distance between the second active region of the first standard cell and the third active region of the second standard cell may be greater than a distance between the first and second active regions of the first standard cell, and may be greater than a distance between the third and fourth active regions of the second standard cell.

Abstract: The present disclosure provides devices and methods for refreshing memory. The method includes receiving information of a memory array in an accelerator coupled to a host device, and when the accelerator operates in a first refresh mode, based on the received information: disabling one or more refresh operations of a memory controller; and refreshing one or more rows of the memory array by the host device.

Abstract: Apparatuses, systems, and methods are disclosed for concurrently programming non-volatile storage cells, such as those of an SLC NAND array. The non-volatile storage cells may be arranged into a first block comprising a first string of storage cells that intersects with a first word line at a first storage cell, a second block comprising a second string of storage cells that intersects with a second word line at a second storage cell, a bit line electrically connectable to the first string and the second string, and controller configured to apply a programming pulse, at an elevated voltage, to the first word line and second word line to concurrently program the first and second storage cells.

Abstract: A memory device includes a plurality of memory chips for writing and reading data in response to a control command and an address signal, and a control logic circuit for transferring the control command and the address signal to the plurality of the memory chips, and receiving a first command from a memory controller to perform a first operation, different from a refresh operation, on at least one of a plurality of the memory chips. The control logic circuit, in response to a refresh command, transmits the first command to at least one of a plurality of the memory chips and performs the first operation during a pre-determined refresh time interval without carrying out the refresh operation.

Abstract: Embodiments of the disclosure are drawn to apparatuses and methods for staggering the timing of targeted refresh operations. A memory device may include a number of memory banks, at least some of which may be simultaneously entered into a refresh mode. A given memory bank may perform an auto-refresh operation or a targeted refresh operation, which may draw less power than the auto-refresh operation. The timing of the targeted refresh operations may be staggered between the refreshing memory banks, such that a portion of the refreshing memory banks are performing a targeted refresh operation simultaneously with a portion of the refreshing memory banks performing an auto-refresh operation.

Abstract: A refresh testing circuit and method are provided. The refresh testing circuit includes an internal clock generator, a counter, and an address detection circuit. The internal clock generator transmits a control clock signal to a refresh controller to generate a bank selection signal and a row address signal for a refresh operation. The counter counts variations of the bank selection signal to generate a count value. The address detection circuit detects whether a value of the row address signal is sequentially increased during the refresh operations to generate a detection signal.

Abstract: Various additional and alternative aspects are described herein. In some aspects, the present disclosure provides a method of controlling a memory of a computing device by an adaptive memory controller. The method includes collecting usage data from the computing device over a first bin, wherein the first bin is associated with a first weight, wherein the first weight is indicative of one or more of a first partial array self-refresh (PASR) setting a first partial array auto refresh (PAAR) setting and a first deep power down (DPD) setting. The method further includes associating the collected data with a second weight, adapting the first bin based on the second weight, wherein the second weight is indicative of one or more of a second PASR, PAAR, and DPD setting. The method further includes controlling the memory during the next first bin based on the second weight.

Abstract: Systems including a processor and a memory device in communication with the processor include an array of non-volatile memory cells configured in a NAND architecture. The array includes a plurality of series-coupled first non-volatile memory cells, each first non-volatile memory cell curving around a first curved side of a substantially vertical pillar and terminating at an isolation region, and a plurality of series-coupled second non-volatile memory cells, each second non-volatile memory cell curving around a second curved side of the substantially vertical pillar and terminating at the isolation region. Respective ones of the first non-volatile memory cells are respectively at same vertical levels as respective ones of the second non-volatile memory cells.

Abstract: Embodiments of the disclosure are drawn to apparatuses and methods for distributed timing of targeted refresh operations. Information stored in volatile memory cells may decay unless refresh operations are performed. A memory device may perform auto-refresh operations, as well as one or more types of targeted refresh operations, where particular rows are targeted for a refresh. Targeted refresh operations may draw less power than an auto-refresh operation. It may be desirable to distribute targeted refresh operations throughout a sequence of refresh operations, to average out a power draw in the memory device. Responsive to an activation of a refresh signal, the memory device may perform a group of refresh operations. At least one refresh operation in each group may be a targeted refresh operation.

Abstract: A memory device having memory cells operates in a normal mode, a first self refresh mode, and a second self refresh mode. The first self refresh mode provides a self refresh operation for retaining data stored in the memory cells without an external command. The time required to return to the normal mode from the first self refresh mode is shorter than a reference time. The second self refresh mode also provides the self refresh operation, but a time required to return to the normal mode from the second self refresh mode is longer than the reference time. The normal mode provides a higher operating voltage to the memory cells than does the second self refresh mode.

Abstract: A semiconductor device according to an aspect of the present invention has: a plurality of memory cells MC; a plurality of word lines WL each coupled to a corresponding one of the plurality of memory cells MC; and a control circuit that intermittently monitors accesses to the plurality of word lines WL, stores/erases some captured row-addresses in a first number of registers, and detects, by comparison with stored addresses, in response to a first number of accesses to one of the word lines WL in a first period of time. According to the present invention, access histories can be precisely analyzed by a small-scale circuit configuration, and measures against, for example, the Row Hammer problem, etc. can be taken.

Abstract: Embodiments of the disclosure are drawn to apparatuses and methods for analog row access tracking. A plurality of unit cells are provided, each of which contains one or more analog circuits used to track accesses to a portion of the wordlines of a memory device. When a wordline in the portion is accessed, the unit cell may update an accumulator voltage, for example by adding charge to a capacitor. A comparator circuit may determine when one or more accumulator voltages cross a threshold (e.g., a reference voltage). Responsive to the accumulator voltage crossing the threshold, an aggressor address may be loaded in a targeted refresh queue, or if the aggressor address is already in the queue, a priority flag associated with that address may be set. Aggressor addresses may be provided to have their victims refreshed in an order based on the number of set priority flags.

Abstract: This application discloses a memory device to retain stored data when receiving a voltage supply having at least a retention voltage level. The retention voltage level varies based on a supply voltage and a temperature of an environment around the memory device. A sensitive circuit can adjust the voltage supply received by the memory device based on the supply voltage and the temperature. The sensitive circuit can alter a memory bias supply voltage for the memory device to adjust the voltage supply towards the retention voltage level. The sensitive circuit can include a temperature dependent circuit to generate a bias voltage based on the supply voltage and the temperature, and an adjustment circuit to alter the memory bias supply voltage based on the bias voltage. The adjustment circuit also can include high temperature circuitry to adjust the memory bias supply voltage based on a leakage current from the memory device.

Abstract: The present disclosure describes a power stage. The power stage includes a metal-oxide semiconductor field-effect transistor (MOSFET) and a driver IC coupled to the MOSFET. The driver IC is configured to switch the MOSFET to an ON-state so that MOSFET conducts a current. The driver IC includes a current monitor circuit that outputs a current monitor signal, which corresponds to the current through the MOSFET when it is in the ON-state. The current monitor signal includes an error caused by a temperature difference between the MOSFET and the driver IC. As a result, the driver IC further includes a compensation circuit that is configured to determine a thermal gradient across the driver IC, and based on the thermal gradient, adjust the current monitor circuit to reduce the error.

Abstract: According to one embodiment, a memory device includes a memory cell including a resistance change memory element and a selector element, a word line, a bit line connected to one end of the memory cell, an operational amplifier including a non-inverting input connected to the bit line, an output circuit including a first terminal connected to an output of the operational amplifier, a second terminal connected to the bit line, and a charge/discharge circuit including a capacitor, a charge circuit and a discharge circuit, each including one end connected to an inverting input of the operational amplifier. At the time of falling of a write voltage for the memory cell, a potential of the other end of the memory cell is set higher than a potential of the other end of the discharge circuit.

Abstract: Disclosed herein is an apparatus that includes a memory cell array including a plurality of memory cells, a first counter circuit configured to periodically update a count value during a first operation mode, a burst clock generator configured to successively generate a burst pulse predetermined times when the count value indicates a predetermined value, and a row address control circuit configured to perform a refresh operation on the memory cell array in response to the burst pulse.

Abstract: The present disclosure provides a control system and a control method for a double data rate synchronous dynamic random access memory (DDR) system in order to reduce power consumption of the DDR system. The system includes a memory and a processor coupled to the memory, and the processor is configured to monitor a working status of each functional system that shares a power domain with the DDR system, determine a target power parameter value and a target clock parameter value of the DDR system according to the working status of each functional system, and control a power parameter and a clock parameter of the DDR system according to the target power parameter value and the target clock parameter value of the DDR system.

Abstract: Apparatuses and methods are disclosed that include ferroelectric memory and for refreshing ferroelectric memory. An example apparatus includes: a word line; a first memory cell coupled to a first digit line and stores a first data on the first digit line responsive to the word line in an active state; a second memory cell coupled to a second digit line and stores a second data on the second digit line responsive to the word line in the active state. The first digit line is coupled to a first power potential and the second digit line is coupled to a second power potential in a refresh operation.

Abstract: A method for performing a refresh operation based on system characteristics is provided. The method includes determining that a current operating condition of a memory component is in a first state. The method also includes detecting a change in the operating condition from the first state to a second state. The method further includes setting a refresh period associated with the memory component based on the change of the operating condition. The refresh period corresponds to a period of time between a first time when a write operation is performed on a segment of the memory component and a second time when a refresh operation is to be performed on the segment. Moreover, the method includes performing the refresh operation according to the refresh period.

Abstract: A memory circuit including a plurality of elementary cells arranged in a plurality of arrays, each including a plurality of rows and a plurality of columns, and wherein: the elementary cells having the same coordinates in the different arrays share a same first conductive track; and in each array, the elementary cells of a same row of the array share a same second conductive track and a same third conductive track.

Abstract: Methods, systems, and devices for refresh-related activation in memory are described. A memory device may conduct a refresh operation to preserve the integrity of data. A refresh operation may be associated with a refresh time where the memory device is unable to execute or issue any commands (e.g., access commands). By posting (e.g., saving) one or more commands and/or row addresses during the refresh time, the memory device may be configured to execute the saved commands and/or re-open one or more rows associated with the saved row addresses at a later time (e.g., upon completion of the refresh operation). Accordingly, fewer commands may be issued to activate the memory cells after the refresh time.

Abstract: Techniques are described for performing a read scan process on a non-volatile memory system in order to determine memory blocks that may have a high bit error rate, where if such blocks are found they can be refreshed. Rather than work through the blocks of a memory system sequentially based on the physical block addresses, the memory system maintains a measure of data quality, such as an estimated or average bit error rate, for multi-block groups. For example, the groups can correspond to regions of memory die in the system. The groups are ranked by their data quality, with the groups being scanned in order of the data quality. The blocks within a group can also be ranked, based on factors such as the program/erase count.

Abstract: Embodiments of an interconnect apparatus enable improved signal integrity, even at high clock rates, increased bandwidth, and lower latency. In an interconnect apparatus for core arrays a sending processing core can send data to a receiving core by forming a packet whose header indicates the location of the receiving core and whose pay load is the data to be sent. The packet is sent to a Data Vortex switch described herein and in the patents incorporated herein. The Data Vortex switch is on the same chip as an array of processing cores and routes the packet to the receiving core first by routing the packet to the processing core array containing the receiving processing core. The Data Vortex switch then routes the packet to the receiving processor core in a processor core array. Since the Data Vortex switches are not crossbar switches, there is no need to globally set and reset the Data Vortex switches as different groups of packets enter the switches.

Abstract: A method of protecting a DRAM memory device from the row hammer effect includes the memory device comprising a plurality of banks composed of memory rows, the method being implemented by at least one logic prevention device configured to respectively associate contiguous sections of rows of a bank with sub-banks and to execute, on each activation of a row of a sub-bank (b) of the memory, an increment step of a required number of preventive refreshments (REFRESH_ACC; REFRESH_ACC/PARAM_D) of the sub-bank (b) using an activation threshold (PARAM_D) of the sub-bank (b). The prevention logic is also configured to execute a preventive refresh sequence of the sub-banks according to their required number of preventive refreshes. A DRAM memory device, a buffer circuit or a controller of such a memory may comprise the logic for preventing the row hammer effect.

Abstract: Disclosed are methods and systems for producing bipole source modeling with reduced computational loads. A method may comprise receiving first electromagnetic data and second electromagnetic data from a first shotpoint and a second of a marine electromagnetic survey, modelling a first electromagnetic field and second electromagnetic field for one or more dipole sources of a bipole source and combining a plurality of data points to provide an approximation of an electromagnetic field for the bipole source. A system may comprise electromagnetic sensors, a bipole source, wherein the bipole source comprise a pair of electrodes that are separated by a distance, wherein the bipole source is configured to generate an electromagnetic field, and a data processing system configured to receive electromagnetic data from a plurality of shotpoints of the bipole source and model electromagnetic fields for one or more dipole sources of the bipole source from the electromagnetic data.

Abstract: Systems and methods for reducing the size of sub-blocks within a physical memory block for a three-dimensional non-volatile memory using buried source lines are described. The physical memory block may be fabricated using dual buried source lines such that sub-blocks within the physical memory block may be individually selected in both a horizontal word line direction and a vertical NAND string direction. The physical memory block may include a plurality of sub-blocks that are individually selectable and that share bit lines and/or source-side select gate lines. The plurality of sub-blocks that are individually selectable may correspond with different portions of the same NAND string in which a first sub-block of the plurality of sub-blocks connects to a drain-side select gate for the NAND string and a second sub-block of the plurality of sub-blocks connects to a source-side select gate for the NAND string.

Abstract: Techniques are disclosed for notifying network control software of new and moved source MAC addresses. In one embodiment, a switch may redirect a packet sent by a new or migrated virtual machine to the network control software as a notification. The switch does not forward the packet, thereby protecting against denial of service attacks. The switch further adds to a forwarding database a temporary entry which includes a “No_Redirect” flag for a new source MAC address, or updates an existing entry for a source MAC address that hits in the forwarding database by setting the “No_Redirect” flag. The “No_Redirect” flag indicates whether a notification has already been sent to the network control software for this source MAC address. The switch may periodically retry the notification to the network control software, until the network control software validates the source MAC address, depending on whether the “No_Redirect” is set.

Abstract: A memory device applies different refresh rates to target data (or objective data) according to data characteristics (i.e., required reliability levels). The memory device includes a memory cell array provided with a plurality of memory cells, a row decoder configured to selectively activate word lines of the memory cell array in response to a row address signal, and a refresh controller configured to output the row address signal in response to the row address signal. The refresh controller controls a refresh ratio of a first storage region and a second storage region contained in the memory cell array in response to a changeable refresh control value.

Abstract: The subject technology performs the following in a storage system including memory dies, where each memory die includes its own temperature sensor. The subject technology determines a temperature of each of the memory dies based on a temperature reading from each memory die's temperature sensor. The subject technology determines whether data is hot data or cold data, where hot data is more likely to be changed after it is written than cold data. In response to determining that the data is hot data, the subject technology stores the data in a memory die with a relatively higher temperature than another one of the memory dies. Further, in response to determining that the data is cold data, the subject technology stores the data in a memory die with a relatively cooler temperature than another one of the memory dies.

Abstract: A method includes forming first and second gate stacks over a substrate. Each of the first and second gate stacks includes a tunneling dielectric layer, a floating gate over the tunneling dielectric layer, a middle dielectric layer over the floating gate, and a control gate over the middle dielectric layer. A conductive layer is formed over the first and second gate stacks. The conductive layer is etched to form a erase gate between the first and second gate stacks. Etching the conductive layer is performed such that a top surface of the erase gate is not higher than a top surface of the control gate and such that the top surface of the erase gate is at least partially curved inwards.

Abstract: A memory includes: first to Nth areas refreshed based on first to Nth refresh control signals, respectively; a control signal generation circuit suitable for generating the second to Nth refresh control signals by sequentially delaying the first refresh control signal, and generating the first refresh control signal by delaying the Nth refresh control signal; an address counter suitable for changing a refresh address, corresponding to each round for activations of the first to Nth refresh control signals, based on the Nth refresh control signal; and a refresh stop circuit suitable for stopping a refresh operation when the round is repeated by a predetermined number.

Abstract: Apparatuses and methods for targeted row refreshes are disclosed herein. In an example apparatus, a predecoder receives a target row address and determines whether a target row of memory associated with the target row address is a primary or a redundant row of memory. The predecoder is further configured to cause one or more rows of memory physically adjacent the primary row of memory to be refreshed if the primary row is the target row or one or more rows of memory physically adjacent the redundant row of memory to be refreshed if the redundant row of memory is the target row of memory.

Abstract: A determination is made by a processing device included in a memory component that an operation to program data to a location in the memory component has failed, the data is programmed to a different location in the memory component by the processing device upon determining the operation has failed, and a notification that the data has been programmed to the different location in the memory component is provided by the processing device to a processing device operatively coupled to the memory component.

Abstract: A memory system includes a first memory chip that includes a first temperature sensor, and a memory controller that includes a second temperature sensor. The memory controller is configured to: perform, at a first timing, a first temperature acquisition process including acquiring a first measured temperature using the first temperature sensor or the second temperature sensor; select one of the first temperature sensor and the second temperature sensor for a second temperature acquisition process based the first measured temperature; and perform, at a second timing later than the first timing, the second temperature acquisition process including acquiring a second measured temperature using the selected one of the first temperature sensor or the second temperature sensor.

Abstract: An apparatus comprises first and second memory regions each to store data using a data storage technology for which retention of data for longer than a predetermined period of time is dependent on a refresh operation for refreshing data in the memory region being performed at a frequency that is greater than or equal to a minimum refresh frequency. The apparatus further comprises at least one controller to control storage of data in the first memory region with the refresh operation performed at a first frequency lower than said minimum refresh frequency when valid data is stored in the first memory region, and to control storage of data in the second memory region with the refresh operation performed at a second frequency that is greater than or equal to said minimum refresh frequency. The at least one controller is configured to communicate with the first memory region via a first memory channel and with the second memory region via a second memory channel.