Storage device and electronic device including heterogeneous volatile memory chip

By using heterogeneous volatile memory chips in storage devices, combining high bandwidth and low power consumption DRAM chips, it solves the problem that traditional storage devices are difficult to meet different application needs at the same time, and achieves an efficient and flexible storage solution.

CN110047522BActive Publication Date: 2024-11-19SAMSUNG ELECTRONICS CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN201811607749.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2017-12-28
Filing Date
2018-12-27
Publication Date
2024-11-19
Estimated Expiration
2038-12-27

AI Technical Summary

Technical Problem

Existing storage devices are difficult to meet the needs of high bandwidth and low power consumption at the same time. Especially in different application scenarios, traditional homogeneous memory chips cannot effectively adjust bandwidth and power consumption, resulting in inefficiency of devices in different applications.

Method used

Using heterogeneous volatile memory chips, by combining high-bandwidth and low-power DRAM chips in storage devices, leveraging different bandwidth and power consumption characteristics, providing storage solutions for different applications, such as stacking different types of DRAM chips and buffers form multi-channel memory packages to achieve efficient data transmission.

Benefits of technology

It realizes storage devices that support both high bandwidth and low power consumption in one package, and is suitable for a variety of application scenarios, improving the flexibility and efficiency of storage devices and reducing manufacturing costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN110047522B_ABST
    Figure CN110047522B_ABST
Patent Text Reader

Abstract

A storage device includes: a first volatile memory chip including a first volatile memory cell array storing first data and receiving or outputting the first data at a first bandwidth; and a second volatile memory chip including a second volatile memory cell array storing second data and receiving or outputting the second data at a second bandwidth different from the first bandwidth.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] Priority is claimed on Korean Patent Application No. 10-2017-0182041 filed on December 28, 2017 in the Korean Intellectual Property Office, the entire contents of which are incorporated herein by reference. Technical Field

[0003] The inventive concepts described herein relate to a memory device and an electronic device including the same, and more particularly, to a memory device including a heterogeneous volatile memory chip and an electronic device including the same. Background Art

[0004] The application processor (AP) may be implemented in the form of a system on chip (hereinafter referred to as "SoC"). SoC may refer to one chip in which various systems are integrated. SoC may include a processor for executing a program according to an application. SoC may include a storage device that communicates with the processor and stores a program to be executed by the processor, a result of the processing, and the like.

[0005] The storage device may be, for example, a dynamic random access memory (DRAM) and may operate as the main memory of the SoC. Characteristics of the storage device, such as capacity, bandwidth, and power consumption, may be associated with the applications supported by the SoC. Since the SoC supports a variety of applications, there may be a need for storage devices suitable for executing the various applications. Summary of the invention

[0006] Embodiments of the inventive concept provide a memory device including a heterogeneous volatile memory chip and an electronic device including the same.

[0007] An embodiment of the inventive concept provides a storage device including: a first volatile memory chip including a first volatile memory cell array storing first data and receiving or outputting the first data at a first bandwidth; and a second volatile memory chip including a second volatile memory cell array storing second data and receiving or outputting the second data at a second bandwidth different from the first bandwidth.

[0008] An embodiment of the present inventive concept provides a storage device including: a first volatile memory die including a first volatile memory cell array storing first data, and a first through silicon via for input / output of the first data; a second volatile memory die including a second volatile memory cell array storing second data, and a second through silicon via for input / output of the second data; and a buffer die receiving the first data through the first through silicon via and outputting the first data at a first bandwidth, and receiving the second data through the second through silicon via and outputting the second data at a second bandwidth different from the first bandwidth.

[0009] An embodiment of the inventive concept provides an electronic device, comprising: a system on chip, including a first processor and a second processor; and a storage device, including a first volatile memory chip communicating with the first processor through a first channel, and a second volatile memory chip communicating with the second processor through a second channel. A first bandwidth of the first channel and a second bandwidth of the second channel are different from each other. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The above and other objects and features of the present inventive concept will become apparent in consideration of the following detailed description of exemplary embodiments with reference to the accompanying drawings.

[0011] Figure 1 FIG. 1 is a diagram illustrating an electronic device according to an embodiment of the inventive concept.

[0012] Figure 2 The diagram shows Figure 1 A block diagram of the first memory chip.

[0013] Figure 3 The diagram shows Figure 1 Block diagram of the second memory chip.

[0014] Figure 4 FIG. 2 illustrates a block diagram of an electronic device according to an embodiment of the inventive concept.

[0015] Figure 5 The figure shows Figure 4 A timing diagram of the operation of a storage device processing a read command of an SoC.

[0016] Figure 6 The diagram shows Figure 4 A flowchart of a method for operating an electronic device.

[0017] Figure 7 FIG. 4 is a block diagram of an electronic device according to another embodiment of the inventive concept.

[0018] Figure 8 FIG. 4 is a block diagram of an electronic device according to another embodiment of the inventive concept.

[0019] Fig. 9 FIG. 4 is a block diagram of an electronic device according to another embodiment of the inventive concept. DETAILED DESCRIPTION

[0020] In the following, embodiments of the inventive concept are described in detail and clearly to the extent that a person having ordinary skill in the art can implement the inventive concept.

[0021] Figure 1 FIG. 1 is a diagram illustrating an electronic device according to an embodiment of the inventive concept. Figure 1 , the electronic device 10 includes a system on chip (hereinafter referred to as “SoC”) 11 , a storage device 12 , and a substrate 13 .

[0022] SoC 11 may control the overall operation of electronic device 10 as, for example, an application processor (AP). SoC 11 may execute a program according to an application supported by electronic device 10, and may receive data associated with program execution from storage device 12 or may transmit a result of program execution to storage device 12.

[0023] The storage device 12 includes a first memory chip 100, a second memory chip 200, and a substrate 300. The first memory chip 100 may include a memory cell array storing data, and may receive and output data at a first bandwidth. The second memory chip 200 may include a memory cell array storing data, and may receive and output data at a second bandwidth. Here, the first bandwidth and the second bandwidth may be different from each other. The bandwidth (i.e., the storage bandwidth) may be a rate at which data can be read from a memory chip or stored in a memory chip by the SoC 11. In some embodiments, the storage device 12 may include more than two memory chips.

[0024] refer to Figure 1 , the second memory chip 200 is stacked on the substrate 300, and the first memory chip 100 is stacked on the second memory chip. However, in other embodiments, Figure 1 Unlike the embodiment of the present invention, the first memory chip 100 may be stacked on the substrate 300, and the second memory chip 200 may be stacked on the first memory chip 100. Here, the order in which the first memory chip 100 and the second memory chip 200 are stacked on the substrate 300 is not limited to Figure 1 Furthermore, one of the first memory chip 100 and the second memory chip 200 may be stacked on a first surface or an upper surface of the substrate 300, and the other of the first memory chip 100 and the second memory chip 200 may be stacked on a second surface or a lower surface opposite to the first surface or the upper surface. Furthermore, the number of memory chips stacked on the substrate 300 is not limited to Figure 1 The quantity shown in the figure.

[0025] refer to Figure 1, the first memory chip 100 and the second memory chip 200 can be electrically connected to the substrate 300 through the wire 400. The wire and the first memory chip 100 and the second memory chip 200 can be protected by molding resin. Pads for wire bonding can be placed on one surface of the substrate 300, and solder balls or solder bumps can be placed on the opposite surface of the substrate 300. For example, the substrate 300 can be a printed circuit board (PCB), a flexible printed circuit board (FPCB), a ceramic substrate, or an interposer. For example, the first memory chip 100 and the second memory chip 200 can be electrically connected to the substrate 300 by flip-chip bonding, solder bumps, conductive vias such as through silicon vias (TSVs), and / or combinations thereof.

[0026] In an embodiment of the inventive concept, each of the first memory chip 100 and the second memory chip 200 may be a volatile memory chip that is a dynamic random access memory (DRAM) chip requiring a refresh operation. For example, the DRAM chip can be a general-purpose DRAM chip, such as a double data rate synchronous dynamic random access memory (DDR SDRAM) chip, a DDR2 SDRAM chip, a DDR3 SDRAM chip, a DDR4 SDRAM chip, a DDR5 SDRAM chip, and the like; a DRAM chip for mobile applications, such as a low power double data rate (LPDDR) SDRAM chip, a LPDDR2 SDRAM chip, a LPDDR3 SDRAM chip, a LPDDR4SDRAM chip, a LPDDR4X SDRAM chip, a LPDDR5 SDRAM chip, and the like; or a DRAM chip providing high bandwidth, such as a graphics double data rate (GDDR) synchronous graphics random access memory (SGRAM) chip, a GDDR2 SGRAM chip, a GDDR3 SGRAM chip, a GDDR4 SGRAM chip, a GDDR5 SGRAM chip, a GDDR6SGRAM chip, a high bandwidth memory (HBM) chip, an HBM2 chip, an HBM3 chip, a WideIO SDRAM chip, and the like.

[0027] According to an embodiment of the inventive concept, the first memory chip 100, the second memory chip 200, and the substrate 300 may be included in one package. The types or kinds of the first memory chip 100 and the second memory chip 200 stacked in one package may be different from each other, and the bandwidth provided by the first memory chip 100 and the bandwidth provided by the second memory chip 200 may be different from each other.

[0028] For example, the first memory chip 100 may be a DRAM chip for mobile applications, and the second memory chip 200 may be a DRAM chip providing a higher bandwidth than the first memory chip 100. The second memory chip 200 may provide a higher bandwidth than the first memory chip 100, and the power consumption of the first memory chip 100 may be less than the power consumption of the second memory chip 200. That is, characteristics such as capacity, bandwidth, and power consumption of the first memory chip 100 and the second memory chip 200 may be different from each other.

[0029] The type or kind of the first memory chip 100 and the type or kind of the second memory chip 200 are not limited to those described above. For example, the first memory chip 100 may be a general-purpose DRAM chip, and the second memory chip 200 may be a DRAM chip that provides a higher (or larger) bandwidth than the first memory chip 100. Moreover, the first memory chip 100 and the second memory chip 200 may be DRAM chips for mobile applications. For example, the first memory chip 100 may be an LPDDR4 chip, and the second memory chip 200 may be an LPDDR5 chip, a GDDR6 chip, etc. that provide a higher bandwidth than the LPDDR4 chip.

[0030] The memory device 12 according to an embodiment of the inventive concept can be manufactured as a package that includes heterogeneous memory chips instead of homogeneous memory chips. As such, even if the memory device 12 is a package, the memory device 12 can be used for both applications requiring high bandwidth and applications requiring low power consumption.

[0031] In an embodiment of the present invention, a package (i.e., a storage device 12) may be, for example, any of a package-on-package (PoP), a ball grid array (BGA), a chip scale package (CSP), a plastic lead chip carrier (PLCC), a molded dual in-line package (PDIP), a die in a waffle package, a die in wafer form, a chip on board (COB), a ceramic dual in-line package (CERDIP), a plastic standard quad flat package (MQFP), a thin quad flat package (TQFP), a small outline integrated circuit (SSI), a shrink small outline package (SSOP), a thin small outline package (TSOP), a system-in-package (SIP), a multi-chip package (MCP), a wafer-level manufactured package (WFP), a wafer-level processed stacked package (WSP), etc. Moreover, the SoC 11 may be implemented by using various combinations of the packages described above.

[0032] A transmission path, i.e., a channel, between the SoC 11 and the memory device 12 may be formed in the substrate 13. For example, the channel may include a command path and an address path for controlling the memory device 12, and a data input / output path between the SoC 11 and the memory device 12. For example, the substrate 13 may be a printed circuit board, a flexible circuit board, a ceramic substrate, or an interposer.

[0033] Figure 2 The diagram shows Figure 1 The block diagram of the first memory chip will be referred to as Figure 1 To describe Figure 2 . refer to Figure 2 , the first memory chip 100 includes a clock pin 111, a command and address (CA) pin 112, a DQ pin 113, a clock buffer 121, a CA buffer 122, a DQ buffer 123, a command decoder 130, a mode register 140, a refresh controller 150, a memory body 161, a memory body controller 162, a data buffer 163, a serializer 171 and a deserializer 172.

[0034] The clock pin 111 may be used to receive a clock from outside the first memory chip 100 (eg, from a host, a memory controller, or Figure 1 The SoC 11) of the first memory chip 100 receives a clock signal CK. The clock signal CK may be a unidirectional signal input to the first memory chip 100, and the clock pin 111 may be an input terminal. For the convenience of description, Figure 2 Only one clock pin 111 is illustrated in the figure. However, in some embodiments, the first memory chip 100 may receive a differential clock signal, and the first memory chip 100 may further include another clock pin for receiving a clock signal (e.g., CKb) having an opposite phase to the clock signal CK (e.g., CKa) input through the clock pin 111.

[0035] The CA pin 112 may be a terminal for receiving a CA signal CA[1:I] from outside the first memory chip 100. The number of CA pins 112 and the number of CA signals CA[1:I] may be "1", that is, they may be the same as each other. Here, "1" may be a natural number and may be determined in advance according to various protocols. The CA signal CA[1:I] may include a command for controlling the first memory chip 100 and an address indicating the location of a memory cell in the memory body 161. The CA signal CA[1:I] may be a unidirectional signal input to the first memory chip 100, and the CA pin 112 may be an input terminal.

[0036] The DQ pin 113 may be a terminal for receiving a DQ signal DQ[1:J] from the outside of the first memory chip 100 or outputting a DQ signal DQ[1:J] to the outside of the first memory chip 100. The number of DQ pins 113 and the number of DQ signals DQ[1:J] may be "J", that is, they may be the same as each other. Here, "J" may be a natural number and may be determined in advance according to various protocols. The DQ signal DQ[1:J] may include write data for a write command or read data for a read command. The DQ signal DQ[1:J] may be a bidirectional signal input to or output from the first memory chip 100, and the DQ pin 113 may be an input / output terminal.

[0037] In an embodiment of the inventive concept, the transmission rate of the DQ signal DQ[1:J] of the DQ pin 113 may be twice the frequency of the clock signal CK. Here, the unit of the transmission rate is bps (bit per second), and the unit of the frequency is Hz.

[0038] The configuration of the first memory chip 100 can be determined according to "J" (which is the number of DQ pins 113) and the number of DQ signals DQ[1:J]. For example, if "J" is 16, the first memory chip 100 can support x16 mode. However, in addition to the x16 mode, the first memory chip 100 can further support x4 and x8 modes. The value of "J" described above is only an example, and "J" can be any value greater than 16, such as 32, 64, 128, 256, 512, 1024, 2048, etc.

[0039] The clock buffer 121 may operate as a receiver (RX) that receives the clock signal CK. The clock buffer 121 may receive the clock signal CK and may output an internal clock signal ICK to the inside of the first memory chip 100. Figure 2 The figure shows that the internal clock signal ICK is output only to the serializer 171 and the deserializer 172. However, although not shown, the internal clock signal ICK may also be output to various components of the first memory chip 100. The first memory chip 100 may process a command in the SDRAM based on the clock signal CK input through the clock pin 111.

[0040] The CA buffer 122 may operate as a receiver that receives the CA signal CA[1:I]. The number of CA buffers 122 may be the same as the number of CA pins 112. The CA buffer 122 may sample or latch the CA signal CA[1:I] at a rising edge or a falling edge of the internal clock signal ICK. The CA buffer 122 may transmit a signal corresponding to a command in the received CA signal CA[1:I] to the command decoder 130. The CA buffer 122 may transmit a signal corresponding to an operation code OPCODE in the received CA signal CA[1:I] to the mode register 140 and may transmit a signal corresponding to an address (e.g., a storage body address BA, a row address RA, and a column address CA) in the received CA signal CA[1:I] to the memory controller 162. The positions of the command, the operation code OPCODE, and the address included in the CA signal CA[1:I] may be determined in advance according to various protocols.

[0041] The DQ buffer 123 may include a receiver that receives the DQ signal DQ[1:J] and a transmitter that transmits the DQ signal DQ[1:J]. Each of the number of receivers and the number of transmitters may be the same as the number of DQ pins 113. The receiver of the DQ buffer 123 may transmit the received DQ signal DQ[1:J] to the deserializer 172. The transmitter of the DQ buffer 123 may receive the DQ signal DQ[1:J] from the serializer 171 and may output the received DQ signal DQ[1:J] to the outside through the DQ pin 113.

[0042] The command decoder 130 may receive the CA signal CA[1:1] from the CA buffer 122. The command decoder 130 may decode a signal corresponding to a command in the CA signal CA[1:1]. For example, the command decoder 130 may decode an activation command, a write command, a read command, a precharge command, and a mode register set (MRS) command among various commands and may control components of the first memory chip 100.

[0043] The command decoder 130 may transmit a bank control signal B_CTRL, a row control signal R_CTRL, a column control signal C_CTRL, and a refresh control signal REF in response to an activate command, a write command, a read command, a precharge command, a refresh command, etc. The command decoder 130 may also transmit a refresh control signal REF to the refresh controller 150. The command decoder 130 may transmit an MRS signal to the mode register 140 in response to an MRS command. In addition to the components described above, the command decoder 130 may control the operation of any other components of the first memory chip 100.

[0044] The mode register 140 may store setting values ​​for various operation modes supported by the first memory chip 100. The mode register 140 may include one or more mode registers. For example, Figure 1 The SoC 11 may set various operation modes of the first memory chip 100 by using the setting values ​​stored in the mode register 140. To this end, the SoC 11 may transmit an MRS command and an operation code OPCODE to the first memory chip 100.

[0045] For example, the settings of various operation modes may include the following operations: setting burst length, setting read burst type, setting column address strobe (CAS) wait time, setting delay locked loop (DLL) enable or reset, setting write recovery (WR) and read to precharge (RTP), setting additional delay (AL), setting RTT_NOM, RTT_WR, and RTT_PARK, setting impedance control of output driver, setting write leveling, setting TDQS enable, setting write cyclic redundancy check (CRC), setting refresh operation, setting CAS write wait time (CWL), setting multi-purpose register (MPR), setting temperature sensor, setting gear reduction mode, setting write and read preamble, setting reference voltage (Vref), setting power-down, setting data bus inversion (DBI), setting data masking, setting on-chip termination (ODT), setting parity check, setting ZQ calibration, and setting various other modes.

[0046] The refresh controller 150 may control the row decoder of the bank controller 162 based on the refresh control signal REF. The refresh operation is necessary to retain the data stored in the bank 161. The refresh operation may be performed by activating a word line of the bank 161. For example, all memory cells of the bank 161 may be activated at least once during a given time (e.g., 32 ms, 64 ms, etc.).

[0047] The refresh controller 150 may transmit a row address REF_RA for a refresh operation to a row decoder of the memory body controller 162. In this case, whenever a refresh command is input from the outside, the row address REF_RA for the refresh operation may be updated. Alternatively, in the case where the first memory chip 100 enters a self-refresh mode, the row address REF_RA for the refresh operation may be updated internally for each given period without an external refresh command. For the update operation described above, the refresh controller 150 may include a counter that generates and updates the row address REF_RA for the refresh operation.

[0048] In an embodiment, the refresh cycle may vary with the temperature of the first memory chip 100. Also, the refresh operation of the first memory chip 100 may be performed in memory bank units (i.e., refresh by memory bank). If the refresh operation of the first memory chip 100 is performed by memory bank unit, when any one memory bank is refreshed, the SoC 11 may access any other memory bank.

[0049] The memory bank 161 is a memory cell array including memory cells arranged repeatedly. The memory cells may be arranged at the intersection of a word line (not shown) and a bit line (not shown) of the memory bank. Here, the memory cells of the memory bank 161 may be volatile memory cells that need to be refreshed in order to retain the data stored therein.

[0050] The number of banks 161 may be determined in advance according to various protocols of the memory chip. When the number of banks 161 becomes larger, the memory cells of the first memory chip 100 may be divided into more banks, and the number of banks accessible by the SoC 11 may increase.

[0051] The memory bank controller 162 may control the memory bank 161 based on the memory bank control signal B_CTRL, the row control signal R_CTRL, the column control signal C_CTRL, and the refresh control signal REF. In an embodiment, the memory bank controller 162 may control each memory bank 161. In other embodiments, one memory bank controller may control two or more memory banks. In other words, one memory bank controller may be shared by two or more memory banks. Each of the memory bank controllers 162 may include a row decoder 162_1 and a column decoder 162_2.

[0052] The row decoder 162_1 may receive a row control signal R_CTRL from the command decoder 130 and a row address RA from the CA buffer 122, and may select a word line WL. Also, the row decoder 162_1 may receive a refresh control signal REF from the command decoder 130 and a row address REF_RA for a refresh operation from the refresh controller 150, and may select a word line WL. For example, in order to drive a word line of a memory bank, the row decoder 162_1 may be arranged along a direction in which the word lines are arranged at regular intervals.

[0053] The column decoder 162_2 may receive the column control signal C_CTRL from the command decoder 130 and the column address CA from the CA buffer 122, and may select the column selection line CSL. Since one or more bit lines (not shown) are connected to the column selection line CSL, the bit line connected to the column selection line CSL may be selected by the column decoder 162_2. For example, in order to drive the bit lines of the memory body, the column decoder 162_2 may be arranged along a direction in which the column selection lines are arranged at regular intervals. Figure 2 Unlike shown in FIG. 1 , the word lines and the column selection lines of the memory bank 161 may cross to be perpendicular to each other.

[0054] The data buffer 163 may transmit write data to the memory bank 161 through the global input / output line GIO or may receive read data from the memory bank 161 through the global input / output line GIO. The number of data buffers 163 may be the same as the number of memory banks 161, and the data buffer 163 may exchange data with the corresponding memory bank 161. Here, the number of bits included in the write / read data may be determined in consideration of factors such as the number of pre-fetch bits, the burst length, the number of DQ pins 113, and the like.

[0055] At least one of the data buffers 163 may be selected according to a read command or a write command by the command decoder 130. The number of data buffers selected in response to a read command or a write command may be determined according to the number of prefetch bits of the first memory chip 100, for example.

[0056] The data buffer selected according to the write command may receive the write data from the deserializer 172, and the selected data buffer may transmit the write data to the selected memory cell. The selected data buffer may drive the global input / output line and the selected memory cell based on the write data.

[0057] The data buffer selected according to the read command can receive and store the read data output from the memory cell selected according to the read command. To this end, the selected data buffer can read out and amplify the voltage of the global input / output line. The selected data buffer can transmit the read data to the serializer 171.

[0058] The serializer 171 may serialize the bits of the read data in response to a read command. The deserializer 172 may deserialize the bits of the write data in a write operation. For example, the serializer 171 may be implemented as a serializer including the same number as the number of DQ signals DQ[1:J], and the deserializer 172 may be implemented as a deserializer including the same number as the number of DQ signals DQ[1:J].

[0059] In an embodiment, the serializer 171 and the deserializer 172 may operate based on the internal clock signal ICK. The serializer 171 may serialize the read data output from the data buffer 163 and generate a DQ signal DQ[1:J] having a transmission rate (i.e., double data rate (DDR)) twice the frequency of the internal clock signal ICK. The deserializer 172 may deserialize the DQ signal DQ[1:J] having a transmission rate twice the frequency of the internal clock signal ICK. That is, the first memory chip 100 may support a DDR mode in which a read operation and a write operation are performed at a transmission rate corresponding to twice the frequency of the clock signal CK.

[0060] In the embodiment, although Figure 2 , but the first memory chip 100 may further receive the data strobe signal DQS together with the write data. In this case, the first memory chip 100 may further include a pin and a buffer for receiving the data strobe signal DQS. The deserializer 172 may operate based on both the internal clock signal ICK and the data strobe signal DQS. Moreover, the first memory chip 100 may generate a data strobe signal DQS aligned with the read data serialized by the serializer 171, and the data strobe signal DQS thus generated may be output to the outside of the first memory chip 100 together with the DQ signal DQ[1:J].

[0061] Figure 3 The diagram shows Figure 1 The block diagram of the second memory chip will be referred to Figure 1 and Figure 2 To describe Figure 3 . refer to Figure 3 , the second memory chip 200 includes a clock pin 211, a CA pin 212, a DQ pin 213, a clock buffer 221, a CA buffer 222, a DQ buffer 223, a command decoder 230, a mode register 240, a refresh controller 250, a memory bank 261, a memory bank controller 262, a data buffer 263, a serializer 271, and a deserializer 272. The components described above can be respectively connected to Figure 2 The clock pin 111, CA pin 112, DQ pin 113, clock buffer 121, CA buffer 122, DQ buffer 123, command decoder 130, mode register 140, refresh controller 150, memory bank 161, memory bank controller 162, data buffer 163, serializer 171 and deserializer 172 of the first memory chip 100 operate similarly and are similar. Hereinafter, a description will be given about the difference between the first memory chip 100 and the second memory chip 200, and a description of similar components may be omitted.

[0062] The bandwidth of the second memory chip 200 may be higher than that of the first memory chip 100. For example, the unit of the bandwidth may be gigabytes per second (GBps), and the bandwidth may increase as the transfer rate per DQ pin becomes higher.

[0063] To provide a higher bandwidth than that of the first memory chip 100 , the second memory chip 200 may additionally include a data clock pin 214 , a data clock buffer 224 , an error detection code (EDC) pin 215 , an EDC buffer 225 , and a phase detector 280 .

[0064] The data clock pin 214 may be a terminal for receiving a data clock signal WCK from the outside of the second storage device 200. The frequency of the data clock signal WCK may be higher than that of the clock signal CK.

[0065] The data clock signal WCK may be a unidirectional signal input to the second memory chip 200, and the data clock pin 214 may be an input terminal. As described with respect to the clock signal CK, the second memory chip 200 may receive a differential clock signal, and thus the second memory chip 200 may further include a data clock pin for receiving a clock signal (e.g., WCKb) having an opposite phase to the data clock signal WCK (e.g., WCKa) input through the data clock pin 214.

[0066] The data clock buffer 224 may operate as a receiver that receives the data clock signal WCK. The data clock buffer 224 may receive the data clock signal WCK and may output an internal data clock signal IWCK to the inside of the second memory chip 200. In detail, in order to make the transmission rate higher, the data clock buffer 224 may transmit the internal data clock signal IWCK to the serializer 271 and the deserializer 272.

[0067] and Figure 2 Unlike the serializer 171 and the deserializer 172 , the serializer 271 and the deserializer 272 may perform serialization and deserialization operations, respectively, based on an internal data clock signal IWCK instead of a clock signal CK.

[0068] The second memory chip 200 may generate a DQ signal DQ[1:N] having a faster transmission rate than the transmission rate of the DQ signal DQ[1:J] of the first memory chip 100 by using the data clock signal WCK at the DQ pin 213. For example, the transmission rate of each of the DQ pins 213 may be at least twice the frequency of the data clock signal WCK. The second memory chip 200 may support a DDR mode, a quad data rate (QDR) mode, etc., for example, with respect to the frequency of the data clock signal WCK.

[0069] In order to provide alignment between the data clock signal WCK and the clock signal CK, the data clock buffer 224 transmits the internal data clock signal IWCK to the phase detector 280. Before the second memory chip 200 processes a read command or a write command, training (WCK2CK training) between the data clock signal WCK and the clock signal CK may be performed in advance. Through the WCK2CK training, Figure 1 The SoC 11 may accurately determine the read latency (RL) and the write latency (WL) of the second memory chip 200 based on the clock signal CK.

[0070] The phase detector 280 may sample the internal data clock signal IWCK at a rising edge or a falling edge of the internal clock signal ICK. The number of times the internal data clock signal IWCK is sampled by the phase detector 280 may be one or more times. The phase detector 280 may repeatedly sample the internal data clock signal IWCK to determine whether the internal data clock signal IWCK is earlier or later than the internal data clock signal IWCK with respect to the internal clock signal ICK. The phase detector 280 may output an EDC signal indicating the result of the determination to the outside (e.g., SoC 11) of the second memory chip 200 through the EDC buffer 225 and the EDC pin 215. Here, the EDC buffer 225 may include a transmitter for outputting an EDC signal, and the EDC pin 215 may be an output terminal for outputting an EDC signal.

[0071] The EDC signal output to the outside can be used for WCK2CK training. For example, the SoC 11 can adjust the phase of the data clock signal WCK based on the EDC signal and can reset a phase-locked loop (PLL) (not shown) or a delay-locked loop (DLL) (not shown) of the second memory chip 200.

[0072] The second memory chip 200 can support a WCK2CK automatic synchronization mode between the data clock signal WCK and the clock signal CK. In the automatic synchronization mode, the delay of the PLL or DLL of the second memory chip 200 can be adjusted, and thus, the phases of the internal data clock signal IWCK and the internal clock signal ICK can be synchronized with each other.

[0073] In other embodiments, the second memory chip 200 may not receive the data clock signal WCK, and may not include the data clock pin 214, the data clock buffer 224, the phase detector 280, the EDC buffer 225, and the EDC pin 215. Instead, in order to provide a high bandwidth, the second memory chip 200 may include a greater number of DQ pins 213 than the number of DQ pins 113 of the first memory chip 100. For example, "N" may be greater than "J".

[0074] When the number of DQ pins 213 increases, the number of bits of data exchanged between the SoC 11 and the second memory chip 200 may increase. For example, it is assumed that the burst length (BL) of the first memory chip 100 and the burst length of the second memory chip 200 are the same as each other, that is, 16. In this case, the number of bits included in the DQ signal DQ[1:J] of the first memory chip 100 may be "J×16", and the number of bits included in the DQ signal DQ[1:N] of the second memory chip 200 may be "NX 16". As described above, because "N" is greater than "J", the second memory chip 200 can provide a higher bandwidth than the first memory chip 100.

[0075] In other embodiments, to provide high bandwidth, the second memory chip 200 may make the transmission rate per DQ pin higher by using the data clock signal WCK, and may include a greater number of DQ pins 213 than the number of DQ pins 113 of the first memory chip 100 .

[0076] The description is given above as the second memory chip 200 providing a higher bandwidth than the bandwidth of the first memory chip 100. However, in some embodiments, in addition to the high bandwidth, the second memory chip 200 may provide a shorter latency than the latency of the first memory chip 100. Here, the latency may include tRCD (RAS to CAS delay), tRP (row precharge time), tRAS (row active time), tRC (row cycle time), tRRD (RAS to RAS delay), tRFC (refresh cycle time), tWR (write recovery time), tWTR (write to read delay), CL (CAS latency), tCCD (CAS to CAS delay), R1, WL, and the like.

[0077] “M” indicating the number of CA signals CA[1:M], the number of CA pins 212, and the number of CA buffers 222 of the second memory chip 200 may be greater than “I” indicating the number of CA signals CA[1:I], the number of CA pins 112, and the number of CA buffers 122 of the first memory chip 100. Therefore, the second memory chip 200 may receive more commands and address bits during the same time period than the first memory chip 100, and may decode commands more quickly than the first memory chip 100.

[0078] The bandwidth of the second memory chip 200 may be different from the bandwidth of the first memory chip 100. The latency of the second memory chip 200 may be different from the latency of the first memory chip 100. The capacity of the second memory chip 200 may be different from the capacity of the first memory chip 100. For example, the number of memory banks 261 of the second memory chip 200 may be different from the number of memory banks 161 of the first memory chip 100, and the capacity of each of the memory banks 261 may be different from the capacity of each of the memory banks 161.

[0079] The power consumption of the second memory chip 200 may be different from the power consumption of the first memory chip 100. For example, the speed at which the volatile memory cells of the memory bank 261 operate may be higher than the speed at which the volatile memory cells of the memory bank 161 operate. In addition, the amount of current required to drive the volatile memory cells of the memory bank 161 may be less than the amount of current required to drive the volatile memory cells of the memory bank 261.

[0080] Figure 4 FIG. 1 shows a block diagram of an electronic device according to an embodiment of the inventive concept. The electronic device 1000 includes a SoC 1100 and a storage device 1200. Here, a first memory chip 1230 and a second memory chip 1240 of the storage device 1200 may be a reference Figures 1 to 3 A first memory chip 100 and a second memory chip 200 are described.

[0081] refer to Figure 4 SoC 1100 may communicate with one memory package (ie, memory device 1200 ) through a first channel CH1 and a second channel CH2 having different bandwidths. SoC 1100 may include a first processor 1110 , a second processor 1120 , a first memory controller 1130 , and a second memory controller 1140 .

[0082] For example, the first processor 1110 may include at least one of a central processing unit (CPU), an image signal processing unit (ISP), and a digital signal processing unit (DSP). The second processor 1120 may support applications different from those supported by the first processor 1110 (e.g., virtual reality (VR), augmented reality (AR), mixed reality (MR), etc.). For example, the amount of calculation required for the application supported by the second processor 1120 may be greater than the amount of calculation required for the application supported by the first processor 1110. Therefore, the second processor 1120 may communicate with a memory chip that provides a higher bandwidth than the memory chip of the first processor 1110. For example, the second processor 1120 may include at least one of a graphics processing unit (GPU), a visual processing unit (VPU), and a neural processing unit (NPU).

[0083] The first memory controller 1130 may provide an interface for performing communication with the first memory chip 1230 through the first channel CH1. Under the control of the first processor 1110, the first memory controller 1130 may transfer data provided from the first processor 1110 to the first memory chip 1230 or may transfer data provided from the first memory chip 1230 to the first processor 1110.

[0084] For example, the first memory controller 1130 may transmit a first clock signal CK1, a first reset signal RESET1, a first clock enable signal CKE1, and a first CA signal CA1[1:1] to the first memory chip 1230 through a first channel CH1. Here, the first reset signal RESET1 may be a signal for resetting the first memory chip 1230, and the first clock enable signal CKE1 may be a signal for determining whether to activate a buffer of the first memory chip 1230 in a power-down mode or a self-refresh mode. Figure 4 In the example, CMD1 / ADD1 may indicate a command and an address included in the first CA signal CA1[1:I].

[0085] The first memory controller 1130 may transmit a first DQ signal DQ1[1:J] including write data to the first memory chip 1230 through the first channel CH1. The first memory controller 1130 may receive a first DQ signal DQ1[1:J] including read data from the first memory chip 1230 through the first channel CH1. Figure 4 In the example, WORD1 may indicate write data or read data of the first memory chip 1230. The first channel CH1 may include a transmission path electrically connected to a clock pin, a CA pin, and a DQ pin of the first memory chip 1230, and may be used in a reference signal such as Figure 1 The first channel CH1 is formed in each of the described substrate 300 and the substrate 13 .

[0086] As in the first memory controller 1130, under the control of the second processor 1120, the second memory controller 1140 may control the second memory chip 1240. For example, the second memory controller 1140 may transmit the second clock signal CK2, the second reset signal RESET2, the second clock enable signal CKE2, and the second CA signal CA2[1:M] to the second memory chip 1240 through the second channel CH2. In order to drive the second memory chip 1240 providing a higher bandwidth than the first memory chip 1230, the second memory controller 1140 may further transmit the second data clock signal WCK2 to the second memory chip 1240 through the second channel CH2.

[0087] Also, the second memory controller 1140 may transmit a second DQ signal DQ2[1:N] including write data to the second memory chip 1240 through the second channel CH2. The second memory controller 1140 may receive a second DQ signal DQ2[1:N] including read data from the second memory chip 1240 through the second channel CH2. Figure 4 In the example, WORD2 may indicate write data or read data of the second memory chip 1240. The second channel CH2 may include a transmission path electrically connected to the clock pin, CA pin, and DQ pin of the second memory chip 1240, and may be used in, for example, reference Figure 1 The second channel CH2 is formed in each of the substrate 300 and the substrate 13 described.

[0088] In an embodiment, Figure 4 Unlike shown in , the first memory controller 1130 and the second memory controller 1140 may transmit the same clock signal CK to the first memory chip 1230 and the second memory chip 1240, respectively. Alternatively, the first memory controller 1130 and the second memory controller 1140 may transmit a common clock signal CK to the memory device 1200, and the common clock signal CK may be shared by the first memory chip 1230 and the second memory chip 1240.

[0089] The bus 1180 may provide a data input / output path between the first processor 1110, the second processor 1120, the first memory controller 1130, and the second memory controller 1140. For example, the bus 1180 may be implemented using Advanced Microcontroller Bus Architecture (AMBA), Advanced High-Performance Bus (AHB), Advanced Peripheral Bus (APB), Advanced eXtensible Interface (AXI), Advanced System Bus (ASB), AXI Coherence Extension (ACE), or a combination thereof (but not limited thereto).

[0090] Figure 5 The figure shows Figure 4 The timing diagram of the operation of the storage device processing the read command of the SoC. Figure 4 To describe Figure 5 .

[0091] The first memory controller 1130 may transmit a first clock signal CK1 to the first memory chip 1230 through a first channel CH1, and the second memory controller 1140 may transmit a second clock signal CK2 to the second memory chip 1240 through a second channel CH2. Here, it is assumed that the frequency of the first clock signal CK1 is the same as the frequency of the second clock signal CK2.

[0092] The first memory controller 1130 may transmit a first CA signal CA1[1:I] synchronized with a first clock signal CK1 to the first memory chip 1230 through a first channel CH1, and the second memory controller 1140 may transmit a second CA signal CA2[1:M] synchronized with a second clock signal CK2 to the second memory chip 1240 through a second channel CH2. The transmission rate of the first CA signal CA1[1:I] may be the same as the frequency of the first clock signal CK1 (i.e., a single data rate (SDR)). The transmission rate of the second CA signal CA2[1:M] may be the same as twice the frequency of the second clock signal CK2 (i.e., a double data rate (DDR)). For a short latency of the second memory chip 1240, the second memory controller 1140 may generate the second CA signal CA2[1:M] faster than the first memory controller 1130.

[0093] Compared with the first memory controller 1130, the second memory controller 1140 may further transmit the second data clock signal WCK2 to the second memory chip 1240 through the second channel CH2. In an embodiment, the frequency of the second data clock signal WCK2 may be twice the frequency of the second clock signal CK2.

[0094] exist Figure 5 In the embodiment, for the convenience of description, it is assumed that the first DQ signal DQ1[1:J] is aligned to the first clock signal CK1, the second DQ signal DQ2[1:N] is aligned to the second data clock signal WCK2, and the second data clock signal WCK2 is aligned to the second clock signal CK2. Moreover, it is assumed that the first memory chip 1230 outputs read data in the DDR mode by using the first clock signal CK1 and the second memory chip 1240 outputs read data in the QDR mode by using the second data clock signal WCK2. In addition, it is assumed that both the first memory chip 1230 and the second memory chip 1240 activate any word line in response to the activation command before the time point T1.

[0095] At time point T1, the first memory chip 1230 may sample (or latch) the first CA signal CA1[1:I] at the rising edge of the first clock signal CK1. As in the first memory chip 1230, at time point T1, the second memory chip 1240 may sample (or latch) the second CA signal CA2[1:M] at the rising edge of the second clock signal CK2.

[0096] At time point T2, the second memory chip 1240 may sample (or latch) the second CA signal CA2[1:M] at the falling edge of the second clock signal CK2. The second memory chip 1240 may decode the second CA signal CA2[1:M] sampled at time points T1 and T2, and may check the second read command RD2 and the read address.

[0097] At time point T3, the first memory chip 1230 may sample the first CA signal CA1[1:I] at a rising edge of the first clock signal CK1. The first memory chip 1230 may decode the first CA signal CA1[1:I] sampled at time points T1 and T3, and may check the first read command RD1 and the read address.

[0098] In an embodiment, as described above, the second memory chip 1240 may provide a shorter latency than that of the first memory chip 1230. To this end, the second memory chip 1240 may perform sampling operations at both rising and falling edges of the second clock signal CK2, thereby reducing the time taken to decode the second read command RD2.

[0099] In order to reduce the number of CA pins and CA buffers described above, the first memory chip 1230 and the second memory chip 1240 may repeatedly perform sampling operations. If the number of the first CA signals CA1[1:I] and the second CA signals CA2[1:M] is large enough, then Figure 5 Unlike shown in FIG. 1 , the first memory chip 1230 and the second memory chip 1240 may respectively perform a sampling operation once to check the first read command RD1 and the second read command RD2 .

[0100] At time point T4, the second memory chip 1240 may output a second DQ signal DQ2[1:N] including second read data corresponding to the second read command RD2 to the second memory controller 1140. For the convenience of description, a timing diagram associated with one of the second DQ signals DQ2[1:N] is shown. Here, the interval between time point T1 and time point T4 may be the RL of the second memory chip 1240. "RL" may include the time when a bit corresponding to the read data is pre-fetched from the memory bank, the time taken to serialize the bit, and the like.

[0101] At time point T5, the second memory chip 1240 may complete the output of the second read data. Here, it is assumed that the burst length BL of the second memory chip 1240 is 16, but the burst length may be changed by the MRS command. In an embodiment, in the remaining period outside the period defined by time points T4 and T5, the second DQ signal DQ2[1:N] may be set to a level corresponding to logic "1".

[0102] At time point T5, the first memory chip 1230 may output a first DQ signal DQ1[1:J] including first read data corresponding to the first read command RD1 to the first memory controller 1130. For the convenience of description, a timing diagram associated with one of the first DQ signals DQ1[1:J] is shown. Here, the interval between time point T1 to time point T5 may be the RL of the first memory chip 1230. That is, the RL of the second memory chip 1240 may be shorter than the RL of the first memory chip 1230.

[0103] At time point T6, the first memory chip 1230 may complete the output of the first read data. As described above, it is assumed that the burst length BL of the first memory chip 1230 is 16, but the burst length may be changed by the MRS command. In an embodiment, in the remaining period outside the period defined by time points T5 and T6, the first DQ signal DQ1[1:J] may be set to a level corresponding to logic "0".

[0104] from Figure 5 It can be understood that the speed of transmitting the second DQ signal DQ2[1:N] through the second channel CH2 is higher than the speed of transmitting the first DQ signal DQ1[1:J] through the first channel CH1. As an example, the bandwidth of the first channel CH1 and the bandwidth of the second channel CH2 are demonstrated in the following Table 1 for comparison.

[0105] [Form 1]

[0106]

[0107] Referring to Table 1, the frequency of the first clock signal CK1 and the frequency of the second clock signal CK2 are 1.75 GHz, and the frequency of the second data clock signal WCK2 is 3.5 GHz. Because the first memory chip 1230 outputs read data in DDR mode by using the first clock signal CK1, the transfer rate per pin is 3.5 Gbps (gigabits per second). In contrast, because the second memory chip 1240 outputs read data in QDR mode by using the second data clock signal WCK2, the transfer rate per pin is 14.0 Gbps. Assuming that both the first memory chip 1230 and the second memory chip 1240 are x16 devices, the bandwidth of the first memory chip 1230 is 7.0 GBps (gigabytes per second), and the bandwidth of the second memory chip 1240 is 28.0 GBps. Since the first memory chip 1230 is assigned to the first channel CH1 and the second memory chip 1240 is assigned to the second channel CH2, the bandwidth of the first memory chip 1230 is the bandwidth of the first channel CH1, and the bandwidth of the second memory chip 1240 is the bandwidth of the second channel CH2. Figure 5 In the example of Table 1, the bandwidth of the second memory chip 1240 may be four times the bandwidth of the first memory chip 1230 .

[0108] In an embodiment, the second memory chip 1240 may output read data in DDR mode by using the second data clock signal WCK2. In this case, in Table 1, the frequency of the second data clock signal WCK2 may be 7.0 GHz. Even if the second memory chip 1240 operates in DDR mode, the bandwidth of the second memory chip 1240 may not change because the frequency of the second data clock signal WCK2 is increased by two times.

[0109] In an embodiment, one or more memory chips may be assigned to each of the first channel CH1 and the second channel CH2. In this case, the bandwidth of the channel may increase in proportion to the number of assigned memory chips.

[0110] Figure 6 The diagram shows Figure 4 A flowchart of an operating method of an electronic device. Figure 4 To describe Figure 6 The operation method of the electronic device may roughly include a mode register setting phase S110, a normal operation phase S120, and a self-refresh phase S130. However, the order of the phases S110 to S130 is not limited to Figure 6For example, the entry into the mode register setting phase S110 may be performed after the normal operation phase S120, and the entry into the self-refresh phase S130 may be performed after the mode register setting phase S110.

[0111] In operation S111, the SoC 1100 transmits an MRS command to the first memory chip 1230 through the first channel CH1. In operation S112, the first memory chip 1230 sets an operation mode for the first channel CH1 in response to the MRS command. The first memory chip 1230 may update a setting value of a mode register in response to the MRS command.

[0112] In operation S113, the SoC 1100 transmits an MRS command to the second memory chip 1240 through the second channel CH2. In operation S114, the second memory chip 1240 sets an operation mode for the second channel CH2 in response to the MRS command. The second memory chip 1240 may update a setting value of the mode register in response to the MRS command.

[0113] The order of operation S111 and operation S113 is not limited to Figure 6 The SoC 1100 may transmit the MRS command to the memory device 1200 separately through the first channel CH1 and the second channel CH2.

[0114] In operation S121, the SoC 1100 transmits a command for accessing the first memory chip 1230 to the first memory chip 1230 through the first channel CH1. In operation S122, the first memory chip 1230 processes the transmitted command. Here, the command may be a command associated with data stored or to be stored in the first memory chip 1230, such as an activation command, a read command, a write command, a precharge command, and the like.

[0115] In operation S123, the SoC 1100 transmits a command for accessing the second memory chip 1240 to the second memory chip 1240 through the second channel CH2. In operation S124, the second memory chip 1240 processes the transmitted command. Here, the command may be a command associated with data stored or to be stored in the second memory chip 1240. The order of operation S121 and operation S123 is not limited to Figure 6 The SoC 1100 may transmit commands to the memory device 1200 separately through the first channel CH1 and the second channel CH2.

[0116] In the normal operation phase S120, the SoC 1100 may issue a refresh command for retaining data stored in the memory device 1200. In this case, the memory device 1200 may perform a refresh operation based on the refresh command issued by the SoC 1100.

[0117] However, if the SoC 1100 does not need to access the memory device 1200 during a certain time, the SoC 1100 may issue a self-refresh command to the memory device 1200. In response to the self-refresh command, each of the first and second memory chips 1230 and 1240 of the memory device 1200 may automatically perform a refresh operation to retain data.

[0118] In an embodiment, the bandwidth of the first memory chip 1230 may be lower than the bandwidth of the second memory chip 1240, but the power consumption of the first memory chip 1230 may be less than the power consumption of the second memory chip 1240. Therefore, the SoC 1100 may copy the data stored in the second memory chip 1240 to the first memory chip 1230, and then may transmit a self-refresh command to the first memory chip 1230 through the first channel CH1.

[0119] In operation S131, the SoC 1100 transmits a read command for copying data stored in the second memory chip 1240 to the second memory chip 1240 through the second channel CH2. In operation S132, the second memory chip 1240 of the memory device 1200 transmits read data to the SoC 1100 through the second channel CH2 in response to the read command.

[0120] In operation S133, the SoC 1100 transmits a write command together with the read data transmitted in operation S132 to the first memory chip 1230 through the first channel CH1. The read data transmitted in operation S132 may be new write data. The first memory chip 1230 of the storage device 1200 may store the new write data, and the data stored in the second memory chip 1240 may be copied to the first memory chip 1230 accordingly.

[0121] In operation S134, the SoC 1100 transmits a self refresh command to the first memory chip 1230 through the first channel CH1. The first memory chip 1230 may retain data stored before operation S133 and data copied in operation S133 in response to the self refresh command.

[0122] In an embodiment, before performing operation S131, the SoC 1100 may determine whether the data stored in the second memory chip 1240 can be stored in the first memory chip 1230, considering the data stored in the first memory chip 1230. The SoC 1100 may perform operations S131 to S134 based on the result of the determination. That is, if it is determined that the data stored in the second memory chip 1240 can be stored in the first memory chip 1230, the SoC 1100 may perform operations S131 to S134.

[0123] Figure 7 FIG. 1 is a block diagram of an electronic device according to another embodiment of the present invention. Figures 2 to 4 To describe Figure 7 . Figure 7 Electronic devices 2000 and Figure 4 The description is given based on the differences between the electronic devices 1000, and for the sake of brevity, descriptions of similar components and operations may be omitted below.

[0124] refer to Figure 4 A description is given of each of the first memory chip 1230 and the second memory chip 1240 supporting one channel. However, referring to Figure 7 , the first memory chip 2230 supports the first channel CH1 and the second channel CH2, and the second memory chip 2240 supports the third channel CH3 and the fourth channel CH4. That is, each of the first memory chip 2230 and the second memory chip 2240 supports one or more channels. According to the number of channels supported by each of the first memory chip 2230 and the second memory chip 2240, each of the first memory chip 2230 and the second memory chip 2240 may further include a reference Figure 2 and Figure 3 The above components described.

[0125] Since the memory device 2200 supports a total of four channels CH1 to CH4, the SoC 2100 includes a first processor 2110, a second processor 2120, first to fourth memory controllers 2130, 2140, 2150 and 2160, a scheduler 2170, and a bus 2180. The first processor 2110, the second processor 2120, and the bus 2180 can be respectively connected to Figure 4 The first processor 1110, the second processor 1120, and the bus 1180 operate similarly and are similar.

[0126] The first memory controller 2130 may transmit the first clock enable signal CKE1 and the first CA signal CA1[1:1] to the first memory chip 2230 through the first channel CH1. The second memory controller 2140 may transmit the second clock enable signal CKE2 and the second CA signal CA2[1:1] to the first memory chip 2230 through the second channel CH2.

[0127] The first memory controller 2130 and the second memory controller 2140 may transmit a clock signal CK12 and a reset signal RESET12 to the first memory chip 2230. Like the clock signal CK12 and the reset signal RESET12, a signal shared by a circuit of the first channel CH1 of the first memory chip 2230 and a circuit of the second channel CH2 of the first memory chip 2230 may not be transmitted through each of the first channel CH1 and the second channel CH2.

[0128] The third memory controller 2150 and the fourth memory controller 2160 may communicate with the second memory chip 2240 through the third channel CH3 and the fourth channel CH4 , and may operate similarly and similarly to the first memory controller 2130 and the second memory controller 2140 .

[0129] The third memory controller 2150 may transmit a third clock enable signal CKE3, a third data clock signal WCK3, and a third CA signal CA3[1:M] to the second memory chip 2240 through a third channel CH3. The fourth memory controller 2160 may transmit a fourth clock enable signal CKE4, a fourth data clock signal WCK4, and a fourth CA signal CA4[1:M] to the second memory chip 2240 through a fourth channel CH4.

[0130] The third memory controller 2150 and the fourth memory controller 2160 may transmit a clock signal CK34 and a reset signal RESET34 to the second memory chip 2240. Like the clock signal CK34 and the reset signal RESET34, a signal shared by a circuit of the third channel CH3 of the second memory chip 2240 and a circuit of the fourth channel CH4 of the second memory chip 2240 may not be transmitted through each of the third channel CH3 and the fourth channel CH4.

[0131] The scheduler 2170 may determine the operation order of the first to fourth memory controllers 2130 to 2160 under the control of the first processor 2110 and the second processor 2120. The scheduler 2170 may allow the first processor 2110 and the second processor 2120 to be connected with the first to fourth memory controllers 2130 to 2160 according to the determined operation order.

[0132] Figure 8 FIG. 1 is a diagram showing an electronic device according to another embodiment of the present invention. Figures 1 to 3 To describe Figure 8 The electronic device 3000 includes a SoC 3100 , a substrate 3200 , and a storage device 3300 .

[0133] The SoC 3100 may be disposed on one surface of the substrate 3200, and solder balls or solder bumps may be disposed on one surface of the SoC 3100. The SoC 3100 and the substrate 3200 may be electrically connected to each other through the solder balls or the bumps. Figure 1 , Figure 4 and Figure 7 The SoC described.

[0134] The substrate 3200 may provide an input / output path between the SoC 3100 and the memory device 3300. For example, the substrate 3200 may be a printed circuit board, a flexible printed circuit board, a ceramic substrate, an interposer, etc. In the case where the substrate 3200 is an interposer, the substrate 3200 may be implemented by using a silicon wafer. A plurality of wires may be implemented within the substrate 3200.

[0135] The substrate 3200 can provide a reference Figure 4 and Figure 7 The wires of the described channel. Figure 8 , the figure shows a conductive line forming a first channel CH1 and a conductive line forming a second channel CH2. However, Figure 8 The number of conductive lines illustrated in FIG. 320 , the number of conductive lines of the first channel CH1 , and the number of conductive lines of the second channel CH2 are merely examples. The conductive lines of the substrate 3200 may form three or more channels.

[0136] The storage device 3300 may include a plurality of memory dies stacked in a vertical direction. For example, the storage device 3300 may be a high bandwidth memory (HBM) device providing high bandwidth. The storage device 3300 may be arranged on one surface of a substrate 3200 on which the SoC 3100 is arranged. Solder balls or solder bumps may be arranged on one surface of the storage device 3300. The storage device 3300 and the substrate 3200 may be electrically connected to each other through solder balls or bumps. The storage device 3300 includes a first memory die 3310, a second memory die 3320, a buffer die 3330, and a through silicon via TSV.

[0137] The first memory die 3310 and the second memory die 3320 may correspond to Figure 2 The first memory chip 100 and Figure 3The first memory die 3310 may include a second memory chip 200. Figure 2 The second memory die 3320 may include a first memory bank 3311 corresponding to the memory bank 161, a first through silicon via for data input / output of the first memory bank 3311, and a first TSV region 3312 in which the first through silicon via is arranged. Figure 3 A second memory bank 3321 corresponding to the memory bank 261 , a second through silicon via for data input / output of the second memory bank 3321 , and a second TSV region 3322 in which the second through silicon via is arranged.

[0138] Here, the first TSV region 3312 may indicate a region in the first memory die 3310 where through silicon vias for communication between the first memory die 3310 and the buffer die 3330 are arranged. As in the above description, the second TSV region 3322 may indicate a region in the second memory die 3320 where through silicon vias for communication between the second memory die 3320 and the buffer die 3330 are arranged.

[0139] Through silicon vias may provide electrical paths between the first memory die 3310, the second memory die 3320, and the buffer die 3330. The first memory die 3310, the second memory die 3320, and the buffer die 3330 may be electrically connected to each other through the through silicon vias. For example, the number of through silicon vias may be from hundreds to thousands, and the through silicon vias may be arranged in a matrix form.

[0140] In an embodiment, Figure 8 , a component for accessing the first storage body 3311 (eg, Figure 2 The remaining components other than the memory bank 161 of the memory bank 162 may be arranged in the buffer die 3330. Also, components for accessing the second memory bank 3321 (e.g., Figure 3 The remaining components other than the memory body 261) can be arranged in the buffer die 3330.

[0141] The buffer die 3330 may be connected to the first TSV region 3312 and the second TSV region 3322 through the through silicon via. The buffer die 3330 may receive data from the outside through the through silicon via and may transmit the received data to the first memory die 3310 and the second memory die 3320. The buffer die 3330 may receive data stored in the first memory die 3310 and the second memory die 3320 through the through silicon via and may output the received data to the outside.

[0142] The buffer die 3330 may include Figure 2The first DQ pin (not shown) corresponding to the DQ pin 113 of the first memory die 3310 and the first buffer circuit 3331 driving the first DQ pin and the first memory die 3310 are connected to the first DQ pin 113 of the first memory die 3310 and the first DQ pin 113 of the first memory die 3310. Figure 3 A second DQ pin (not shown) corresponding to the DQ pin 213 and a second buffer circuit 3332 that drives the second DQ pin and the second memory die 3320.

[0143] The first buffer circuit 3331 may include a circuit for accessing the first memory bank 3311 and a circuit for driving the first DQ pin (ie, Figure 2 The first buffer circuit 3331 may further include: a deserializer (not shown) that deserializes data transmitted from the first DQ pin and transmits the deserialized data to the first TSV region 3312; and a serializer (not shown) that serializes data transmitted from the first TSV region 3312 and transmits the serialized data to the first DQ pin.

[0144] The serializer and the deserializer may be circuits for reducing the number of DQ signals, which increases as the first memory die 3310 and the second memory die 3320 are stacked, and the serializer and the deserializer may be different from the reference 3300. Figure 2 and Figure 3 Described serializers 171 and 271 and deserializers 172 and 272. As in the first buffer circuit 3331, the second buffer circuit 3332 may include a circuit for accessing the second memory bank 3321, a circuit for driving the second DQ pin, and a serializer and deserializer for reducing the number of DQ signals.

[0145] exist Figure 8 In the embodiment, the bandwidth of the second channel CH2 may be higher than the bandwidth of the first channel CH1, and the transmission rate of each of the second DQ pins of the second buffer circuit 3332 may be higher than the transmission rate of each of the first DQ pins of the first buffer circuit 3331. Furthermore, the number of the second DQ pins may be greater than the number of the first DQ pins.

[0146] In other embodiments, Figure 8 The difference shown in Figure 2 All components of the first memory chip 100 may be arranged in the first memory die 3310, and Figure 3All components of the second memory chip 200 may be arranged in the second memory die 3320. In this case, the buffer die 3330 may include a circuit that buffers signals transmitted through the first channel CH1 and the second channel CH2 and transmits the buffered signals to the first memory die 3310 and the second memory die 3320, or buffers signals transmitted from the first memory die 3310 and the second memory die 3320 and outputs the buffered signals through the first channel CH1 and the second channel CH2.

[0147] In other embodiments, Figure 8 Unlike those shown in the figure, the buffer die 3330 can drive the first memory die 3310 and the second memory die 3320, and can operate similarly and simultaneously to the first memory die 3310 or the second memory die 3320. The buffer die 3330 may further include a memory bank and a circuit for driving the memory bank, and the memory bank included in the buffer die 3330 may be assigned to any channel.

[0148] Fig. 9 FIG. 4 shows a block diagram of an electronic device according to another embodiment of the inventive concept. The electronic device 4000 may be implemented to be capable of using or supporting a mobile industry processor interface. For example, the electronic device 4000 may be, but is not limited to, a server, a computer, a smart phone, a tablet, a personal digital assistant (PDA), a digital camera, a portable multimedia player (PMP), a wearable device, an Internet of Things (IoT) device, and the like.

[0149] The electronic device 4000 includes a SoC 4100 and a storage device 4200. The SoC 4100 includes a first processor 4110 and a second processor 4120, and may be implemented substantially as described above. Figure 1 , Figure 4 , Figure 7 and Figure 8 The memory device 4200 includes a first memory chip 4230 and a second memory chip 4240, and may be implemented substantially as described in reference Figure 1 , Figure 4 , Figure 7 and Figure 8 The storage devices described are the same.

[0150] The electronic device 4000 includes a display 4310 that communicates with the SoC 4100. The SoC 4100 may communicate with a DSI device 4315 through a display serial interface (DSI). For example, an optical deserializer DES may be implemented in the DSI device 4315.

[0151] The electronic device 4000 includes an image sensor 4320 that communicates with the SoC 4100. The SoC 4100 may communicate with a CSI device 4325 through a camera serial interface (CSI). For example, an optical serializer SER may be implemented in the CSI device 4325.

[0152] The electronic device 4000 further includes a radio frequency (RF) chip 4330 that communicates with the SoC 4100. The RF chip 4330 includes a physical layer 4331, a DigRF slave device 4332, and an antenna 4333. For example, the physical layer 4331 of the RF chip 4330 and the SoC 4100 can exchange data with each other through a DigRF interface proposed by the MIPI Alliance.

[0153] The electronic device 4000 further includes an embedded / card storage 4340. The embedded / card storage 4340 may store data provided from the SoC 4100 and may permanently store data provided from the storage device 4200. The electronic device 4000 may communicate with an external system, for example, through Worldwide Interoperability for Microwave Access (WiMAX) 4350, a Wireless Local Area Network (WLAN) 4360, an Ultra Wideband (UWB) 4370, and the like.

[0154] remove Fig. 9 In addition to the components illustrated in the figure, any other components (e.g., speakers, microphones, A-GPS, etc.) may be additionally included in the electronic device 4000. As the functions supported by the electronic device 4000 increase, many components should be arranged in the limited area of ​​the electronic device 4000. According to an embodiment of the inventive concept, the SoC 4100 can communicate with a storage device 4200 through at least two channels CH1 and CH2, rather than communicating with storage devices with different bandwidths and transfer speeds. The storage device 4200 can provide memory chips with different bandwidths and different transfer rates. Therefore, the area and cost of the electronic device 4000 can be reduced.

[0155] The memory device according to the embodiment of the present invention can provide both a memory chip with low power consumption and a memory chip providing high bandwidth, which is implemented in one package, connected to the SoC and implemented using the SoC. Therefore, the cost of manufacturing the memory device can be reduced and the efficiency of data input / output between the SoC and the memory device can be improved.

[0156] While the inventive concept has been described with reference to exemplary embodiments thereof, it will be apparent to those skilled in the art that various changes and modifications can be made thereto without departing from the spirit and scope of the inventive concept as set forth in the appended claims.

Claims

1. A storage device, include: A first volatile memory chip including a first volatile memory cell array storing first data and configured to receive or output the first data at a first bandwidth; as well as a second volatile memory chip including a second volatile memory cell array storing second data and configured to receive or output the second data at a second bandwidth different from the first bandwidth, Wherein, the first volatile memory chip further comprises: A first clock pin, receiving a first clock; a first CA pin receiving a first command and a first address for the first data in synchronization with the first clock; and a first DQ pin, receiving or outputting the first data at a first transmission rate according to the first command, and Wherein, the second volatile memory chip further comprises: A second clock pin, receiving a second clock; a second CA pin, receiving a second command and a second address for the second data in synchronization with the second clock; a second DQ pin to receive or output the second data at a second transmission rate different from the first transmission rate according to the second command; and A data clock pin receives a data clock, wherein the frequency of the data clock is higher than the frequency of the second clock.

2. The storage device according to claim 1, in, The second bandwidth is higher than the first bandwidth, and The second transmission rate is higher than the first transmission rate.

3. The storage device according to claim 1, in, The first transmission rate is twice the frequency of the first clock, and Wherein, the second transmission rate is at least twice the frequency of the data clock.

4. The storage device according to claim 1, in, The second bandwidth is higher than the first bandwidth, and The number of the second DQ pins is greater than the number of the first DQ pins.

5. The storage device according to claim 1, further comprising: include: The substrate includes a first transmission path of a first channel respectively connected to the first clock pin, the first CA pin and the first DQ pin, and a second transmission path of a second channel respectively connected to the second clock pin, the second CA pin and the second DQ pin.

6. The storage device according to claim 5, in, The second volatile memory chip is stacked on the substrate, and The first volatile memory chip is stacked on the second volatile memory chip.

7. A storage device, include: A first volatile memory die including a first volatile memory cell array storing first data and a first through silicon via for input / output of the first data; a second volatile memory die comprising a second volatile memory cell array storing second data and a second through silicon via for input / output of the second data; as well as a buffer die configured to receive the first data through the first through silicon via and output the first data at a first bandwidth, and to receive the second data through the second through silicon via and output the second data at a second bandwidth different from the first bandwidth, Wherein, the buffer die comprises: A first DQ pin receives or outputs the first data at a first transmission rate; A second DQ pin receives or outputs the second data at a second transmission rate; a first buffer circuit configured to drive the first DQ pin, wherein the first buffer circuit includes a first serializer configured to serialize first data transmitted through the first through silicon via and transmit the serialized first data to the first DQ pin; and A second buffer circuit is configured to drive the second DQ pin, wherein the second buffer circuit includes a second serializer configured to serialize second data transmitted through the second through silicon via and transmit the serialized second data to the second DQ pin.

8. The storage device according to claim 7, in, The second bandwidth is higher than the first bandwidth, and The second transmission rate is higher than the first transmission rate.

9. The storage device according to claim 7, in, The second bandwidth is higher than the first bandwidth, and The number of the second DQ pins is greater than the number of the first DQ pins.

10. An electronic device, include: A system on chip, comprising a first processor and a second processor; as well as a storage device comprising a first volatile memory chip communicating with the first processor via a first channel and a second volatile memory chip communicating with the second processor via a second channel, wherein a first bandwidth of the first channel and a second bandwidth of the second channel are different from each other, and Wherein, the first volatile memory chip further comprises: A first clock pin, receiving a first clock; a first CA pin receiving a first command and a first address for first data in synchronization with the first clock; and a first DQ pin, receiving or outputting the first data at a first transmission rate according to the first command, and Wherein, the second volatile memory chip further comprises: A second clock pin, receiving a second clock; a second CA pin for receiving a second command and a second address for second data in synchronization with the second clock; a second DQ pin to receive or output the second data at a second transmission rate different from the first transmission rate according to the second command; and A data clock pin receives a data clock, wherein the frequency of the data clock is higher than the frequency of the second clock.

11. The electronic device according to claim 10, in, The system on chip further comprises: a first memory controller configured to transmit a first clock and a first command synchronized with the first clock to the first volatile memory chip through the first channel under the control of the first processor; and The second memory controller is configured to transmit a second clock and a second command synchronized with the second clock to the second volatile memory chip through the second channel under the control of the second processor.

12. The electronic device according to claim 11, in, The system on chip further comprises: A scheduler is configured to determine an operation order of the first memory controller and the second memory controller.

13. The electronic device according to claim 11, in, The second bandwidth is higher than the first bandwidth, and The second memory controller further transmits a data clock to the second volatile memory chip through the second channel, and a frequency of the data clock is higher than a frequency of the second clock.

14. The electronic device according to claim 13, in, The system on chip is configured to determine whether data stored in the second volatile memory chip can be stored in the first volatile memory chip.

15. The electronic device according to claim 10, in, The memory device is a memory package in which the first volatile memory chip and the second volatile memory chip are stacked.