Storage system and method of operating the same
By dynamically managing the power of computing devices in the storage system, the problem of improper power consumption and performance management in the prior art is solved, and efficient resource allocation and performance optimization are achieved.
Patent Information
- Application Number
- CN201911218412.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2018-12-07
- Filing Date
- 2019-12-02
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2039-12-02
AI Technical Summary
When managing the power of computing equipment, existing storage systems are difficult to dynamically adjust to optimize performance and power consumption, and the intervention of the host may lead to improper allocation of hardware resources, affecting performance and power consumption management.
A storage system, including nonvolatile memory and a storage controller, is provided to optimize power consumption and performance by monitoring the operating status of a computing device, dynamically manage its power, adjusting driving voltage, clock signal frequency and hardware resource utilization.
It realizes reducing power consumption and maximizing hardware resources for data processing without reducing computing device performance, improving system performance and efficiency.
Smart Images

Figure CN111290564B_ABST
Abstract
Description
[0001] CROSS-REFERENCE TO RELATED APPLICATIONS
[0002] This application claims the benefit of Korean Patent Application No. 10-2018-0157475, entitled “Memory System and Operating Method Thereof,” filed on December 7, 2018, in the Korean Intellectual Property Office, which is hereby incorporated by reference in its entirety. Technical Field
[0003] Embodiments relate to a memory system, and more particularly to a memory system including a computing device and a method of managing power of the memory system. Background Art
[0004] Flash memory is a non-volatile memory that can retain data stored therein even when power is off. Widely used storage devices including flash memory, such as embedded multimedia cards (eMMC), universal flash memory (UFS), solid-state drives (SSDs), or memory cards, have been effectively used to store or move large amounts of data. Computing systems tend to be heterogeneous systems, including a host responsible for overall system management and a computing device assigned to a specific task for accelerated processing. The computing device can be implemented as a storage acceleration platform installed in the storage device, or can form a module with the storage device. Summary of the invention
[0005] According to one aspect, a storage system is provided, comprising: a nonvolatile memory that stores data; a computing device that performs data processing on input data provided from the nonvolatile memory or a host outside the storage system; and a controller that controls write operations and read operations on the nonvolatile memory, monitors an operating state of the computing device when the computing device is performing data processing, and dynamically manages power of the computing device based on the monitoring results.
[0006] According to another aspect, a method for operating a storage system including a computing device and a storage controller is provided, the method comprising: performing data processing by the computing device on received input data; receiving, by the storage controller, status information about an operating status of the computing device while the computing device is performing the data processing; and managing, by the storage controller, power of the computing device based on the received status information while the computing device is performing the data processing.
[0007] According to another aspect, a storage system is provided, comprising: a storage device including a non-volatile memory and a storage controller, the storage controller being used to control writing and reading for the non-volatile memory; and a computing device, being used to communicate with the storage controller and perform data processing on received input data, wherein the storage controller manages power of the computing device based on an operating state of the computing device. BRIEF DESCRIPTION OF THE DRAWINGS
[0008] Features will become apparent to those skilled in the art by describing in detail exemplary embodiments with reference to the accompanying drawings, in which:
[0009] Figure 1 illustrates a storage system and a data processing system according to an embodiment;
[0010] Figure 2 A storage system according to an example embodiment is illustrated;
[0011] Figure 3 A storage controller according to an example embodiment is illustrated;
[0012] Figure 4 A computing device according to an example embodiment is illustrated;
[0013] Figure 5 A method of operating a storage system according to an example embodiment is illustrated;
[0014] Figure 6 A method of operating a storage system according to an example embodiment is illustrated;
[0015] Fig. 7A and Figure 7B illustrates a diagram showing an example of dynamically adjusting a voltage level of a driving voltage and a frequency of a clock signal provided to a computing device in a memory system according to an embodiment;
[0016] Fig. 8A and Figure 8B An example of adjusting hardware utilization of computing devices in a storage system according to an embodiment is illustrated;
[0017] Fig. 9 An example of adjusting the transmission rate of input data to be transmitted to a computing device in a storage system according to an embodiment is illustrated;
[0018] FIG. 10A to FIG. 10D illustrates a table of management data according to an example embodiment;
[0019] Fig.11 A method of operating a storage system according to an example embodiment is illustrated;
[0020] Fig.12 A flow chart illustrating a method of operating a storage system according to an example embodiment;
[0021] Fig.13 illustrates a storage system and a data processing system according to an embodiment;
[0022] Fig.14 A storage system according to an example embodiment is illustrated;
[0023] Fig.15 illustrates a data processing system according to an example embodiment;
[0024] Fig.16 illustrates a data processing system according to an example embodiment; and
[0025] Fig.17 A solid state drive (SSD) system according to an example embodiment is illustrated. DETAILED DESCRIPTION
[0026] Figure 1 The storage system 100 and the data processing system 1000 according to the embodiment are illustrated. The data processing system 1000 can be embedded in an electronic device or implemented by an electronic device. The electronic device can be implemented as, for example, a personal computer (PC), a data server, an ultra-mobile PC (UMPC), a workstation, a netbook, a network attached storage (NAS), a smart TV, an Internet of Things (IoT) device, or a portable electronic device. The portable electronic device can be a laptop computer, a mobile phone, a smart phone, a tablet PC, a personal digital assistant (PDA), an enterprise digital assistant (EDA), a digital still camera, a digital video camera, an audio device, a portable multimedia player (PMP), a personal navigation device (PND), an MP3 player, a handheld game controller, an e-book, a wearable device, etc.
[0027] refer to Figure 1 , the data processing system 1000 may include a storage system 100 and a host 200 .
[0028] The host 200 manages the overall operation of the data processing system 1000. The host 200 can store data in the storage system 100 and read data from the storage system 100. For example, the host 200 can send a write command and data to be written or a read command to the storage system 100. In addition, the host 200 can assign tasks and data to the storage system 100 and control the storage system 100 so that the storage system 100 performs a task by processing the data. For example, the host 200 can send a data processing command for performing a task to the storage system 100 together with data to be processed in the storage system 100, or send a data processing command for data stored in the storage system 100 to the storage system 100.
[0029] The host 200 may be implemented by a central processing unit (CPU), a processor, a microprocessor, an application processor (AP), etc., and according to an embodiment, the host 200 may be implemented by a system on chip (SoC).
[0030] According to an example embodiment, the storage system 100 may include a storage controller 110, a non-volatile memory (NVM) 120, and a computing device 130. The storage system 100 may be referred to as a storage device. The storage system 100 may store or process data in response to a command from a host 200. According to an embodiment, the storage system 100 may be implemented by a storage acceleration platform for accelerating data processing. For example, the storage system 100 may be a solid state drive (SSD), an embedded multimedia card (eMMC), an embedded UFS storage device, a UFS storage library, a compact flash (CF) card, a secure digital (SD) card, a micro SD card, a mini SD card, an extreme digital (XD) card, or a memory stick.
[0031] The NVM 120 may store data. The NVM 120 may store data provided from the host 200 or data provided from the computing device 130. The NVM 120 may include a memory cell array (MCA) including a non-volatile memory cell capable of maintaining the stored data even when the power of the memory system 100 is cut off, and the MCA may be divided into a plurality of memory blocks. The plurality of memory blocks may have a two-dimensional (2D) horizontal structure, in which the memory cells are arranged two-dimensionally on the same plane (or layer), or a three-dimensional (3D) vertical structure, in which the non-volatile memory cells are arranged three-dimensionally. The memory cell may be a single-level cell (SLC) storing 1-bit data, a multi-level cell (MLC) storing 2-bit data, a three-level cell (TLC) storing 3-bit data, or a four-level cell (QLC) storing 4-bit data. However, the memory cell is not limited thereto, and the memory cell may store 5 or more bits of data.
[0032] According to some embodiments, NVM 120 may include a plurality of dies, each of which includes an MCA, or NVM 120 may include a plurality of chips. For example, NVM 120 may include a plurality of chips, and each of the plurality of chips may include a plurality of dies. According to an embodiment, NVM 120 may include a plurality of channels, each of which includes a plurality of chips.
[0033] According to an embodiment, NVM 120 may be a NAND flash memory device. However, NVM 120 may be implemented by a resistive memory device such as a resistive random access memory (RAM), a phase change RAM, and a magnetic RAM. Hereinafter, the embodiment is described based on the assumption that NVM 120 is a NAND flash memory device.
[0034] The computing device 130 is used to perform data processing on the received data, and can perform tasks assigned by the host 200. According to an embodiment, the computing device 130 can perform parallel data processing on the received data (i.e., input data) by driving an application. The application may include multiple data operations related to task execution, such as arithmetic operations, convolution operations, polling operations, etc. For example, when the computing device 130 performs a task based on a neural network, the application may include a neural network model. Based on at least one of various types of neural networks, the neural network model may include multiple data operations, and input and output the size, weight, bias, etc. of the multiple data operations, and various types of neural networks include convolutional neural networks (CNN), regional CNN (R-CNN), regional candidate networks (RPN), recurrent neural networks (RNN), stack-based deep neural networks (S-DNN), state-space dynamic neural networks (S-SDNN), deconvolution networks, deep belief networks (DBN), restricted Boltzmann machines (RMB), full convolution networks, long short-term memory (LSTM) networks, and classification networks.
[0035] For example, the computing device 130 may be implemented by a graphics processing unit (GPU), a digital signal processor (DSP), a field programmable gate array (FPGA), a neural processing unit (NPU), etc. However, the computing device 130 is not limited thereto, and the computing device 130 may be implemented by various types of acceleration circuits (accelerators) to perform data processing required to perform the assigned tasks, such as parallel data operations.
[0036] The storage controller 110 may manage the overall operation of the storage system 100 and control the NVM 120 and the computing device 130 so that operations according to commands received from the host 200 are performed. For example, the storage controller 110 may control the NVM 120 in response to a write or read command from the host 200 so that data is written to or read from the NVM 120, and control an erase operation of the NVM 120. In addition, the storage controller 110 may manage various operations of the NVM 120, such as garbage collection, bad block management, read recycling, and read replacement, and manage the power of the NVM 120.
[0037] In addition, the storage controller 110 may control the computing device 130 so that the computing device 130 performs a task. The storage controller 110 may provide the computing device 130 with data and / or applications received from the host 200 or stored in the NVM 120. The storage controller 110 may configure or adjust the operating environment of the computing device 130. Specifically, the storage controller 110 may dynamically manage the power of the computing device 130 when the computing device 130 performs a task.
[0038] The storage controller 110 may include a power manager 30. The power manager 30 may monitor the real-time operating status of the computing device 130, such as utilization, temperature, operating time, clock signal frequency, driving voltage level, etc., and dynamically control the power of the computing device 130 based on the monitoring results. The utilization rate represents the percentage of hardware resources used per unit time, that is, the percentage of hardware resources used for actual data processing in the hardware resources for processing data. For example, the computing device 130 may include multiple processing elements ( Figure 2 PE), and at least some of the plurality of processing elements may be activated (or turned on) to process data. The utilization rate may represent a percentage of processing elements that perform data processing among the activated processing elements.
[0039] According to an embodiment, the power manager 30 may adjust at least one of a driving voltage or a clock signal frequency provided to the computing device 130. According to an embodiment, the power manager 30 may adjust a transmission rate of data to be provided to the computing device 130. According to an embodiment, the power manager 30 may adjust a utilization rate of hardware resources of the computing device 130. For example, the power manager 30 may determine a plurality of processing elements ( Figure 2 The method further comprises: determining the number of at least some processing elements in the PEs to be activated to perform data processing, and activating the at least some processing elements.
[0040] In addition, the power manager 30 may analyze the size of data (i.e., data volume) to be provided to the computing device 130 and the characteristics of the application, and configure the operating environment of the computing device 130 based on the analysis result. For example, the power manager 30 may calculate the data processing volume, i.e., the data operation volume, by analyzing the application characteristics and the data size, and determine the number of at least some of the processing elements to be activated among the plurality of processing elements based on the data operation volume. The power manager 30 may manage the power of the computing device 130 based on the data operation volume and the operating state of the computing device 130 when performing a task.
[0041] Figure 2 A storage system 100 according to an example embodiment is illustrated. Figure 2 , the memory system 100 may include a memory controller 110, an NVM 120, a computing device 130, a buffer memory 140, a voltage generator (VGEN) 150, and a clock signal generator (CGEN) 160. Reference will not be repeated. Figure 1 Description of storage controller 110 , NVM 120 , and computing device 130 .
[0042] The buffer memory 140 may operate as a buffer in the storage system 100 to temporarily store data DT. The buffer memory 140 may store data DT received from the host 200 or read from the NVM 120, and store data generated by the computing device 130, such as a task execution result, that is, output data DT_O of the computing device 130. For example, when the computing device 130 executes a task based on data stored in the NVM 120, the storage controller 110 may read the data DT from the NVM 120 and store the data DT in the buffer memory 140. When the computing device 130 executes the task, the storage controller 110 may send the data DT stored in the buffer memory 140 to the computing device 130 as input data DT_I. When the computing device 130 completes the task, the storage controller 110 may temporarily store the task execution result, that is, output data DT_O of the computing device 130, in the buffer memory 140, and send the output data DT_O to the host 200 in response to a request from the host 200.
[0043] The buffer memory 140 may be implemented by a volatile memory, such as a dynamic RAM (DRAM), a static RAM (SRAM), etc., or by various types of non-volatile memories, such as a resistive non-volatile memory, such as an MRAM, a PRAM, a resistive RAM (ReRAM), etc., a flash memory, a nano floating gate memory (NFGM), a polymer RAM (PoRAM), a ferroelectric RAM (FRAM), etc. In the present embodiment, the buffer memory 140 is provided outside the memory controller 110, but the buffer memory 140 may also be provided inside the memory controller 110.
[0044] The VGEN 150 may generate various voltages used in the memory system 100. For example, the VGEN 150 may be implemented by a power management integrated circuit (PMIC), a power management unit (PMU), etc. The VGEN 150 may generate a voltage based on an external voltage provided from the outside of the memory system 100 (e.g., the host 200). According to an embodiment, the VGEN 150 may generate a driving voltage of the NVM 120, that is, a first driving voltage Vd1, and a driving voltage of the computing device 130, that is, a second driving voltage Vd2, and provide the first driving voltage Vd1 and the second driving voltage Vd2 to the NVM 120 and the computing device 130, respectively. The VGEN 150 may adjust the voltage levels of the first driving voltage Vd1 and / or the second driving voltage Vd2 under the control of the power manager 30.
[0045] CGEN 160 may generate various clock signals used in storage system 100. For example, CGEN 160 may be implemented by an oscillator, a phase-locked loop (PLL), a delay-locked loop (DLL), etc. CGEN 160 may generate a clock signal used by NVM 120, i.e., a first clock signal CLK1, and a clock signal used by computing device 130, i.e., a second clock signal CLK2. CGEN 160 may adjust the frequency of the first clock signal CLK1 and / or the second clock signal CLK2 under the control of power manager 30.
[0046] As referenced above Figure 1 As described, when the computing device 130 is executing a task, the power manager 30 of the storage controller 110 can monitor the operating state of the computing device 130, and dynamically manage the power of the computing device 130 based on the monitoring result. The power manager 30 can receive the state information IF_S from the computing device 130 multiple times periodically or non-periodically, and manage the power of the computing device 130 based on the state information IF_S. For example, the state information IF_S may include at least one of the utilization, temperature, or operating time of the computing device 130, and may also include the clock signal frequency and the driving voltage level. The clock signal frequency and the driving voltage level may be received from the VGEN 150 and the CGEN 160, respectively.
[0047] The storage controller 110 may control other components of the storage system 100, such as the NVM 120, the computing device 130, the buffer memory 140, the VGEN 150, and the CGEN 160. The storage controller 110 may control write and read operations of the NVM 120 and data processing of the computing device 130 in response to a command CMD from the host 120.
[0048] The power manager 30 may generate a power control signal CS_PM for power control of the computing device 130 based on the state information IF_S, and transmit the power control signal CS_PM to at least one of the computing device 130 , the VGEN 150 , or the CGEN 160 .
[0049] According to an embodiment, the power manager 30 may dynamically adjust the voltage level of the second driving voltage Vd2 and / or the frequency of the second clock signal CLK2 of the computing device 130. The power manager 30 may send a power control signal CS_PM to the VGEN 150 and / or the CGEN 160 indicating a change in the voltage level of the second driving voltage Vd2 and / or the frequency of the second clock signal CLK2.
[0050] According to an embodiment, the power manager 30 may determine the number of at least some of the processing elements PE to be activated to perform data processing among the plurality of processing elements included in the computing unit 131. In this case, the processing element PE is an operation circuit for processing unit operations, which may be implemented by a core of a CPU, an arithmetic logic unit (ALU) of a GPU, a lookup table (LUT) of an FPGA, etc.
[0051] The power manager 30 may send an index range indicating the processing elements PE to be activated as a power control signal CS_PM to the computing device 130. Accordingly, the number of processing elements PE to be activated to perform data processing may be dynamically changed.
[0052] According to an embodiment, a first path P1 and a second path P2 may be provided between the storage controller 110 and the computing device 130, data such as input data DT_I and output data DT_O may be sent and received through the first path P1, and various types of control signals including state information IF_S and power control signal CS_PM may be sent and received through the second path P2. Different interface schemes may be applied to the first path P1 and the second path P2. A peripheral component interconnect express (PCIe) scheme may be used for the first path P1, and an inter-integrated circuit (I2C) interface scheme may be used for the second path P2. In an implementation, the first path P1 and the second path P2 may be isomorphic paths to which the same interface scheme is applied.
[0053] According to an embodiment, the power manager 30 may adjust the transmission rate of the input data DT_I. When the transmission rate of the input data DT_I increases, the power consumption of the computing device 130 may also increase, and when the transmission rate of the input data DT_I decreases, the power consumption of the computing device 130 may also decrease. However, when the transmission rate of the input data DT_I is lower than the data processing rate of the computing device 130, the performance of the computing device 130 may decrease. Therefore, the power manager 30 may determine the transmission rate of the input data DT_I based on the state information IF_S of the computing device 130, so that the power consumption of the computing device 130 is reduced but the performance of the computing device 130 is not deteriorated. The storage controller 110 may send the input data DT_I to the computing device 130 based on the determined transmission rate.
[0054] As reference Figure 1 and Figure 2As described, in the storage system 100, the storage controller 110 may configure the operating environment of the computing device 130 and dynamically manage the power of the computing device 130. When the computing device 130 operates based on a constant voltage and frequency or uses a plurality of processing elements PE to the maximum without considering the complexity of the application and the amount of data, the power consumption of the computing device 130 may increase, and the performance of the computing device 130 may decrease. However, in the storage system 100 according to the embodiment, the operating environment of the computing device 130 is based on the complexity of the application and the amount of data, and the power is dynamically managed based on the real-time operating state of the computing device 130, and therefore, the power consumption of the computing device 130 can be reduced without performance deterioration.
[0055] Furthermore, in the storage system 100 according to the embodiment, the storage controller 110 manages power, and the host 200 only determines task allocation and task execution time points, and therefore, intervention of the host 200 can be minimized, thereby reducing the workload of the host 200.
[0056] At the same time, if the computing device 130 directly manages power while executing a task, the hardware resources of the computing device 130 are used for power management, and therefore the performance of the computing device 130 may be reduced, but in the storage system 100 according to the embodiment, the storage controller 110 dynamically manages the power of the computing device 130, so the hardware resources of the computing device 130 can be allocated to the drive acceleration of the application, that is, parallel data processing, to the maximum extent, thereby improving the performance of the computing device 130.
[0057] Figure 3 A memory controller 110 is illustrated according to an example embodiment. Figure 3 , the memory controller 110 may include a processor 10, a RAM 20, a power manager 30, a host interface (I / F) circuit 40, a buffer I / F circuit 50, a device I / F circuit 60, and a NAND I / F circuit (or a NAND flash I / F) 70. The components of the memory controller 110, such as the processor 10, the RAM 20, the power manager 30, the host I / F circuit 40, the buffer I / F circuit 50, the device I / F circuit 60, and the NAND I / F circuit, may communicate with each other via a bus 80.
[0058] The processor 10 may include a CPU, a microprocessor, etc., and controls the overall operation of the memory controller 110. According to an embodiment, the processor 10 may be implemented by a multi-core processor, such as a dual-core, quad-core processor, etc.
[0059] The RAM 20 operates under the control of the processor 10 and may be used as an operating memory, a buffer memory, a cache memory, etc. For example, the RAM 20 may be implemented by a volatile memory such as a DRAM, an SRAM, etc., or by a non-volatile memory such as a PRAM, a flash memory, etc. The RAM 20 may store firmware and data for controlling the storage controller 110. The stored firmware and data may be driven or processed by the processor 10. The software hierarchical architecture of the storage controller 110 implemented by firmware may include an interface layer, a flash translation layer (FTL), etc.
[0060] The power manager 30 can dynamically manage the storage system including the storage controller 110 ( Figure 2 As described above, the power manager 30 can configure the computing device ( Figure 2 The power manager 30 can be implemented by software (or firmware), hardware or a combination thereof. When the power manager 30 is implemented by software, the program instructions forming the power manager 30 can be loaded onto the RAM 20 and executed by the processor 10.
[0061] The host I / F circuit 40 communicates with the external host 200 under the control of the processor 10. The host I / F circuit 40 may adopt at least one of various interface schemes, for example, Universal Serial Bus (USB), Advanced Technology Attachment (ATA), Serial ATA (SATA), Parallel ATA (PATA), Serial Attached Small Computer System Interface (SAS), High-Speed Interchip Interconnect (HSIC), Small Computer System Interface (SCSI), Peripheral Component Interconnect (PCI), PCI Express (PCIe), UFS, SD, MMC, eMMC, Dual In-line Memory Module (DIMM), Registered DIMM (RDIMM), Load Reduced DIMM (LRDIMM), Enhanced Small Disk Interface (ESDI), Integrated Drive Electronics (IDE), etc.
[0062] The buffer I / F circuit 50 provides an interface between the storage controller 110 and the buffer memory 140. Data to be stored in the NVM 120, data read from the NVM 120, or output data of the computing device 130 ( Figure 2 The DT_O) can be temporarily stored in the buffer memory 140 through the buffer I / F circuit 50.
[0063] The device I / F circuit 60 provides an interface between the storage controller 110 and the computing device 130. In addition, the device I / F circuit 60 can provide a storage controller 110 and VGEN ( Figure 2 150) and between the storage controller 110 and the CGEN ( Figure 2 The interface between 160).
[0064] The NAND I / F circuit 70 may provide an interface between the storage controller 110 and the NVM 120. For example, a mapping table, write data, and read data may be transmitted and received between the storage controller 110 and the NVM 120 through the NAND I / F circuit 70. According to an embodiment, the number of NAND I / F circuits 70 may correspond to the number of nonvolatile memory chips included in the storage system 100 or to the number of channels between the storage controller 110 and the NVM 120. According to an embodiment, the NAND I / F circuit 70 may provide a control path between the storage controller 110 and the computing device 130, such as a second path ( Figure 2 of P2).
[0065] In an implementation, the memory controller 110 may further include an error detection and correction (ECC) engine or the like to encode data to be stored in the NVM 120 and decode data read from the NVM 120 .
[0066] Figure 4 A computing device 130 is illustrated according to an example embodiment. Figure 4 , the computing device 130 may include a computing unit 131 , a memory 132 , a control logic 133 , a temperature sensor 134 , and an I / F circuit 135 .
[0067] The computing unit 131 may include a plurality of processing elements PE and execute tasks assigned to the computing device 130. An activated processing element PE among the plurality of processing elements PE, for example, a processing element PE in an on state, may execute data processing.
[0068] The memory 132 may store data received from the outside or data generated inside the computing device 130. For example, the memory 132 may store input data DT_I and intermediate data and / or output data DT_O generated when the computing unit 131 processes at least a portion of the input data DT_I. The memory 132 may also store received applications. The memory 132 may be implemented by a volatile memory, such as a DRAM, SRAM, etc., or by a non-volatile memory, such as a resistive non-volatile memory, such as an MRAM, a PRAM, or a ReRAM, etc., a flash memory, a NFGM, a PoRAM, a FRAM, etc.
[0069] The control logic 133 may control the computing unit 131 and the memory 132 so as to perform tasks based on the application. In addition, the control logic 133 may calculate the utilization of the hardware resource (i.e., the computing unit 131) and the running time of the unit data processing, and generate information about the utilization and the running time, i.e., the utilization information IF_UL and the running time information IF_R.
[0070] The temperature sensor 134 may sense the temperature inside the computing device 130 and generate temperature information IF_TP. The temperature information IF_TP may be sent to the storage controller 110. When the computing device 130 executes a task, the processing element PE operates and thus generates heat. The temperature of the computing device 130 may affect the performance and durability of the computing device 130. An increase in the temperature of the computing device 130 may cause an increase in the power consumption of the computing device 130. When the temperature increases above a threshold, the computing device 130 may be damaged. Therefore, the power and heat throttling of the computing device 130 may be dynamically managed based on the temperature sensed by the temperature sensor 134.
[0071] The I / F circuit 135 provides an interface between the memory controller 110 and the computing device 130. Through the I / F circuit 135, a control signal such as a power control signal CS_PM and input data DT_I may be received from the memory controller 110, and status information such as utilization information IF_UL, runtime information IF_R, and temperature information IF_TP or output data DT_O may be sent to the memory controller 110.
[0072] although Figure 4 The computing device 130 is shown to include one I / F circuit 135, but may include multiple I / F circuits 135. For example, when different interface schemes are applied to the data transmission path and the control signal and status information transmission path, the computing device 130 may include at least two I / F circuits 135.
[0073] Figure 5 An operating method of the memory system 100 according to an example embodiment is illustrated. Figure 5 A method of configuring the operating environment of computing device 130 is illustrated.
[0074] refer to Figure 5, a data processing command CMD_DP may be received from the host 200. According to an embodiment, the data processing command CMD_DP may include an indication of allocating an application APP and data DT associated with a task to be executed by the computing device 130. The storage controller 110 may read the application APP and data DT from the NVM 120 and store the application APP and data DT in the buffer memory 140. The storage controller 110 may send the application APP and data DT to the computing device 130. The data DT may be stored in the memory 132 of the computing device 130 as input data DT_I. According to an embodiment, the host 200 may provide at least one of the application APP or the data DT, and as described above, the application APP and data DT sent from the host 200 may be temporarily stored in the buffer memory 140 and then sent to the computing device 130.
[0075] In this case, the power manager 30 of the storage controller 110 may analyze the characteristics of the application APP and the size (ie, data volume) of the data DT, and configure the operating environment of the computing device 130 based on the analysis result.
[0076] According to an embodiment, the power manager 30 may calculate the data processing amount, that is, the data operation amount, by analyzing the characteristics of the application APP and the size of the data DT, and determine the number of at least some of the processing elements PE (that is, the processing elements PE to be activated) among the multiple processing elements PE on which data processing is to be performed based on the data operation amount. The power manager 30 may generate an index range control signal CS_IR indicating the number of at least some of the processing elements PE to be activated, and send the index range control signal CS_IR as a power management control signal to the computing device 130. The computing device 130 may activate at least some of the processing elements PE corresponding to the determined number among the multiple processing elements PE in response to the index range control signal CS_IR.
[0077] According to an embodiment, the power manager 30 may determine the voltage level of the second driving voltage Vd2 and the frequency of the second clock signal CLK2 provided to the computing device 130 based on the analysis result. In one implementation, the voltage level of the second driving voltage Vd2 and the frequency of the second clock signal CLK2 may be determined according to preset initial values, respectively.
[0078] The execution command CMD_EX may be received from the host 200, and the memory controller 110 may send the execution command CMD_EX or a control signal corresponding to the execution command CMD_EX to the computing device 130. In response to the execution command CMD_EX, the computing device 130 may perform a task by performing data processing on the input data DT_I based on the application APP.
[0079] Figure 6 An operating method of the memory system 100 according to an example embodiment is illustrated. Figure 6 A method of dynamically managing power of computing device 130 is illustrated.
[0080] refer to Figure 6 , the computing device 130 may perform data processing on the input data DT_I. The activated processing element PE, that is, the processing element PE in the turned-on state, may perform data processing on the input data DT_I according to the assigned application. The intermediate processing result DT_MO generated according to the data processing may be stored in the memory 132 and reused in the data processing. The to-be-processed data DT stored in the buffer memory 140 may be sent to the computing device 130 as the input data DT_I in a specific unit (e.g., a packet unit).
[0081] When the computing device 130 is operating, that is, when a task is being executed, the storage controller 110 may manage the power of the computing device 130. The storage controller 110 may monitor the operating state of the computing device 130 when the computing device 130 is executing a task. Specifically, the power manager 30 may monitor the operating state of the computing device 130 by receiving state information IF_S from the computing device 130. The state information IF_S may include at least one of utilization information IF_UL, runtime information IF_R, or temperature information IF_TP. The power manager 30 may also receive voltage level information of the second driving voltage Vd2 and frequency information of the second clock signal CLK2. According to an embodiment, the storage controller 110 may receive the state information IF_S periodically or non-periodically.
[0082] The power manager 30 may manage the power of the computing device 130 based on the state information IF_S. As an example, the power manager 30 may dynamically adjust at least one of the voltage level of the second driving voltage Vd2 and the frequency of the second clock signal CLK2 based on the utilization information IF_UL. For example, the power manager 30 may increase the frequency and the voltage level when the utilization of the computing device 130 is lower than the target value. The power manager 30 may send a voltage control signal CS_V and a frequency control signal CS_F indicating a voltage level change and a frequency change to VGEN 150 and CGEN 160, respectively. VGEN 150 and CGEN 160 may change the voltage level of the second driving voltage Vd2 and the frequency of the second clock signal CLK2 based on the voltage control signal CS_V and the frequency control signal CS_F, respectively.
[0083] The power manager 30 may also adjust at least one of the voltage level of the second driving voltage Vd2 and the frequency of the second clock signal CLK2 based on the temperature information IF_TP. For example, the power manager 30 may reduce the frequency and voltage level when the temperature of the computing device 130 is higher than the target value. Accordingly, the thermal throttling of the computing device 130 may be dynamically managed.
[0084] As another example, the power manager 30 may adjust the utilization of the hardware resources of the computing device 130 based on the utilization information IF_UL. For example, the power manager 30 may manage the number of at least some of the processing elements PE to be activated among the plurality of processing elements PE. For example, the power manager 30 may increase the number of at least some of the processing elements PE to be activated when the utilization is lower than the target value. The power manager 30 may send an index range control signal CS_IR indicating a change value of the number of at least some of the processing elements PE to be activated to the computing device 130. The computing unit 131 may increase the number of at least some of the processing elements PE to be activated based on the index range control signal CS_IR.
[0085] As another example, the power manager 30 may adjust the transmission rate of the input data DT_I based on the runtime information IF_R. For example, when the runtime is faster than a reference value, the transmission rate of the input data DT_I may be reduced.
[0086] Fig. 7A and Figure 7B is a diagram showing an example of dynamically adjusting a voltage level and a clock signal frequency of a driving voltage provided to a computing device in a memory system according to an embodiment. Figure 6 and Fig. 7A , the power manager 30 may dynamically change the voltage level of the driving voltage (second driving voltage Vd2) to be provided to the computing device 130. Figure 6 and Figure 7B , the power manager 30 may dynamically change the frequency of the clock signal (the second clock signal CLK2 ) to be provided to the computing device 130 .
[0087] As referenced above Figure 6 As described, the power manager 30 may periodically or non-periodically receive the state information IF_S and change the voltage level of the driving voltage based on the state information IF_S. For example, the power manager 30 may receive the state information IF_S at time point t1 and change the voltage level of the driving voltage from the first level VL1 to the second level VL2 based on the state information IF_S, and thereafter, receive the state information IF_S at time point t2 and change the voltage level of the driving voltage from the second level VL2 to the third level VL3 based on the state information IF_S.
[0088] As referenced above Figure 6 As described, the power manager 30 may receive IF_S periodically or non-periodically and change the frequency of the clock signal (i.e., the second clock signal CLK2) based on the status information IF_S. For example, the power manager 30 may receive the status information IF_S at time point t1 and change the frequency of the clock signal from the first frequency F1 to the second frequency F2 based on the status information IF_S, and thereafter, receive the status information IF_S at time point t2 and change the frequency of the clock signal from the second frequency F2 to the third frequency F3 based on the status information IF_S.
[0089] As described above, the power manager 30 may dynamically change the voltage level of the driving voltage and the frequency of the clock signal provided to the computing device 130 according to the operating state of the computing device 130 .
[0090] Fig. 8A and Figure 8B An example of adjusting hardware utilization of a computing device in a storage system according to an embodiment is shown. Fig. 8A and Figure 8B , the computing unit 131 included in the computing device may include a plurality of processing elements PE. The power manager ( Figure 6 30) can be based on the state information ( Figure 6 IF_S) reduces the number of at least some processing elements PE to be activated (such as Fig. 8A ), or the number of at least some of the processing elements PE to be activated may be increased (as shown in Figure 8B The power manager 30 may adjust the number of at least some of the processing elements PE to be activated according to the operating state of the computing device 130 .
[0091] Fig. 9 An example of adjusting the transmission rate of input data to be transmitted to a computing device in a storage system according to an embodiment is shown. Figure 6 and Fig. 9 , the input data DT_I may be transmitted from the buffer memory 140 to the computing device 130 in a specific unit, and the power manager 30 may adjust the transmission rate.
[0092] For example, the input data DT_I can be sent in packet units. The first to sixth packets PK0 to PK5 can be sent sequentially to the computing device 130. The power manager 30 can receive the status information IF_S at time point t1 and reduce the transmission rate based on the status information IF_S, so that the transmission period of the third packet PK2 and the fourth packet PK3 is T2, which is longer than the period T1 of the previous transmission of the first packet PK0 and the second packet PK1. Thereafter, the power manager 30 can receive the status information IF_S at time point t2 and increase the transmission rate based on the status information IF_S. After time point t2, the transmission rate of the input data DT_I can be increased, so that the transmission period of the fifth packet PK4 and the sixth packet PK5 is T3, which is shorter than the period T2 of the transmission of the third packet PK2 and the fourth packet PK3.
[0093] refer to Figure 6 , the power manager 30 may manage power based on the management data MDT. The management data MDT may be stored in the NVM 120, and the memory controller 110 may load the management data MDT into the buffer memory 140 or the internal RAM ( Figure 3 20), and manage power based on the state information IF_S and the reference management data MDT. The management data MDT is data for power management of the computing device 130, which may be a data table, metadata, a bit file, etc. For example, the management data MDT may include scaling factors corresponding to utilization, driving voltage, and clock signal frequency.
[0094] FIG. 10A to FIG. 10D is a table illustrating management data according to an example embodiment. Figures 10A to 10D In , A, B and C are the values of the frequency scaling factor SF_F, and D, E and F are the values of the voltage scaling factor SF_V.
[0095] refer to Fig. 10A , the management data MDTa may include a frequency scaling factor SF_F and a voltage scaling factor SF_V corresponding to the utilization rate UL. For example, when the utilization rate UL of the computing device 130 is 20%, the power manager 30 may increase the frequency of the clock signal (i.e., the second clock signal CLK2) by BMHz (megahertz) and / or increase the voltage level of the driving voltage (i.e., the second driving voltage Vd2) by EV (volts) based on the frequency scaling factor SF_F and the voltage scaling factor SF_V, respectively.
[0096] refer to Fig. 10B, the management data MDTb may include a frequency scaling factor SF_F and a voltage scaling factor SF_V corresponding to the difference between the current utilization and the target utilization (ie, the utilization increment value DUL). For example, when the current utilization is 80% and the target utilization is 60%, the power manager 30 may reduce the frequency of the second clock signal CLK2 by B MHz and / or reduce the voltage level of the second driving voltage Vd2 by EV based on the frequency scaling factor SF_F and the voltage scaling factor SF_V corresponding to the utilization increment value DUL (ie, -20%).
[0097] refer to Fig. 10C , the management data MDTc may include a frequency scaling factor SF_F and a voltage scaling factor SF_V corresponding to the temperature TP. For example, when the temperature of the computing device 130 is 50°, the power manager 30 may reduce the frequency of the second clock signal CLK2 by C MHz and / or reduce the voltage level of the second driving voltage Vd2 by FV based on the frequency scaling factor SF_F and the voltage scaling factor SF_V corresponding to the temperature TP of 50°.
[0098] refer to Fig. 10D , the management data MDTd may include a frequency scaling factor SF_F and a voltage scaling factor SF_V corresponding to the difference between the current temperature and the target temperature (ie, the temperature increment value DTP). For example, when the current temperature is 50° and the target temperature is 30°, the temperature increment value DTP is -20°, and the power manager 30 may reduce the frequency of the second clock signal CLK2 by A MHz and / or reduce the voltage level of the second driving voltage Vd2 by DV based on the frequency scaling factor SF_F and the voltage scaling factor SF_V corresponding to the temperature increment value DTP (ie, -20°).
[0099] The method of managing the power of the computing device 130 based on the status information IF_S and the management data performed by the power manager 30 has been described as an example. However, the power manager 30 may manage the power of the computing device 130 in various ways based on the status information IF_S, and the management data may be configured in various formats.
[0100] Fig.11 An operating method of the memory system 100 according to an example embodiment is illustrated. Fig.11 The method of operation of the storage system 100 after the computing device 130 completes the task is illustrated.
[0101] refer to Fig.11 When the task is completed, the task execution result (ie, output data DT_O) may be stored in the memory 132. The memory controller 110 may store the output data DT_O from the computing device 130 in the buffer memory 140.
[0102] The memory controller 110 may control the computing device 130 so that the computing device 130 is in a low power state (idle state or shutdown state). For example, the memory controller 110 may send a state switching signal CS_IDLE to the computing device 130 indicating switching to an idle state. According to an embodiment, the memory controller 110 may send a voltage control signal CS_V and a frequency control signal CS_F to VGEN 150 and CGEN 160, respectively, to control the second driving voltage Vd2 and the second clock signal CLK2, so that the voltage level of the second driving voltage Vd2 and the frequency of the second clock signal CLK2 provided to the computing device 130 have minimum values, respectively. Accordingly, the computing device 130 may switch to an idle state and the power consumption of the computing device 130 may be minimized.
[0103] The memory controller 110 may transmit the output data DT_O to the host 200. For example, the memory controller 110 may transmit the output data DT_O to the host 200 in response to a request from the host 200.
[0104] Fig.12 is a flow chart of an operation method of a storage system according to an example embodiment. Figure 1 The method of operating the storage system 100, therefore, the above reference Figures 1 to 11 The description made also applies to the present embodiment.
[0105] refer to Fig.12 In operation S110, the storage controller 110 may receive a data processing command from the host 200. The data processing command may include an instruction on allocating an application and input data associated with a task to be performed by the computing device 130. According to an embodiment, the storage controller 100 may read the application and input data from the NVM 120. According to an embodiment, the application and input data may be provided from the host 200.
[0106] In operation S120, the storage controller 110 may configure the operating environment of the computing device 130 based on the application and input data to be sent to the computing device 130. According to an embodiment, the storage controller 110 may analyze the characteristics of the application and the amount of input data. The storage controller 110 may calculate the amount of data operations based on the analysis result, and determine the hardware utilization of the computing device 130 based on the amount of data operations. For example, the storage controller 110 may determine (or adjust) the number of at least some of the processing elements to be activated to perform data processing among the plurality of processing elements included in the computing device 130.
[0107] In operation S130, the computing device 130 may perform data processing based on the received application and input data. The computing device 130 may perform data processing in the operating environment configured in operation S120.
[0108] In operation S140, the memory controller 110 may receive an operating state of the computing device 130, such as state information, while the computing device 130 is performing data processing. The memory controller 110 may receive the operating state of the computing device 130 by periodically or non-periodically monitoring the computing device 130. The memory controller 110 may receive information about characteristics of the computing device 130, such as utilization, operating time, temperature, etc., and may also receive a voltage level of a driving voltage and a frequency of a clock signal of the computing device 130. According to an embodiment, the computing device 130 may periodically send the information to the memory controller 110 in response to a request of the memory controller 110.
[0109] In operation S150, the memory controller 110 may manage the power of the computing device 130 based on the operating state of the computing device 130. Figure 6 to Figure 1 0, the memory controller 110 may dynamically manage the power of the computing device 130 based on various data indicating the operating state of the computing device 130. The computing device 130 may perform data processing in an operating environment that is changed according to the power management of the memory controller 110. For example, the memory controller 110 may change the level of a driving voltage of the computing device 130 and / or the frequency of a clock signal of the computing device 130 based on the operating state. Accordingly, the computing device 130 may perform data processing based on the changed voltage level of the driving voltage and the changed clock signal frequency.
[0110] When the data processing is completed, the computing device 130 may send the data processing result to the storage controller 110 and switch to a low power state in operation S170. According to an embodiment, the storage controller 110 may store the received data processing result in the buffer memory 140 and send a control signal for controlling the computing device 130 to a low power state (e.g., an idle state or a shutdown state) to the computing device 130.
[0111] In operation S180, the memory controller 110 may transmit the data processing result to the host 200. According to an embodiment, the memory controller 110 may transmit the data processing result to the host 200 in response to a request of the host 200.
[0112] Fig.13 The memory system 100a and the data processing system 1000a according to the embodiment are illustrated. Fig.13, the data processing system 1000a may include a host 200 and a storage system 100a, and the storage system 100a may include a storage controller 110, an NVM 120, and a computing device 130a. Fig.13 The configuration and operation of the data processing system 1000a are similar to Figure 1 Configuration and operation of data processing system 1000.
[0113] Fig.13 and Figure 1 The only difference is that the computing device 130a can directly communicate with the host 200. The computing device 130a can communicate with the storage controller 110 through the first path P1a and communicate with the host 200 through the second path P2a. The computing device 130a may include a first interface IF1 for communicating with the storage controller 110 and a second interface IF2 for communicating with the host 200.
[0114] The communication and operation between the storage controller 110 and the computing device 130a may be similar to that described above with reference to Figures 1 to 12 The storage controller 110 may monitor the operating status of the computing device 130a and dynamically manage the power of the computing device 130a.
[0115] The computing device 130a may receive data DT or commands CMD not stored in the NVM 120 from the host 200 through the second path P2a. For example, the computing device 130a may receive real-time data DT (e.g., log data) from the host 200 through the second path P2a, and process the received data DT. According to an embodiment, the computing device 130a may receive an application through the second path P2a. The computing device 130a may send a data processing result directly to the host 200 through the second path P2a. Alternatively, the computing device 130a may send a data processing result to the storage controller 110 through the first path P1a, and the storage controller 110 may send the data processing result to the host 200.
[0116] Fig.14 A storage system 100b according to an example embodiment is illustrated. Fig.14 , the storage system 100b may include a storage controller 110b, an NVM 120, and a computing device 130. The storage controller 110b may include a first controller 111 and a second controller 112. The storage system 100b may also include a buffer memory, a VGEN, a CGEN, and the like.
[0117] The first controller 111 may manage the general operation of the storage system 100b, and the second controller 112 may manage the write / read operation of the NVM 120. The first controller 111 may manage the power of the storage system 100b, for example, dynamically manage the power of the computing device 130. The first controller 111 may include a power manager 30, and the power manager 30 may periodically or non-periodically receive the state information IF_S indicating the operation state of the computing device 130, and generate a power control signal CS_PM for controlling the power of the computing device 130 based on the state information IF_S. The detailed operation of the power manager 30 has been described above.
[0118] The second controller 110 may be between the first controller 111 and the NVM 120 and may schedule a programming operation of the NVM 120. For example, the first controller 111 may communicate with the host 200 and send write data received from the host 200 to the second controller 112. The second controller 112 may buffer the data received from the first controller 111 and schedule programming of the data. According to an embodiment, the first controller 111 may convert a logical address into a physical address and send the physical address to the second controller 112. The second controller 112 may control the NVM 120 so that the write data is programmed in a storage area of the NVM 120 corresponding to the physical address.
[0119] According to an embodiment, each of the first controller 111 and the second controller 112 may be implemented by a semiconductor device (e.g., a semiconductor chip) and include at least one interface circuit for communicating with an external device. The first controller 111 may include an interface circuit for communicating with the host 200, the second controller 112, and the computing device 130, and the second controller 112 may include an interface circuit for communicating with the first controller 111 and the NVM 120.
[0120] Fig.15 FIG. 1 illustrates a data processing system 1000c according to an example embodiment. Fig.15 , the data processing system 1000c may include a host 200, a storage device 100c, and a computing device 300. The storage device 100c may include a storage controller 110 and an NVM 120. Each of the storage device 100c and the computing device 300 may be implemented by a semiconductor device (e.g., a semiconductor chip or a semiconductor module), and the storage device 100c and the computing device 300 may be referred to as a storage system.
[0121] The storage device 100c may communicate with the host 200 and directly communicate with the computing device 300 through the direct path DP. The storage device 100c may send data DT to the computing device 300 through the direct path DP and dynamically manage the power of the computing device 300. The power manager 30 included in the storage controller 110 may periodically or non-periodically receive the state information IF_S of the computing device 300, and dynamically manage the power of the computing device 300 based on the state information IF_S. The power manager 30 may generate a power control signal CS_PM for controlling the power of the computing device 300, and send the power control signal CS_PM to the computing device 300.
[0122] The computing device 300 may perform tasks based on the tasks and data assigned from the host 200. That is, the computing device 300 may perform data processing. The computing device 300 may receive applications and input data for performing tasks from the host 200. According to an embodiment, the computing device 300 may directly communicate with the host 200 and receive applications and input data from the host 200.
[0123] Fig.16 A data processing system 1000d according to an example embodiment is illustrated. Fig.16 , the data processing system 1000d may include a plurality of (first to fourth) storage systems 100_1, 100_2, 100_3, and 100_4 and a host 200. The plurality of storage systems 100_1, 100_2, 100_3, and 100_4 and the host 200 may communicate with each other via a bus 600.
[0124] The plurality of storage systems 100_1, 100_2, 100_3, and 100_4 may include storage controllers 110_1, 110_2, 110_3, and 110_4 and NVMs 120_1, 120_2, 120_3, and 120_4, respectively. At least one of the plurality of storage systems 100_1, 100_2, 100_3, and 100_4, for example, the first storage system 100_1 may include a computing device 130. Figure 1 , Fig.13 and Fig.14 The described storage systems 100 , 100 a , and 100 b may be used as a first storage system 100_1 , and the power manager 30 in the storage controller 110_1 may dynamically manage the power of the computing device 130 based on the operating state of the computing device 130 .
[0125] The computing device 130 may perform data processing on data stored in the first storage system 110_1, that is, data stored in the NVM 120_1, or data provided from another storage system 100_2, 100_3, or 100_4. For example, the second storage system 100-2 may send data DT to the first storage system 110_1, and the computing device 130 may perform data processing on the data DT received from the second storage system 100_2. According to an embodiment, the second storage system 100_2 may directly send the data DT to the first storage system 110_1 through a direct memory access (DMA) mechanism without passing through the host 200.
[0126] According to an embodiment, each of the plurality of storage systems 100_1, 100_2, 100_3, and 100_4 may be implemented by a module and installed in a slot of a mainboard. For example, the plurality of storage systems 100_1, 100_2, 100_3, and 100_4 may be installed in a PCIe slot and communicate with the host 200 and other storage systems via a PCIe bus. Because at least one of the plurality of storage systems 100_1, 100_2, 100_3, and 100_4, such as the first storage system 100_1, includes a computing device 130, a separate slot is not required to install the computing device 130. Accordingly, the data processing system 1000d according to the embodiment may include an acceleration function and maintain data storage capacity. In addition, since the storage controller 110_1 of the first storage system 100_1 dynamically manages the power of the computing device 130, the performance of the computing device 130 may be improved, and the workload of the host 200 for power management may be reduced.
[0127] Fig.17 An SSD system 2000 is illustrated according to an example embodiment. Fig.17 , the SSD system 2000 may include a host 2100 and an SSD 2200. The SSD 2200 may send and receive a signal SGL to and from the host 2100 through a signal connector, and receive power PWR through a power connector. The SSD 2200 may include an SSD controller 2210, a plurality of storage devices 2230, 2240, and 2250, a computing device 130, and an auxiliary power supply 2220. Figure 1 , Fig.13 and Fig.14The storage systems 100, 100a, and 100b of the SSD 2200 may be used as the SSD 2200. The power manager 30 included in the SSD controller 2210 may dynamically manage the power of the computing device 130 based on the operating state of the computing device 130. The auxiliary power supply 2220 may generate voltages V of various currents to be used by the SSD 2200, such as a driving voltage of the computing device 130. The power manager 30 may adjust the voltage level of the driving voltage of the computing device 130 by controlling the auxiliary power supply 2220 based on the operating state of the computing device 130. In addition, the power manager 30 may dynamically adjust the frequency of the clock signal of the computing device 130 or the hardware resource utilization of the computing device 130 based on the operating state of the computing device 130. Accordingly, the performance of the computing device 130 may be improved, and the workload of the host 2100 for power management may be reduced.
[0128] The storage system according to the above-described embodiments may be installed in or applied to not only the SSD system 2000 but also a memory card system, a computing system, a UFS, and the like.
[0129] An embodiment provides a storage system including a computing device, wherein a storage controller dynamically manages power of the computing device according to an operating state of the computing device, and also provides an operating method of the storage system.
[0130] Embodiments are described and shown in the figures in the form of functional blocks, units, modules and / or methods. It will be appreciated by those skilled in the art that these blocks, units, modules and / or methods are physically implemented by electronic (or optical) circuits, such as logic circuits, discrete devices, microprocessors, hard-wired circuits, memory elements, wiring connections, etc., which can be formed using semiconductor-based processing techniques or other manufacturing techniques. In the case where blocks, units, modules and / or methods are implemented by microprocessors or similar means, they can be programmed using software (e.g., microcode) to perform the various functions discussed herein, and can be optionally driven by firmware and / or software. Alternatively, each block, unit, module and / or method can be implemented by dedicated hardware, or as a combination of dedicated hardware that performs part of the functions and processors (e.g., one or more programmed microprocessors and their associated circuits) that perform other functions. Similarly, each block, unit and / or module of the embodiment can be physically separated into two or more interactive discrete blocks, units and / or modules without departing from the scope of the present disclosure. Further, the blocks, units and / or modules of the embodiment can be physically merged into more complex blocks, units and / or modules without departing from the scope of the present disclosure.
[0131] Example embodiments have been described herein, and although specific terms are employed, they are used and interpreted in a general and descriptive sense only, and not for limiting purposes. In some instances, it will be apparent to one of ordinary skill in the art at the time of filing this application that features, characteristics, and / or elements described in conjunction with a particular embodiment may be used alone or in combination with features, characteristics, and / or elements described in conjunction with other embodiments, unless otherwise noted. Accordingly, those skilled in the art will appreciate that various changes in form and detail may be made without departing from the spirit and scope of the invention as set forth in the following claims.
Claims
1. A storage system, include: Non-volatile memory, storing data; a computing device configured to receive an application and input data provided from the nonvolatile memory or a host external to the storage system, and perform data processing on the input data according to the application; as well as The controller is configured to control write operations and read operations of the non-volatile memory, monitor the operation status of the computing device when the computing device is performing data processing, and dynamically manage the power of the computing device according to the monitoring result.
2. The storage system according to claim 1, wherein the controller is further configured to periodically or non-periodically receive information about at least one of hardware utilization, temperature, or operating time from the computing device multiple times, and dynamically perform power management for the computing device based on the received information. 3 . The memory system of claim 2 , wherein the controller is further configured to adjust at least one of a voltage level of a driving voltage of the computing device or a frequency of a clock signal of the computing device. 4 . The memory system of claim 3 , wherein the controller is further configured to adjust at least one of a voltage level of the driving voltage or a frequency of the clock signal based on the received information about hardware utilization. 5 . The memory system according to claim 3 , wherein the controller is further configured to receive real-time temperature information, and dynamically adjust at least one of a voltage level of the driving voltage or a frequency of the clock signal based on the received real-time temperature information.
6. The storage system of claim 2, wherein the computing device comprises a plurality of processing elements, and The controller is further configured to determine, based on the received information, a number of at least some of the processing elements included in the computing device to be activated to perform data processing. 7 . The storage system according to claim 2 , wherein the controller is further configured to send the input data to the computing device when the computing device is performing data processing, and adjust a transmission rate of the input data.
8. The storage system of claim 2, wherein the controller is further configured to dynamically perform power management based on the received information and management data preconfigured for power management of the computing device.
9. The memory system of claim 8, wherein the management data comprises a voltage scaling factor of a driving voltage and a frequency scaling factor of a clock signal, the voltage scaling factor and the frequency scaling factor being determined for each of a plurality of values regarding hardware utilization.
10. The storage system of claim 1, wherein the controller is further configured to analyze characteristics of the application and a data volume of the input data, the characteristics and the data volume being provided to the computing device for data processing, and the controller is further configured to determine an initial hardware utilization of the computing device based on the analysis result. 11 . The storage system according to claim 1 , wherein the controller is further configured to receive a data processing result from the computing device after data processing is completed, and send the data processing result to the host. 12 . The storage system of claim 1 , wherein the controller is further configured to control the computing device to be in a low power state after data processing is completed.
13. The storage system according to claim 1, further comprising a first path and a second path between the computing device and the controller, The controller is further configured to send the input data to the computing device through the first path, and send a control signal for power management to the computing device through the second path.
14. The storage system according to claim 1, wherein the controller include: a first controller configured to communicate with the host and manage power of the computing device; as well as The second controller is configured to control writing and reading of data to and from the non-volatile memory.
15. The storage system of claim 1, wherein the computing device comprises hardware resources implemented as a field programmable gate array (FPGA).
16. A method for operating a storage system comprising a computing device and a storage controller, the method comprising: include: receiving, by the computing device, an application and input data, and performing data processing on the input data according to the application; receiving, by the storage controller, state information about an operating state of the computing device while the computing device is performing data processing; as well as While the computing device is performing data processing, power of the computing device is managed by the storage controller based on the received status information.
17. The operating method according to claim 16, wherein receiving status information include: Information on at least one of hardware utilization, temperature, or operating time is received from the computing device multiple times periodically or aperiodically.
18. The operating method of claim 17, wherein managing power comprises changing at least one of the following based on the information received multiple times: a voltage level of a driving voltage of the computing device, a frequency of a clock signal of the computing device, or a number of at least some processing elements to be activated among a plurality of processing elements included in the computing device.
19. The operating method according to claim 16, further comprising: include: An operating environment of the computing device is configured by the storage controller based on the application and the input data.
20. The operating method according to claim 16, further comprising: include: receiving, by the storage controller, a processing result according to data processing from the computing device; as well as The computing device is switched to a low power state.
21. A storage system, include: A storage device comprising a non-volatile memory and a storage controller, wherein the storage controller is configured to control writing and reading to the non-volatile memory; as well as a computing device configured to communicate with the storage controller, receive an application and input data, and perform data processing on the input data according to the application, The storage controller is further configured to manage power of the computing device based on an operating state of the computing device when the computing device is performing data processing.
22. The storage system of claim 21, wherein the storage controller is further configured to periodically or non-periodically receive operating status information from the computing device multiple times while the computing device is performing data processing, and dynamically change a voltage level of a driving voltage of the computing device and a frequency of a clock signal of the computing device based on the received operating status information.
23. The storage system of claim 22, wherein the storage controller is further configured to dynamically adjust the number of processing elements for performing data processing among a plurality of processing elements included in the computing device based on the received operation status information.
24. The memory system of claim 22, wherein the memory controller is further configured to adjust a transmission rate of the input data based on the received operation status information.
25. The storage system of claim 21, wherein the computing device is configured to receive the input data from the external device by communicating with the external device.
