Integrated circuits and computing systems, and methods for operating integrated circuits

By optimizing the frequency and voltage dynamically through the DVFS controller and environmental information in the integrated circuit, the problem of failure to consider real-time factors in integrated circuit power management is solved, thus improving performance and efficiency.

CN114879831BActive Publication Date: 2026-11-17SAMSUNG ELECTRONICS CO LTD
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202210108567.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2021-02-05
Filing Date
2022-01-28
Publication Date
2026-11-17
Estimated Expiration
2042-01-28

AI Technical Summary

Technical Problem

Existing integrated circuits fail to effectively consider real-time dynamic factors in power management, leading to increased power management complexity and impacting performance and efficiency.

Method used

By integrating a voltage and frequency scaling (DVFS) controller, the workload, frequency, and voltage of the sub-block are dynamically adjusted, and power management is optimized in conjunction with environmental information, including a clock management unit (CMU) and a power management unit (PMU) to achieve DVFS operation.

Benefits of technology

It improves the performance and power efficiency of integrated circuits, adapts to environmental changes, optimizes power management, and reduces complexity.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114879831B_ABST
    Figure CN114879831B_ABST
Patent Text Reader

Abstract

This disclosure provides an integrated circuit and a computing system, as well as a method of operating the integrated circuit. The integrated circuit and computing system perform dynamic voltage and frequency scaling (DVFS) operation. The integrated circuit includes a plurality of sub-blocks and a dynamic voltage and frequency scaling (DVFS) controller. The DVFS controller is configured to output the workload of the plurality of sub-blocks, determine a first frequency corresponding to the workload, determine a first voltage corresponding to the first frequency, and provide a second frequency from at least one frequency corresponding to the first voltage to the plurality of sub-blocks.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] Cross-references to related applications

[0002] This application is based on and claims priority to Korean Patent Application No. 10-2021-0016840, filed with the Korean Intellectual Property Office on February 5, 2021, the disclosure of which is incorporated herein by reference in its entirety. Technical Field

[0003] This disclosure relates to integrated circuits, and more specifically, to an integrated circuit and computing system that perform dynamic voltage and frequency scaling (DVFS) operations, as well as a method of operating the integrated circuit. Background Technology

[0004] An integrated circuit is a collection of electronic circuits built on semiconductor materials to perform calculations. Electronic devices such as mobile devices, computers, laptops, and navigation systems are examples of devices that use integrated circuits. Computing devices, such as mobile devices, have become more compact.

[0005] As the number of processors and Internet Protocol (IP) devices included in mobile devices increases to improve their performance, the complexity of power management in mobile devices also increases. Therefore, while the performance of electronic devices increases, so does the demand for power management.

[0006] In some cases, power supply settings are based on fixed lookup tables without considering other factors. Therefore, there is a need in the art for improved power management based on dynamic factors that change in real time. Summary of the Invention

[0007] According to this disclosure, an integrated circuit and a computing system that perform dynamic voltage and frequency scaling (DVFS) operation taking environmental factors into account, as well as a method for operating the integrated circuit, wherein the operating voltage is optimized relative to performance.

[0008] According to one aspect of this disclosure, an integrated circuit includes: a plurality of sub-blocks and a dynamic voltage and frequency scaling (DVFS) controller configured to output the workload of the plurality of sub-blocks, determine a first frequency corresponding to the workload, determine a first voltage corresponding to the first frequency, and provide a second frequency among at least one frequency corresponding to the first voltage to the plurality of sub-blocks.

[0009] According to one aspect of this disclosure, a method of operating an integrated circuit includes: calculating the workload of a plurality of sub-blocks; determining a first frequency corresponding to the workload; calculating a first voltage corresponding to the first frequency; and providing a second frequency from at least one frequency corresponding to the first voltage to the plurality of sub-blocks.

[0010] According to one aspect of this disclosure, a computing system includes: a plurality of sub-blocks; a memory storing a dynamic voltage and frequency scaling (DVFS) table indicating clock signal information and power supply voltage information based on the workload of the plurality of sub-blocks; an environment management unit configured to generate environment information for the computing system; and a DVFS controller configured to modify the DVFS table stored in the memory based on the environment information, calculate the workload of the plurality of sub-blocks, and perform DVFS operations on the plurality of sub-blocks based on the workload and the modified DVFS table.

[0011] According to one aspect of this disclosure, an integrated circuit includes: at least one sub-block configured to process instructions; a power management unit (PMU) configured to receive a power voltage control signal indicating the magnitude of a power supply voltage and to generate a power supply voltage based on the power voltage control signal; a clock management unit (CMU) configured to receive a clock control signal indicating a frequency and to generate a clock signal at the indicated frequency based on the clock control signal; and a dynamic voltage and frequency scaling (DVFS) controller configured to determine the workload of the at least one sub-block, identify a first frequency based on the workload, determine a first voltage corresponding to the first frequency, send the power voltage control signal to the PMU to set the power supply voltage to the first voltage, identify a second frequency based on the first voltage, and send the clock control signal to the CMU to set the clock control signal to the second frequency. Attached Figure Description

[0012] Embodiments of this disclosure will be more clearly understood from the following detailed description taken in conjunction with the accompanying drawings, in which:

[0013] Figure 1 This is a block diagram of a system according to an exemplary embodiment of the present disclosure;

[0014] Figure 2 This is a block diagram of a computing system according to an exemplary embodiment of the present disclosure;

[0015] Figure 3 It is a graph showing the relationship between operating frequency and operating voltage according to an exemplary embodiment of the present disclosure;

[0016] Figure 4 This is a flowchart of a method for operating an Internet Protocol (IP) device according to an exemplary embodiment of the present disclosure;

[0017] Figure 5 This is a flowchart of a method for operating an Internet Protocol (IP) device according to an exemplary embodiment of the present disclosure;

[0018] Figure 6 It is a graph showing the relationship between operating frequency and operating voltage for different systems according to exemplary embodiments of the present disclosure;

[0019] Figure 7 This is a table showing the relationship between operating frequency and operating voltage for different systems according to exemplary embodiments of this disclosure;

[0020] Figure 8 It is a graph showing the relationship between operating frequency and operating voltage for two different temperatures according to an exemplary embodiment of the present disclosure;

[0021] Figure 9 This is a table showing the relationship between operating frequency and operating voltage for different systems according to exemplary embodiments of this disclosure;

[0022] Figure 10 This is a flowchart of a method for operating an IP device according to an exemplary embodiment of the present disclosure;

[0023] Figure 11 This is a block diagram of a system according to an exemplary embodiment of the present disclosure; and

[0024] Figure 12 This is a block diagram of a communication apparatus including an application processor according to an exemplary embodiment of the present disclosure. Detailed Implementation

[0025] This disclosure relates to integrated circuits, and more specifically, to an integrated circuit and computing system that perform dynamic voltage and frequency scaling (DVFS) operations, as well as a method of operating the integrated circuit.

[0026] In some embodiments, this disclosure establishes a target voltage based on the workload and sets the highest frequency (from a set of frequencies corresponding to the target voltage) as the target frequency. As a result, high performance is achieved with the same power consumption. Additionally, DVFS can be performed taking into account environmental information of the mobile device. For example, the voltage and clock frequency based on the workload can vary based on aging, temperature, or humidity. As a result, improved performance can be based on constantly changing environmental information.

[0027] DVFS operations are performed in mobile devices and personal computers. Therefore, the Clock Management Unit (CMU) changes the clock frequency, and the Power Management Unit (PMU) changes the voltage. A target voltage is set to a single voltage value, where multiple frequencies correspond to the target voltage. As a result, the target voltage and target frequency are established based on a fixed DVFS value, regardless of the multiple frequencies. In some cases, the application processor of a mobile device can perform power management by controlling the voltage via DVFS operations that control the frequency and voltage in the application processor.

[0028] This disclosure discloses an integrated circuit and a computing system for performing Dynamic Voltage and Frequency Scaling (DVFS) operation, as well as a method for operating the integrated circuit. The integrated circuit includes multiple sub-blocks and a Dynamic Voltage and Frequency Scaling (DVFS) controller. The DVFS controller is configured to output the workload of the multiple sub-blocks, determine a first frequency corresponding to the workload, determine a first voltage corresponding to the first frequency, and provide a second frequency from at least one frequency corresponding to the first voltage to the multiple sub-blocks.

[0029] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings.

[0030] Figure 1 This is a block diagram of system 10 according to an exemplary embodiment of the present disclosure. System 10 may include Internet Protocol (IP) device 100, clock management unit (CMU) 210, power management unit (PMU) 220, environment management unit (EMU) 230, and memory 300.

[0031] System 10 can correspond to various types of data processing devices, and can correspond to, for example, mobile devices. Additionally or alternatively, system 10 can correspond to laptop computers, mobile phones, smartphones, tablet computers (PCs), personal digital assistants (PDAs), enterprise digital assistants (EDAs), digital still cameras, digital video cameras, portable multimedia players (PMPs), personal navigation devices or portable navigation devices, handheld game consoles, mobile internet devices (MIDs), wearable computers, Internet of Things (IoT) devices, Internet of Everything (IoE) devices, or e-books.

[0032] System 10 may include various types of memory devices. For example, memory 300 may correspond to various types of semiconductor memory devices, and depending on the implementation, may include dynamic random access memory (DRAM) such as double data rate (DDR) SDRAM, low power double data rate (LPDDR) SDRAM, graphics double data rate (GDDR) SDRAM, RAMBUS dynamic random access memory (RDRAM), etc. Furthermore, memory 300 may also include any one of flash memory, phase-change RAM (PRAM), magnetoresistive RAM (MRAM), resistive RAM (ReRAM), and ferroelectric RAM (FeRAM).

[0033] In some implementations, system 10 may be implemented as a system-on-a-chip (SoC). The SoC may include a system bus that applies a protocol with a standard bus specification, and may include various IP devices connected to the system bus. The standard specification for the system bus may include the Advanced Microcontroller Bus Architecture (AMBA) protocol from advanced RISC machines (ARM). Bus types for the AMBA protocol may include Advanced High Performance Bus (AHB), Advanced Peripheral Bus (APB), Advanced Extensible Interface (AXI), AXI4, AXI Coherence Extensions (ACE), etc. Alternatively or additionally, other types of protocols may be applied, such as SONICs' u-network, IBM's CoreConnect, the Open Core Protocol of OCP-IP, etc.

[0034] IP device 100 may be a mobile phone, etc., and may include multiple sub-blocks 110 and a Dynamic Voltage and Frequency Scaling (DVFS) controller 120. The sub-blocks may be configured to process instructions.

[0035] IP device 100 can be designed as an integrated circuit implemented with multiple transistors. For example, IP device 100 may include a central processing unit (CPU), a graphics processing unit (GPU), a neural processing unit (NPU), or a graphics signal processor (ISP). Figure 1 Only one IP device (e.g., IP device 100) is shown in the diagram, but the type and number of IP devices included in system 10 are not limited thereto. In some embodiments, the DVFS controller 120 may be external to IP device 100.

[0036] Multiple sub-blocks 110 can process instructions independently of each other. Each sub-block may include a CPU core, GPU core, NPU core, or ISP core. Because the IP device 100 includes multiple cores, the system 10 may be referred to as a multi-core processor. Sub-blocks may also be referred to as sub-functional blocks.

[0037] Each of the multiple sub-blocks 110 can process instructions based on a clock signal CLK and a power supply voltage VDD. The performance of each sub-block depends on the clock signal CLK and the power supply voltage VDD. As the magnitude of the power supply voltage VDD increases and the frequency of the clock signal CLK increases, the performance of the IP device 100 can be improved, but the power consumption of the IP device 100 can also increase. For ease of description, in this specification, the frequency of the clock signal CLK can be referred to as the operating frequency, and the magnitude of the power supply voltage VDD can be referred to as the operating voltage.

[0038] The DVFS controller 120 can determine the operating state of various functional blocks in the system 10, and based on the determination result, send a clock control signal CTRL_CLK and a power supply voltage control signal CTRL_VDD to the CMU 210 and / or PMU 220 to control the operating frequency and / or operating voltage of each functional block. For example, the DVFS controller 120 can control the operating frequency and operating voltage supplied to multiple sub-blocks 110.

[0039] For example, the DVFS controller 120 may send a clock control signal CTRL_CLK and a power supply voltage control signal CTRL_VDD to control the operating conditions of the IP device 100 in each predetermined sampling period. For example, the DVFS controller 120 may send a clock control signal CTRL_CLK to control the frequency of the clock signal CLK, and a power supply voltage control signal CTRL_VDD to control the level of the power supply voltage VDD.

[0040] In this embodiment, the DVFS controller 120 can calculate the workload of the operations performed by the system 10 and / or the IP device 100, and can perform DVFS operations in response to the obtained workload by referring to a DVFS table stored in the memory 300. The DVFS controller 120 can obtain from the DVFS table a target voltage and / or a target frequency for maximizing the performance of the IP device 100 under the corresponding workload. In this embodiment, the DVFS controller 120 can obtain a target frequency corresponding to the workload and a target voltage corresponding to the target frequency.

[0041] The DVFS controller 120 determines the workload of sub-block 110, identifies a first frequency f1 based on the workload, determines a first voltage V1 corresponding to the first frequency f1, sends a power supply voltage control signal CTRL_VDD to PMU 220 to set the power supply voltage VDD to the first voltage V1, identifies a second frequency f2 based on the first voltage V1, and sends a clock control signal CTRL_CLK to CMU 210 to set the clock control signal CTRL_CLK to the second frequency f2. The DVFS controller 120 is referenced... Figure 1 , Figure 2 and Figure 3 Examples of the corresponding elements described or including aspects thereof.

[0042] Furthermore, one or more target frequencies may correspond to a target voltage value. When IP device 100 has a fixed operating voltage value, the performance or power efficiency of IP device 100 increases with the increase of operating frequency. Therefore, when the operating voltage of IP device 100 is determined to be a first voltage value, determining the maximum frequency among one or more frequencies corresponding to the first voltage value as the operating frequency can improve the performance of IP device 100.

[0043] Therefore, in this embodiment, the DVFS controller 120 can calculate the target voltage corresponding to the operating load through DVFS operation, and set the maximum frequency among one or more frequencies corresponding to the target voltage as the target frequency. Thus, while maximizing power efficiency, the performance of the IP device 100 can be improved.

[0044] Additionally or alternatively, environmental factors of the IP device 100 (e.g., temperature, humidity, etc.) may vary over time. Environmental factors can affect the operation of the IP device 100 and DVFS operation. For example, under the same workload, the target voltage and / or target frequency determined by DVFS operation may vary depending on the temperature or aging of the IP device 100. Therefore, in this embodiment, the DVFS controller 120 may take environmental factors into account when performing DVFS operation. This will be referred to below. Figures 5 to 9 Detailed description. The values ​​of each environmental factor below can be referred to as environmental information.

[0045] In this embodiment, DVFS controller 120 may refer to hardware capable of performing DVFS functions and DVFS operations, or it may refer to computer program code capable of performing DVFS functions and DVFS operations. However, DVFS controller 120 is not limited to these, but may refer to an electronic recording medium (e.g., a processor) having computer program code capable of performing DVFS functions and DVFS operations. In other words, DVFS controller 120 may refer to a combination of functions and / or a combination of structures for implementing the hardware and / or driving the hardware of this disclosure.

[0046] The CMU 210 can generate a clock signal CLK and control the frequency of the clock signal CLK based on the clock control signal CTRL_CLK. For example, the CMU 210 may include an oscillator for generating the clock signal CLK based on the clock control signal CTRL_CLK. The CMU 210 may also be referred to as a clock generator or clock generation circuit.

[0047] Alternatively, the CMU 210 may receive a clock control signal CTRL_CLK and generate a clock signal CLK at an indicated frequency based on the clock control signal CTRL_CLK. The clock control signal CTRL_CLK may indicate a frequency.

[0048] PMU 220 can generate a power supply voltage VDD and can control the magnitude of the power supply voltage VDD based on a power supply voltage control signal CTRL_VDD. In an embodiment, PMU 220 may include a switching regulator for generating the power supply voltage VDD based on the power supply voltage control signal CTRL_VDD, and may include a power management integrated circuit (PMIC).

[0049] Alternatively, the PMU 220 may receive a power supply voltage control signal CTRL_VDD and generate a power supply voltage VDD based on the power supply voltage control signal CTRL_VDD. The power supply voltage control signal CTRL_VDD indicates the magnitude of the power supply voltage VDD.

[0050] EMU 230 can collect environmental information INFO_E from system 10 and provide the environmental information INFO_E to IP device 100. In an embodiment, EMU 230 may include multiple thermal sensors and use the temperatures sensed by each thermal sensor to generate temperature information.

[0051] In this implementation, EMU 230 can inspect the aging of system 10 and generate aging information for system 10. EMU 230 can measure the degree of degradation of multiple semiconductor devices included in system 10 and generate aging information indicating whether the aging degree of system 10 is high or low. For example, aging information can be generated based on threshold voltage changes, timing changes, operating voltages, operating temperatures, etc. of multiple semiconductor devices.

[0052] EMU 230 can generate environmental information INFO_E, which includes information about the performance or characteristics of the semiconductor devices in system 10. Alternatively, environmental information INFO_E can be generated by reflecting the real-time changes in the performance or characteristics of the semiconductor devices. In other words, environmental information INFO_E can include static environmental information and dynamic environmental information, as referenced... Figure 5 Detailed description.

[0053] Memory 300 is accessible by IP device 100, and IP device 100 can store data in memory 300 or read data stored in memory 300. Memory 300 may include volatile memory devices such as SRAM, DRAM, etc., or may include non-volatile memory devices such as flash memory, resistive random access memory (RRAM), etc. In some embodiments, memory 300 may store a DVFS table. Figure 1 In this embodiment, the memory 300 may be located outside the IP device 100. However, this is not the only example; the memory 300 may be included within the IP device 100.

[0054] System 10 is not limited to Figure 1 The components shown may also include various types of function blocks for communicating with the IP device 100, such as input / output (IO) interface blocks, universal serial bus (USB) master blocks, USB slave blocks, etc.

[0055] Figure 2 This is a block diagram illustrating a computing system 20 according to an exemplary embodiment of the present disclosure.

[0056] Reference Figure 2 The computing system 20 may include a DVFS controller 400, a memory 500, a CMU 610, a PMU 620, and an EMU 700. The DVFS controller 400, memory 500, CMU 610, PMU 620, and EMU 700 may respectively correspond to... Figure 1 The components include the DVFS controller 120, memory 300, CMU 210, PMU 220, and EMU 230. Therefore, redundant descriptions will be omitted.

[0057] The DVFS controller 400 may include a DVFS regulator module (or DVFS regulator) 410, a CMU driver 420, and a PMU driver 430. The DVFS regulator module 410 controls the DVFS operation of the computing system 20. In one embodiment, the DVFS regulator module 410 calculates the workload of the computing system 20 and obtains a DVFS table 510 from the memory 500 to determine a target voltage and / or target frequency corresponding to the workload. In another embodiment, the DVFS regulator module 410 may modify the DVFS table 510 stored in the memory 500 based on information INFO1 and INFO2 collected from the EMU 700, and perform DVFS operations. In this embodiment, the order of setting the target frequency, setting the target voltage corresponding to the target frequency, and setting the target frequency again by the DVFS regulator module 410 is described, but the order of DVFS operations is not limited to this.

[0058] The DVFS controller module 410 can control the CMU driver 420 and the PMU driver 430 based on a determined target voltage and / or a determined target frequency. Hereinafter, the term "module" may refer to hardware capable of performing functions and operations according to the names of the respective modules, or it may refer to computer program code capable of performing functions and operations. However, "module" is not limited to these, but may refer to an electronic recording medium (e.g., a processor) carrying computer program code capable of performing functions and operations. In other words, a module may refer to a combination of functions and / or a combination of structures for executing the hardware of this disclosure and / or for driving the hardware.

[0059] The CMU driver 420 can generate a clock control signal CTRL_CLK under the control of the DVFS controller module 410, and provide the clock control signal CTRL_CLK to the CMU 610.

[0060] The PMU driver 430 can generate a power supply voltage control signal CTRL_VDD under the control of the DVFS controller module 410, and provide the power supply voltage control signal CTRL_VDD to the PMU 620.

[0061] The memory 500 may include a DVFS table 510. The DVFS table 510 may include operating voltages and operating frequencies corresponding to various workloads. The DVFS table 510 may include values ​​that can be hard-coded in the memory 500, or it may include values ​​that can be modified by software coding. The DVFS table 510 may be modified by the DVFS controller module 410.

[0062] exist Figure 2 In this system, a single DVFS table 510 is provided, but multiple DVFS tables 510 can also be provided. For example, multiple DVFS tables 510 can be generated depending on changes in the environment of the computing system 20. In an implementation, the DVFS controller module 410 can select any one of the multiple DVFS tables 510 that can be pre-generated according to the runtime environment and perform DVFS operations.

[0063] The CMU 610 can generate a clock signal CLK based on the clock control signal CTRL_CLK and provide the clock signal CLK to the components within the computing system 20.

[0064] The PMU 620 can generate a power supply voltage VDD based on the power supply voltage control signal CTRL_VDD and provide the power supply voltage VDD to the components within the computing system 20.

[0065] The EMU 700 may include a Temperature Monitoring Unit (TMU) 710 and a Hardware Performance Monitoring Unit (HPM) 720, and may also include various modules and / or sensors capable of measuring the environment of the computing system 20. The EMU 700 may generate environmental information INFO_E based on the information generated by the TMU 710 and HPM 720, and provide the environmental information INFO_E to the DVFS controller 400. In other words, the environmental information INFO_E may include temperature information INFO1 and / or performance information INFO2.

[0066] The TMU 710 may include multiple thermal sensors and can use the temperatures measured by the multiple thermal sensors to generate temperature information INFO1 and provide the temperature information INFO1 to the DVFS controller 400. The multiple thermal sensors may be arranged within the computing system 20 and also within the DVFS controller 400. Each of the multiple thermal sensors can sense the ambient temperature of a specified component.

[0067] The HPM 720 measures the performance of components within the computing system 20 under operating conditions. Alternatively, the HPM 720 can measure component performance regardless of the operating environment. The HPM 720 generates performance information INFO2 indicating the performance of multiple semiconductor devices (e.g., semiconductor device speed, operating voltage, threshold voltage, etc.). Figure 2 In this context, the HPM720 is located outside the DVFS controller 400, but the placement of the HPM720 is not limited to any particular location.

[0068] In addition, the DVFS controller module 410 can perform operations such as calculating workload, collecting information from EMU 700, and modifying DVFS table 510 according to a preset cycle.

[0069] Figure 3 This is a curve showing the relationship between operating frequency and operating voltage according to an exemplary embodiment of the present disclosure. Figure 11 .

[0070] Reference Figure 3 ,curve Figure 11 DVFS tables can be drawn (e.g., Figure 2 (of 510). When in Figure 1 System 10 or Figure 2 When performing DVFS operation in computing system 20, the operating frequency of system 10 or computing system 20 and the corresponding operating voltage can be determined.

[0071] The operating frequency that exhibits optimal performance under the workload can be determined. The operating voltage can be determined corresponding to the determined operating frequency. In this case, when CMUs 210 and 610 can set their operating frequencies at intervals smaller than those set by PMUs 220 and 620, a single operating voltage can correspond to multiple operating frequencies. In an embodiment, a single first voltage V1 can correspond to multiple frequencies corresponding to a first frequency range fr1.

[0072] In this scenario, at the same voltage, performance and power efficiency improve with increasing clock frequency. Therefore, the second frequency f2 can be the highest frequency among those corresponding to the first voltage V1 and can be determined as the operating frequency. For example, this can be achieved when the first frequency f1 is set as the target frequency and when the second frequency f2 is set as the target frequency, and when the first voltage is also set as the target voltage. Therefore, when the first voltage V1 is determined as the target voltage, performance can be improved compared to voltage by setting the second frequency f2 instead of the first frequency as the target frequency.

[0073] Furthermore, when the first voltage V1 is set to the target voltage according to the workload, the multiple frequencies corresponding to the first voltage V1 can vary according to environmental factors of system 10 or computing system 20. In an embodiment, the second frequency f2 can be the highest among the frequencies corresponding to the first voltage V1, and can increase with the increase of the temperature of system 10. For example, the second frequency f2 can decrease with the increase of the aging degree of system 10 and the increase of the usage time of system 10.

[0074] Figure 4 This is a flowchart illustrating an operation method of an Internet Protocol (IP) device according to an exemplary embodiment of this disclosure. (See also...) Figure 2 and Figure 4 The workload calculation operation can be performed in operation S110. In this embodiment, the workload of computing system 20 can be obtained. In this embodiment, the DVFS controller module 410 can calculate the workload of components (e.g., multiple sub-blocks) within system 10 or computing system 20.

[0075] The operation of determining a first frequency corresponding to the obtained workload can be performed in operation S120. In an embodiment, the DVFS controller module 410 can refer to the DVFS table 510 stored in the memory 500 to determine the first frequency corresponding to the workload.

[0076] In operation S130, the operation of setting a first voltage corresponding to a first frequency as a target voltage can be performed. In an embodiment, the DVFS regulator module 410 can refer to the DVFS table 510 and set the first voltage corresponding to the first frequency as the target voltage of the components in the calculation system 20. For example, the DVFS regulator module 410 can control the PMU driver 430 to generate a power supply voltage control signal CTRL_VDD based on the first voltage and provide the power supply voltage control signal CTRL_VDD to the PMU 620.

[0077] In operation S140, the operation of setting a second frequency among the frequencies corresponding to the first voltage as the target frequency can be performed. In this embodiment, the DVFS controller module 410 can set the second frequency by referring to the DVFS table 510. The second frequency, which serves as the target frequency for the components within the calculation system 20, is the maximum value among a plurality of frequencies corresponding to the first voltage.

[0078] The second frequency can be provided to multiple sub-blocks in operation S150. Figure 1 The operation of (110). In an embodiment, the DVFS controller module 410 can control the CMU driver 420 to generate a clock control signal CTRL_CLK based on a second frequency and provide the clock control signal CTRL_CLK to the CMU 610.

[0079] Figure 5 This is a flowchart of an operation method of an IP device according to an exemplary embodiment of the present disclosure.

[0080] Refer to together Figure 2 and Figure 5 The workload calculation operation can be performed in operation S210. In this embodiment, the DVFS controller module 410 can calculate multiple sub-blocks (e.g., Figure 1 The workload of components within the computing system 20 (110) or the computing system 20.

[0081] The operation of collecting environmental information can be performed in operation S220. In this embodiment, the DVFS controller module 410 can collect environmental information INFO_E of the computing system 20 from the EMU 700. The environmental information INFO_E may include dynamic environmental information and static environmental information. For example, dynamic environmental information may include at least one of the following: utilization rate of multiple sub-blocks 110, size of data to be processed, data location, and operation frequency. Static environmental information may include at least one of the following: number of sub-blocks 110, operating capacity, size of memory 500, and bandwidth of memory 500.

[0082] The INFO_E operation for collecting environmental information can be performed according to a preset cycle, and can also be performed during DVFS operations.

[0083] The operation of modifying the DVFS table 510 based on environmental information INFO_E can be performed in operation S230. In this embodiment, the DVFS controller module 410 can modify the DVFS table 510 stored in the memory 500 based on the collected environmental information INFO_E. The operation of modifying the DVFS table 510 can be performed when collecting environmental information INFO_E, can be performed according to a preset period, or can be performed when performing DVFS operations.

[0084] In this implementation, the memory 500 may include multiple different DVFS tables based on the environmental factors included in the environmental information INFO_E. The DVFS controller module 410 may select the DVFS table 510 corresponding to the collected environmental information INFO_E from the multiple DVFS tables 510.

[0085] In this implementation, the memory may include operating voltages and operating frequencies for various workloads, which enable DVFS operation based on environmental factors. Therefore, the DVFS controller module 410 can modify the DVFS table 510 stored in the memory 500 based on the operating voltage and operating frequency corresponding to the collected environmental information INFO_E. For example, the DVFS table 510 can be modified based on temperature, as shown in reference... Figure 8 and Figure 9Detailed description.

[0086] In operation S240, DVFS operations can be performed based on the modified DVFS table. In this embodiment, the DVFS controller module 410 can calculate the workload and set the operating frequency and operating voltage corresponding to the obtained workload as the target frequency and target voltage of the calculation system 20 based on the modified DVFS table or a DVFS table selected from multiple DVFS tables.

[0087] Figure 6 These are curves illustrating the relationship between operating frequency and operating voltage for different systems according to exemplary embodiments of this disclosure. Figure 12 ,and Figure 7 Table 13 shows the relationship between target frequency and target voltage for different systems according to exemplary embodiments of this disclosure.

[0088] Refer to together Figure 6 and Figure 7 DVFS table 510 is based on the computing system (e.g., Figure 2 The target voltage and target frequency can be changed (20%).

[0089] For example, using the reference curve Figure 12 Table 13 describes a scenario where the workload of the first system SYSTEM1 and the second system SYSTEM2 is calculated, and as a result, the target frequency of the first system SYSTEM1 and the second system SYSTEM2 is determined to be 400MHz.

[0090] In the first system SYSTEM1, the operating voltage corresponding to the target frequency of 400MHz can be 700mV, and the frequency corresponding to 700mV can be a frequency within the second frequency range fr2. Therefore, according to this embodiment, the DVFS controller 400 of the first system SYSTEM1 can set any frequency (e.g., 430MHz) between 400MHz and 430MHz as the target frequency.

[0091] Furthermore, in the second system SYSTEM2, the operating voltage corresponding to the target frequency of 400MHz can be 687.5mV, and the frequency corresponding to 687.5mV can be a frequency within the third frequency range fr3. Therefore, according to this embodiment, the DVFS controller 400 of the second system SYSTEM2 can set any frequency (e.g., 410MHz) between 400MHz and 410MHz as the target frequency.

[0092] Furthermore, even within a single system (e.g., the first system SYSTEM1), the DVFS table varies depending on environmental factors; therefore, the target voltage and target frequency can change. This will be referenced... Figure 8 and Figure 9 describe.

[0093] Figure 8 Graph 14 shows the relationship between operating frequency and operating voltage at different temperatures according to an exemplary embodiment of the present disclosure. Figure 9 Table 15 shows the relationship between operating frequency and operating voltage for different temperatures according to an exemplary embodiment of this disclosure.

[0094] Refer to together Figure 2 and Figure 8 The temperature of the first system 10 can vary depending on the operating environment. As a result, DVFS operations can be performed using different DVFS tables depending on the temperature of the operating environment where the workload is obtained.

[0095] For example, the DVFS table itself used during DVFS operations can be modified based on the temperature. Additionally or alternatively, information included in the DVFS table can be modified. In an implementation, at a first temperature T1, the DVFS table can be modified based on information in the columns of Table 15 corresponding to the first temperature T1, and at a second temperature T2, the DVFS table can be modified based on information in the columns of Table 15 corresponding to the second temperature T2.

[0096] In one implementation, the memory 300 may include a DVFS table corresponding to a first temperature T1 and a DVFS table corresponding to a second temperature T2. In this case, the DVFS controller 400 may select either the DVFS table corresponding to the first temperature T1 or the DVFS table corresponding to the second temperature T2 based on the ambient temperature of the operating environment, and perform DVFS operations based on the selected DVFS table.

[0097] For example, the case of obtaining the workload of the first system 10 and determining the target frequency corresponding to the workload as 400MHz is described with reference to graph 14 and table 15.

[0098] When the temperature of the first system 10 when it receives a workload is a first temperature T1, the operating voltage corresponding to the target frequency of 400MHz can be 700mV, and the frequency corresponding to 700mV can be a frequency within the second frequency range fr2. Therefore, according to this embodiment, the DVFS controller 400 of the first system 10 can set any frequency (e.g., 430MHz) between 400MHz and 430MHz as the target frequency.

[0099] Furthermore, when the temperature of the first system 10 at the time of obtaining the workload is the second temperature T2, the operating voltage corresponding to the target frequency of 400MHz can be 700mV, and the frequency corresponding to 700mV can be a frequency within the fourth frequency range fr4. Therefore, according to this embodiment, the DVFS controller 400 of the first system 10 can set any frequency (e.g., 450MHz) between 400MHz and 450MHz as the target frequency. Alternatively, multiple DVFS tables based on temperature can be stored in the memory 500 before performing the above operations.

[0100] Figure 10 This is a flowchart illustrating an operation method of an IP device according to an exemplary embodiment of the present disclosure. (See also...) Figure 2 and Figure 10 The operation of calculating the workload can be performed in operation S310. The operation of collecting environmental information INFO_E can be performed in operation S320. The operation of modifying DVFS table 510 based on environmental information INFO_E can be performed in operation S330. In an embodiment, the operation of selecting a single appropriate DVFS table 510 from multiple DVFS tables stored in memory 500 based on environmental information INFO_E can be performed instead of operation S330.

[0101] In operation S340, the operation of determining a first frequency corresponding to the obtained workload based on the modified DVFS table 510 can be performed. In this embodiment, the DVFS controller module 410 may refer to the modified DVFS table 510 stored in the memory 500 and determine the first frequency corresponding to the workload. The operation of determining the first frequency corresponding to the obtained workload based on a DVFS table selected from multiple DVFS tables according to the environmental information INFO_E can be performed.

[0102] In operation S350, the operation of setting the first voltage corresponding to the first frequency as the target voltage can be performed. In operation S360, the operation of setting the second frequency among the frequencies corresponding to the first voltage as the target frequency can be performed. Thereafter, the operation of providing the second frequency to multiple sub-blocks (e.g., Figure 1 The operation of 110).

[0103] Figure 11 This is a block diagram illustrating a system 30 according to an exemplary embodiment of the present disclosure. System 30 may be implemented as a mobile phone, smartphone, portable computer, personal digital assistant (PDA), enterprise digital assistant (EDA), digital still camera, digital video camera, portable multimedia player (PMP), personal navigation device or portable navigation device, handheld game console or handheld device (such as an e-book).

[0104] System 30 may include SoC 3100 and memory device 3200. SoC 3100 may include CPU 3110, GPU 3120, NPU 3130, ISP 3140, memory interface (MIF) 3150, clock management unit (CMU) 3160, and power management unit (PMU) 3170. CPU 3110, GPU 3120, NPU 3130, and ISP 3140 may be referred to as main IP devices, and MIF 3150 may be referred to as secondary IP devices. At least one of CPU 3110, GPU 3120, NPU 3130, ISP 3140, and MIF 3150 may be as described above. Figures 1 to 10 The described IP device implementation. Therefore, at least one of CPU 3110, GPU 3120, NPU 3130, ISP 3140, and MIF 3150 may include multiple sub-blocks and may include a DVFS controller for performing DVFS operations based on a DVFS table stored in memory device 3200 and collected environmental information. The DVFS controller included in CPU 3110, GPU 3120, NPU 3130, ISP 3140, and MIF 3150 can control CMU 3160 or PMU 3170, and CPU 3110, GPU 3120, NPU 3130, ISP 3140, and MIF 3150 can process one or more instructions by receiving a clock signal CLK from CMU 3160 and a power supply voltage from PMU 3170.

[0105] CPU 3110 can process or execute commands and / or data stored in memory device 3200 in response to clock signal CLK generated by CMU 3160.

[0106] GPU 3120 can obtain image data stored in memory device 3200 in response to clock signal CLK generated by CMU 3160. Based on the image data provided from MIF 3150, GPU 3120 can generate data to be output through display device (not shown) and can encode the image data.

[0107] NPU 3130 can refer to any device used to execute machine learning models. NPU 3130 can be a hardware block designed to execute machine learning models. Machine learning models can include models based on artificial neural networks, decision trees, support vector machines, regression analysis, Bayesian networks, genetic algorithms, etc. As a non-limiting example, artificial neural networks can include convolutional neural networks (CNNs), region-with-convolutional-neural networks (R-CNNs), region proposal networks (RPNs), recurrent neural networks (RNNs), stacked deep neural networks (S-DNNs), state-space dynamic neural networks (S-SDNNs), deconvolutional networks, deep belief networks (DBNs), restricted Boltzmann machines (RBMs), fully convolutional networks, long short-term memory (LSTM) networks, and classification networks.

[0108] The ISP 3140 can perform signal processing operations on raw data received from an image sensor (not shown) external to the SoC 3100 and generate digital data with improved image quality.

[0109] The MIF 3150 provides an interface for an external memory device 3200 to the SoC 3100. The memory device 3200 may include dynamic random access memory (DRAM), phase-change random access memory (PRAM), resistive random access memory (ReRAM), or flash memory. The memory device 3200 may include multiple DVFS tables depending on environmental factors. Each of the DVFS tables can be modified by a DVFS controller.

[0110] The CMU 3160 generates a clock signal and provides it to the components of the SoC 3100. The CMU 3160 may include clock generation devices such as a phase-locked loop (PLL), a delay-locked loop (DLL), a crystal, etc. The PMU 3170 converts external power to internal power and supplies the internal power to the components of the SoC 3100.

[0111] Figure 12 This is a block diagram illustrating a communication device 40 including an application processor 4010 according to an exemplary embodiment of the present disclosure.

[0112] Reference Figure 12 The communication device 40 may include an application processor 4010, a memory device 4020, a display 4030, an input device 4040, and a radio transceiver 4050. The application processor 4010 may be as described above. Figures 1 to 11 An implementation of at least one IP device described.

[0113] Radio transceiver 4050 can transmit or receive radio signals via antenna 4060. For example, radio transceiver 4050 can convert radio signals received via antenna 4060 into signals that can be processed by application processor 4010.

[0114] Therefore, the application processor 4010 can process the signal output from the radio transceiver 4050 and send the processed signal to the display 4030. Alternatively, the radio transceiver 4050 can convert the signal output from the application processor 4010 into a radio signal and send the radio signal to an external device via the antenna 4060.

[0115] The input device 4040 is a device through which control signals for operation of the application processor 4010 or data to be processed by the application processor 4010 can be input, and can be implemented as a pointing device such as a touchpad and computer mouse, a small keyboard or keyboard.

[0116] According to an implementation, the application processor 4010 may include a plurality of sub-blocks and a DVFS controller for controlling the plurality of sub-blocks. (Refer to the above...) Figures 1 to 11 As described, the DVFS controller can perform DVFS operations based on environmental information of the components within the communication device.

[0117] although Figure 12 Although not shown, the communication device 40 may also include a CMU for providing clock signals to various components disposed in the communication device 40 and a PMU for providing power supply voltages to various components disposed in the communication device 40. The CMU may output a clock signal with an regulated frequency under the control of the DVFS controller, and the PMU may output a power supply voltage with an regulated magnitude under the control of the DVFS controller.

[0118] Although this disclosure has been specifically shown and described with reference to its embodiments, it will be understood that various changes in form and detail may be made therein without departing from the spirit and scope of the appended claims.

Claims

1. An integrated circuit, comprising: Multiple sub-blocks; as well as A dynamic voltage and frequency scaling controller is configured to output the workload of the plurality of sub-blocks, determine a first frequency corresponding to the workload, determine a first voltage corresponding to the first frequency, and provide the plurality of sub-blocks with a second frequency from a plurality of frequencies corresponding to the first voltage. The second frequency is the highest frequency among the plurality of frequencies corresponding to the first voltage.

2. The integrated circuit according to claim 1, wherein, The integrated circuit is connected to the memory. The dynamic voltage and frequency scaling controller is further configured to determine the first frequency based on a dynamic voltage and frequency scaling table stored in the memory, and to calculate the first voltage. The dynamic voltage and frequency scaling table includes frequency information corresponding to the workload and voltage information corresponding to the frequency information.

3. The integrated circuit according to claim 2, wherein, The dynamic voltage and frequency scaling controller is also configured to collect environmental information of the integrated circuit from the environmental monitoring unit and modify the dynamic voltage and frequency scaling table based on the environmental information.

4. The integrated circuit according to claim 3, wherein, The environmental information includes at least one of the temperature information of the integrated circuit and the aging information of the integrated circuit.

5. The integrated circuit according to claim 3, wherein, The dynamic voltage and frequency scaling controller is also configured to determine a third frequency corresponding to the workload based on a modified dynamic voltage and frequency scaling table, calculate a second voltage corresponding to the third frequency, and provide a fourth frequency among at least one frequency corresponding to the second voltage to the plurality of sub-blocks.

6. The integrated circuit according to claim 4, wherein, The second frequency is the frequency of the clock signal applied to the plurality of sub-blocks, and The first voltage has the magnitude of the power supply voltage applied to the plurality of sub-blocks.

7. The integrated circuit according to claim 1, wherein, Each of the plurality of sub-blocks includes any one of a central processing unit core, a graphics processing unit core, a neural processing unit core, and a graphics signal processor core.

8. A method of operating an integrated circuit, the method comprising the steps of: Calculate the workload of multiple sub-blocks; Determine the first frequency corresponding to the workload; Calculate the first voltage corresponding to the first frequency; as well as Provide the second frequency from a plurality of frequencies corresponding to the first voltage to the plurality of sub-blocks. The step of providing the second frequency to the plurality of sub-blocks includes: determining the maximum frequency among the plurality of frequencies corresponding to the first voltage as the second frequency.

9. The method according to claim 8, wherein, The step of determining the first frequency corresponding to the workload includes: Environmental information of the multiple sub-blocks is collected from the environmental monitoring unit; Based on the environmental information, the dynamic voltage and frequency scaling table stored in the memory is modified to a first dynamic voltage and frequency scaling table; and The first frequency is determined based on the first dynamic voltage and frequency scaling table.

10. The method according to claim 8, wherein, The magnitude of the second frequency increases with the increase of the temperature of the integrated circuit.

11. The method according to claim 8, wherein, The magnitude of the second frequency decreases as the aging degree of the integrated circuit increases.

12. The method according to claim 8, wherein, The step of providing the second frequency to the plurality of sub-blocks includes: providing the second frequency as the frequency of a clock signal applied to the plurality of sub-blocks.

13. A computing system, comprising: Multiple sub-blocks; The memory stores a dynamic voltage and frequency scaling table indicating clock signal information and power supply voltage information based on the workload of the plurality of sub-blocks; An environment management unit is configured to generate environmental information for the computing system; as well as A dynamic voltage and frequency scaling controller is configured to modify the dynamic voltage and frequency scaling table stored in the memory based on the environmental information, calculate the workload of the plurality of sub-blocks, and perform dynamic voltage and frequency scaling operations on the plurality of sub-blocks based on the workload of the plurality of sub-blocks and the modified dynamic voltage and frequency scaling table. The dynamic voltage and frequency scaling controller is further configured to determine a first frequency corresponding to the workload, calculate a first voltage corresponding to the first frequency, and provide a second frequency to the plurality of sub-blocks as the maximum frequency among the plurality of frequencies corresponding to the first voltage.

14. The computing system according to claim 13, wherein, The dynamic voltage and frequency scaling controller is also configured to provide the plurality of sub-blocks with any one of at least one frequency corresponding to the workload, based on a modified dynamic voltage and frequency scaling table.

15. The computing system according to claim 13, wherein, The second frequency includes the frequency of the clock signal applied to the plurality of sub-blocks.

16. The computing system according to claim 13, wherein, The environmental information includes static environmental information and dynamic environmental information determined based on the running time of the multiple sub-blocks.

17. The computing system according to claim 16, wherein, The dynamic environmental information includes at least one of the temperature information of the computing system and the aging information of the computing system.

Citation Information

Patent Citations

  • Apparatus for damping vibration of strip

    KR1020210016840A

  • Reliability and performance of a system-on-a-chip by predictive wear-out based activation of functional components

    US20110173432A1

  • Dynamic frequency scaling in multi-processor systems

    US20160327999A1

  • Multiprocessor systems having processors with different processing capabilities connecting to a clock generator

    US20170269629A1