Context-aware power grid
By designing a configurable charge storage network and control system in integrated circuits, dynamically adjusting the configuration and charging parameters of the charge storage device, the problem that existing capacitors cannot adapt to changes in circuit activity is solved, and more efficient charge storage and lower Power consumption.
Patent Information
- Application Number
- CN201980016977.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2018-03-06
- Filing Date
- 2019-03-04
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2039-03-04
AI Technical Summary
The size of existing capacitors is usually adjusted to handle the worst-case activity level of the integrated circuit, resulting in increased area, power consumption, and charging time, and inability to effectively adapt to the situational changes in circuit operation.
A configurable charge storage network and control system is designed to dynamically adjust the configuration and charging parameters of the charge storage device to meet different power needs through a scenario-based controller monitoring circuit activity level and power supply voltage.
It realizes that while ensuring the normal operation of the circuit, reduce the size and power consumption of the capacitor, improve charge storage efficiency, adapt to different power needs, and reduce charging time.
Smart Images

Figure CN111801868B_ABST
Abstract
Description
Background Art
[0001] Charge storage devices such as capacitors can be used to power an entire integrated circuit chip, or a portion of a circuit on a chip. The size (capacitance) of a power capacitor determines the amount of charge it can hold, which in turn determines the amount of power it can provide to a circuit. Summary of the invention
[0002] A configurable charge storage network and control system are described that adjust the charge storage device configuration of the charge storage network based on the needs of an integrated circuit. The adjustment can be for current needs or expected needs of the integrated circuit.
[0003] A system is described that includes a circuit (e.g., an integrated circuit); a power network coupled to the circuit to provide a core voltage to the circuit; and a scenario-based controller that monitors a power supply voltage level of the power supply, monitors a core voltage level of the core voltage, and monitors activity of the circuit to derive an activity level of the circuit. The scenario-based controller can adjust a capacitance of the power network or a charging parameter associated with the power network based on the activity level of the circuit to correspond to a power demand associated with the activity level.
[0004] A context-aware power network may include at least one charge storage unit; and a context-based controller including at least one input receiving a context of circuit operation; and at least one control signal output coupled to the at least one charge storage unit, wherein the context-based controller determines an activity level of the circuit based on the context of the circuit operation received via the at least one input, and generates at least one control signal to be output via the at least one control signal output based on a power demand corresponding to the determined activity level.
[0005] This Summary is provided to introduce some concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter. BRIEF DESCRIPTION OF THE DRAWINGS
[0006] Figure 1A and Figure 1B shows a simplified configuration of a target circuit using a capacitor as a power supply ( Figure 1A ) and an example activity profile associated with the target circuit ( Figure 1B ).
[0007] Figure 2A and Figure 2B An example configurable charge storage architecture is shown.
[0008] Figure 3 An example charge storage cell is shown.
[0009] Figure 4 Another example charge storage cell is shown.
[0010] Figures 5A-5D Shown is the Figure 1B An example activity profile shows an example application of a configurable charge storage network powered by .
[0011] Fig. 6A and Figure 6B Shown is the Figure 1B
[0046] Another example application of an example activity profile powered configurable charge storage network.
[0012] Figure 7 A network of charge storage cells is shown.
[0013] Figure 8 Another example network of charge storage cells is shown.
[0014] Fig. 9A and Fig. 9B Some example architectures of configurable charge storage networks are shown.
[0015] Fig.10 An example implementation of a context-aware power network applied to clocked charge domain logic cells is shown.
[0016] Fig.11 An example of a system with a context-based controller that also includes clock control is shown. DETAILED DESCRIPTION
[0017] A configurable charge storage network and control system are described that adjust the charge storage device configuration of the charge storage network based on the needs of an integrated circuit. The adjustment can be for current needs or expected needs of the integrated circuit.
[0018] Figure 1A A simplified configuration of a target circuit using a charge storage device as a power source is shown; and Figure 1B An example activity profile associated with a target circuit is shown. Figure 1A , the target circuit 100 can use the charge accumulated on the charge storage device 110 (represented by the power capacitor C1) to operate independently of the power supply 120. The target circuit 100 may require different amounts of charge depending on the activity level. Figure 1BAs shown, the target circuit 100 may have a low power demand during a time period TA1 , a medium power demand during a time period TA2 , a high power demand during a time period TA3 , and a low power demand during a time period TA4 .
[0019] Among the various considerations for selecting an appropriate capacitance to power a target circuit, selecting an appropriately sized capacitor (when used as a charge storage device for the target circuit) may be considered one of the key considerations. The size of the capacitor, or the amount of capacitance available from the charge storage device (for cases where the charge storage device may not be a capacitor), is selected to be able to power the circuit while maintaining the voltage output of the capacitor or other charge storage device within an acceptable range for normal operation of the associated circuit. For example, when a capacitor is used as a charge storage device, the size of the capacitor is typically selected to be large enough to withstand the worst-case activity level (usually high current demand) of the circuit. While a large capacitor can provide enough power to service larger circuits / intense activity, the size of the capacitor has other effects, such as increased manufacturing cost due to the larger silicon area that the capacitor may occupy; increased power consumption due to the higher current / power required to charge the capacitor; and increased time required to charge the capacitor to capacity.
[0020] like Figure 1B As shown in the example of , the activity of the target circuit is typically not constant; and may vary in intensity. Power capacitors are typically sized to handle the worst-case activity level. This forces chip designers to potentially select capacitors that are larger in size (and therefore have greater power and performance overhead) than would be required during periods of lower activity levels (where smaller capacitors can be used). Advantageously, in various embodiments, the described configurable charge storage network and control system that adjusts the configuration of the charge storage devices to minimize one or more of area, power consumption, and charging time can be implemented.
[0021] The described configurable charge storage architecture is able to sense the capacitance level and adapt it to the context of circuit operation. The context can be set based on time, operating mode, voltage level, current consumption and / or other operating parameters.
[0022] Figure 2A and Figure 2B An example configurable charge storage architecture is shown. Figure 2A and Figure 2B, the scenario-based controller 200 can receive any number of inputs about circuit operation and generate appropriate control signals (e.g., CTRL S1, CTRL S2, CTRL S3, ...) for controlling one or more charge storage units (CSUs) 220. For example, multiple operating parameters can be monitored and provided as inputs about circuit operation (including the operation of the target circuit 230). The scenario can be derived, for example, by monitoring the power supply voltage level (Vsup) on the voltage source output by the power supply 240, for example, via tap 251, and monitoring the core voltage level Vcore on the core voltage provided by the CSU 220, for example, via tap 252. In some cases, the scenario-based controller 200 generates a control signal based on a voltage drop or gradient of the voltage source or core voltage over time. In some cases, the scenario-based controller 200 generates a control signal based on the voltage difference between the voltage source and the core voltage. In some cases, the controller can generate an appropriate signal by combining or accessing a signal generator and applying an appropriate pulse width or frequency modulation. In some cases, the controller can generate an appropriate signal by connecting a DC voltage to a control output line.
[0023] The scenario may also or alternatively be derived, for example, by monitoring the activity of the circuit (e.g., actual activity data or command activity). For example, the scenario may be derived by monitoring a command bus or a data bus associated with the target circuit 230 (monitoring represented by tap 255). In some cases, the scenario-based controller 200 may monitor a command or control signal sequence (e.g., from a command bus associated with the target circuit 230) and / or a data sequence (e.g., via a data bus of the target circuit 230). The scenario-based controller may use the command or control signal sequence or the data sequence to determine the activity level (and corresponding power demand) of the target circuit 230. In some cases, the scenario-based controller 200 may receive an activity indicator (e.g., an activity level indicator) from, for example, an activity controller or other controller or processor used in the system. The activity indicator may directly or indirectly indicate the power level required by the target circuit 230. In some cases, the activity indicator may directly indicate the required power level by being one or more signals indicating high, medium, or low power demand (or other number of levels). In some cases, the activity indicator may indirectly indicate the required power level by being a signal that the scenario-based controller 200 may use to derive the appropriate power level. In some cases, the activity controller performs monitoring of the command or data bus and determines a corresponding activity level to provide an activity indicator to the context-based controller 200 .
[0024] In some cases, the scenario-based controller 200 may use a combination of any of the above scenarios to determine the appropriate control signal. The charge storage unit 220 may receive an appropriate control signal (e.g., a CTRL signal) that corresponds to one or more of CTRL S1, CTRL S2, CTRL S3, and any other control signal for operating the CSU 220. Using the control signal, the scenario-based controller 200 may adjust the capacitance of the charge storage unit 220, or a charging parameter (e.g., charging frequency) associated with the charge storage unit 220. In some cases, for example, with respect to Fig.11 In more detail, the scenario-based controller 200 can adjust the voltage level or frequency of the target circuit based on power demand, such as reducing the operating frequency of the target circuit in the case of higher power demand; or, for example, postpone or delay some operations when engaging the target circuit to prevent parallel occurrence. In some cases, some non-essential operations can be scheduled at a later time to manage (and possibly stagger) peak power demand.
[0025] The scenario-based controller 200 may include or have access to registers or other memory that stores appropriate mappings for power requirements and corresponding control signals based on received scenario inputs. For example, where the target circuit 230 includes at least a portion of an encryption circuit, an indication of an AES start command may indicate to the scenario-based controller 200 that the target circuit 230 will consume more power in order to perform encryption functions; and a register or other memory may be accessed to identify the appropriate control signal to be applied based on the identification that the AES start command has been found on the command bus.
[0026] Figure 3 An example charge storage unit is shown; and Figure 4 Another example charge storage unit is shown. Figure 3 The charge storage unit CSU-1 320 may include a plurality of charge storage devices, represented by parallel capacitors C1, C2, and C3; and switches S1, S2, S3, connecting the corresponding charge storage devices to a power supply (Vsup) and a target circuit (not shown) to provide a core voltage (Vcore). Figure 4 The example charge storage unit CSU-2 420 further includes switches S4 and S5, which can be used to controllably isolate the charge storage unit 420 from the power supply (Vsup). It should be understood that the number of charge storage devices (and corresponding switches) can be greater or less than the number shown. In addition, in some cases, more than one type of charge storage device can be included in a single charge storage unit.
[0027] Figures 5A-5D Shown is the Figure 1BAn example activity profile of an example application of a configurable charge storage network powered by Figure 5A , for time TA1, the context-aware controller may have determined that nominal (low level) activity is (or will be) occurring based on the context, and generates a control signal to connect a charge storage device corresponding to a smaller capacitor. For example, most switches 520 may remain open, so that one charge storage device 521 is coupled to the target circuit 530. In some cases, the charge storage device 521 is coupled some time before the target circuit operation begins to accumulate charge from the power supply 540.
[0028] refer to Figure 5B , for time TA2, the context-aware controller may have determined that a medium level of activity is (or will be) occurring based on the context, and generates control signals to connect a sufficient number of charge storage devices to support the appropriate voltage requirements of the circuit for the predefined medium level of activity. For example, some switches 520 may remain open so that two (or more) charge storage devices 521, 522 are coupled to the target circuit 530. In some cases, the charge storage devices 521, 522 are coupled for a period of time before the target circuit operation begins in order to accumulate charge from the power supply 540.
[0029] refer to Figure 5C , for time TA3, the context-aware controller may have determined that a high level of activity is being (or will be) performed based on the context, and generates control signals to connect a sufficient number of charge storage devices to support the appropriate voltage requirements of the circuit for the predefined high level of activity. For example, all switches may be closed so that all charge storage devices 521, 522, 525 are coupled to the target circuit 530. In some cases, the charge storage devices 521, 522, 525 are coupled some time before the target circuit operation begins in order to accumulate charge from the power supply 540.
[0030] refer to Figure 5D , for time TA4, the context-aware controller may have determined based on the context that a nominal (low level) activity is (or will be) taking place, and generates a control signal to connect the charge storage device 521 corresponding to the smaller capacitor, as described with respect to Figure 5A As the operation of the target circuit continues, the context-aware controller can determine the activity level of the target circuit according to the context (as described in reference Figure 2A and Figure 2B as described), and generates appropriate control signals to connect and disconnect the charge storage device target circuit of the charge storage unit.
[0031] It should be understood that more or fewer activity levels can be defined for the charge storage unit, so that more or fewer charge storage devices are grouped for connection to the target circuit. In addition, the number of charge storage devices coupled for a particular activity level can be selected according to design choice based on the power requirements of the target circuit and other design considerations (such as parasitic effects, etc.) used to manufacture the chip.
[0032] Fig. 6A and Figure 6B Shown is the Figure 1B Another example application of a configurable charge storage network powered by an example activity profile. Fig. 6A , the charge storage unit 600 may include a single charge storage device 610, and a corresponding switch 620 that directly or indirectly connects the charge storage device 610 to the target circuit 630 and the power supply 640 (e.g., other switches may be included between the power supply 640 and the charge storage unit 600). Instead of using multiple charge storage devices in parallel to increase capacitance, the charge storage unit 600 uses a certain time to operate to charge the charge storage device 610.
[0033] Figure 6B Shown for Figure 1B 610 for each activity level. As can be seen, for time TA1, the context-aware controller may have determined based on the context that nominal (low level) activity is (or will be) occurring, and generated a control signal so that switch 620 connects charge storage device 610 for the nominal charging time. For time TA2, the context-aware controller may have determined based on the context that a medium level of activity is (or will be) occurring, and generated a control signal so that switch 620 connects charge storage device 610 for a longer time than occurred for TA1. In some cases, another control signal may be applied to refresh the charge storage device (see, e.g., Figure 6B 620 connects charge storage device 610 for a longer period than occurred for TA1 and T2. Then, for time TA4, the context-aware controller may have determined based on the context that a nominal (low level) activity is (or will be) occurring, and generates a control signal so that switch 620 connects charge storage device 610 as described with respect to TA1.
[0034] Despite Figure 6B620. Although a pulse width modulated signal is shown in FIG. 64, in some cases, various charging of the charge storage device is achieved using frequency modulation, or a combination of frequency modulation and pulse width modulation. For example, frequency modulation can be used so that during high power demand activities, the frequency of the control signal of switch 620 can be increased.
[0035] In some cases, the charge storage unit may include a selectable resistance path, wherein the resistance of the charge path may be varied based on a determined activity level of the target circuit. For example, a lower resistance path may be selected during a time period when the target circuit requires more power, while a higher resistance path may be selected when nominal power is required.
[0036] In some cases, the charge storage cells may be constructed from standardized cells to facilitate layout during the design and / or layout and / or routing phases of the chip.A network of charge storage cells may be created to provide the required capacitance.
[0037] Figure 7 An example network of charge storage cells is shown; and Figure 8 Another example network of charge storage cells is shown. Figure 7 , the charge storage network 700 may include a plurality of charge storage units (720-1, 720-2, ..., 720-N) connected in parallel. Figure 8 , the charge storage network 800 may include a plurality of parallel charge storage units (820-1, 820-2, . . . , 820-N), and switches CSX and CXY for isolating the charge storage network 800 from a power source.
[0038] The CSU of Network 700 and Network 800 can be connected to Figure 3 CSU-1 320, Figure 4 6. In some cases, a combination of any of the described CSUs may be included in the same network.
[0039] Fig. 9A and Fig. 9B Some example architectures of configurable charge storage networks are shown. Fig. 9A, the charge storage network 900 can be controlled by a scenario-based controller 910. Similar to what is described with respect to the scenario-based controller 200 of FIG. 2 , the scenario-based controller 910 can receive any number of inputs regarding circuit operation and generate appropriate control signals for controlling the charge storage network 900. For example, the scenario can be derived by monitoring the voltage level (Vsup) on the voltage source and monitoring the core voltage level Vcore on the core voltage. The scenario can also or alternatively be derived, for example, by monitoring a command bus or a data bus. In addition, in some cases, the scenario-based controller 910 can receive an activity indicator from, for example, an activity controller or other controller or processor used in the system. To support the charge storage network, the scenario-based controller 910 can generate control signals (e.g., signals CTRL S1, CTRL S2, CTRL S3, ..., CTRL SN) for more than one charge storage unit forming the charge storage network. For the case where the charge storage network is configured as the charge storage network 800, the scenario-based controller 910 can also generate CTRL SX and CTRL SY signals for controlling switches CSX and CSY. Reference Fig. 9B , for which the charge storage network 940 includes such as Figure 4 In the case of a charge storage unit such as CSU-2 420 described above, scenario-based controller 950 may further generate control signals for connecting and disconnecting the charge storage unit from the power source at a level more granular with the configuration of charge storage network 800. Of course, in some cases, a charge storage unit such as CSU-2 420 may be implemented in charge storage network 800, and scenario-based controller 950 may further generate CTRLSX and CTRLSY signals for controlling switches CSX and CSY.
[0040] In some cases, each charge storage unit has a corresponding context-based controller. In some cases, each charge storage network has a corresponding context-based controller. In some cases, a context-based controller can provide control signals for multiple charge storage networks. In still other cases, a single context-based controller is provided for the entire chip.
[0041] The described context-aware power network can be applied to time-controlled charge domain logic and other safety systems. Fig.10 An example implementation of a context-aware power network applied to a timed charge domain logic unit is shown. Fig.10, charge storage network 1000 and context-based controller 1010 may be used to implement a charge storage device of an independent power supply 1020 of a circuit or unit 1030. Charge storage network 1000 may be any configuration of charge storage units and charge storage networks implemented as described herein.
[0042] In a system incorporating timed charge domain logic, a charge storage device is controllably coupled to a circuit and / or cell of a circuit to provide external power source independent power to the circuit and / or cell. An example configuration for facilitating such external power source independent power is shown using five switches (S1, S2, S3, S4, and S5), which in this case are controlled by a controller 1040 to connect and disconnect the charge storage network 1000 from an external power source 1050 and circuits / cells 1030. Separate from the safety operations performed by the controller 1040, a scenario-based controller ("CSN controller") 1010 monitors any of a number of signals to determine an appropriate capacitance to use as part of the charge storage network 1000. In some cases, the scenario-based controller adjusts charging parameters (e.g., charging frequency) associated with the power network. For example, the context of the context-based controller 1010 can be derived by monitoring the command bus or data bus (represented by tap 1051), monitoring the supply voltage level on the voltage / power source 1050 (via tap 1053B), monitoring the core voltage level of the core voltage (via tap 1052), or a combination thereof. The output of the tap can be directly or indirectly input to the CSN controller 1010 via inputs 1054, 1055, 1056, 1057. Of course, the input can be any of the inputs described with respect to controllers 200, 910, and 950.
[0043] As described above, a scenario-based controller can adjust the frequency of the target circuit based on power demand, for example, by reducing the operating frequency of the target circuit to reduce consumption of the charge storage unit under higher power demand conditions; or by, for example, deferring or delaying some operations when engaging the target circuit to prevent them from occurring in parallel.
[0044] The dynamic power consumption of a circuit is proportional to capacitance, voltage, and frequency. In certain cases, if higher switching activity creates higher power demands (due to increased capacitance switched during (one or more) clock cycles), the frequency of these operations can be reduced (by reducing the clock frequency) to keep power consumption manageable. Of course, it is still desirable to have the clock frequency at an appropriate frequency to meet timing constraints. In some cases, some non-essential operations can be scheduled at a later time to manage (and possibly stagger) peak power demands.
[0045] When the frequency is reduced in the case of higher power demand, the operating frequency of the target circuit can be reduced, for example, by adjusting the operating clock of the target circuit. In some cases, the frequency of the clock signal is reduced. In some cases, the operating clock can be suspended and / or a different clock signal (e.g., with different characteristics such as jitter, noise, slew rate, etc.) can be applied.
[0046] Clock gating is one way to delay an operation. Another way is to use a control circuit to abort a particular operation. Delaying of operations can be performed in conjunction with a scheduler or timer to manage and / or stagger operations. In some cases, a context-based controller gates a circuit's clock signal based on the circuit's activity level. In some cases, a context-based controller adjusts the circuit's operating frequency based on the circuit's activity level.
[0047] Fig.11 An example of a system having a scenario-based controller that also includes clock control is shown. In addition to generating control signals for one or more charge storage units 1120, the scenario-based controller 1100 can also adjust the clock of the target circuit 1130 based on the activity level of the target circuit 1130. In the example shown, the scenario-based controller 1100 can include or communicate with some or all components of the clock control circuit 1140. Through the clock control circuit 1140, the clock (CLK) of the target circuit 1130 can be adjusted by gating or by adjusting its frequency or other characteristics. Here, the clock control circuit 1140 includes a selector component (represented by MUX 1142) and a clock gating circuit (represented by AND gate 1144).
[0048] The selector component ("clock selector") enables the scenario-based controller 1100 to change the clock provided to the target circuit 1130, which can be used to change the frequency of the target circuit operation. One of the multiple clocks can be selected to adjust the operating frequency of the circuit. For example, the enable signal of MUX 1142 can be used to select two or more clock signals. In the example, two clock signals CLK1 and CLK2 are shown. Based on the activity level of the circuit, the selector component can select CLK1 to output to CLK or select CLK2 to output to CLK. The clock gating circuit can be used to pause the clock. For example, a static voltage (high or low) (e.g., via an enable signal) can be selected to output to the CLK line to pause the clock. It should be understood that although only an AND gate is shown, the clock gating circuit can be formed by at least one sequential element and a combinational element. In addition, the selector component and the clock gating circuit can be independent circuits, or can be configured in series in any order.
[0049] The enable signal for the selector component and / or the clock gating circuit can be generated or obtained from the same signal used to generate the control signal for the (one or more) charge storage unit. For example, the scenario can be derived by monitoring the voltage level (Vsup) on the voltage source and monitoring the core voltage level Vcore on the core voltage. In some such cases, the scenario-based controller 1100 selects / generates the appropriate enable signal based on the voltage drop or gradient of the voltage source or core voltage over time. In some cases, the scenario-based controller 1100 selects / generates the appropriate enable signal based on the voltage difference between the voltage source and the core voltage. The scenario can also or alternatively be derived by monitoring the activity of the circuit (e.g., actual activity data or command activity). For example, the scenario can be derived by monitoring the command bus or data bus associated with the target circuit 1130. In some cases, the scenario-based controller 1100 monitors a command or control signal sequence (e.g., from a command bus associated with the target circuit 1130) and / or a data sequence (e.g., via the data bus of the target circuit 1130). The context-based controller can use a sequence of command or control signals or a sequence of data to determine the activity level (and corresponding power demand) of the target circuit 1130. In some cases, the context-based controller 1100 can receive an activity indicator from, for example, an activity controller or other controller or processor used in the system. In some cases, the context-based controller 1100 can use a combination of any of the above scenarios to determine the appropriate enable signal.
[0050] For implementations that use a context-aware controller to suspend specific operations of a target circuit, any combination of the above scenarios used in a clock gating scenario can be used to determine the appropriate suspend / enable operation signal from the context-aware controller to the target circuit or other circuitry that controls the operation of the target circuit.
[0051] In any of the charge storage cells described herein, the charge storage device may be, but is not limited to, a MOSFET (metal oxide semiconductor field effect transistor) (where the gate is connected to the source, drain, and (in some cases) the body terminals), a MOS (metal oxide semiconductor) capacitor, a MIM (metal insulator metal) capacitor, a MOM (metal oxide metal) capacitor, a discrete capacitor, or a charge coupled device.
[0052] In any charge storage unit described herein, the switch can be implemented by a transistor. Any suitable transistor known in the art can be used, for example, a bipolar junction transistor, a metal oxide semiconductor field effect transistor (MOSFET) or a combination thereof. Each MOSFET used can be a p-type MOSFET (PMOS) or an n-type MOSFET (NMOS). In some cases, a transmission gate configuration can be used. In some cases, a diode can be used as one or more switches. In other cases, a switch based on a micro-electromechanical system (MEMS) can be used.
[0053] As shown herein, in a first example, a system includes: a circuit; a power network coupled to the circuit to provide a core voltage to the circuit; and a scenario-based controller that monitors a power supply voltage level of a power supply, monitors a core voltage level of a core voltage, and monitors activity of the circuit to derive an activity level of the circuit; and based on the activity level of the circuit, adjusts a capacitance of the power network or a charging parameter associated with the power network to correspond to a power demand associated with the activity level. The scenario-based controller may operate as described in any implementation of the present invention.
[0054] In a second example according to the first example, the power network comprises at least one charge storage unit, each of the at least one charge storage unit comprises: a plurality of charge storage devices connected in parallel through respective switches controlled by the context aware controller.
[0055] In a third example according to the second example, the power network further comprises: a network-level switch controllably coupling and decoupling the at least one charge storage unit to and from the supply voltage.
[0056] In a fourth example according to the first example, the power network includes: at least one charge storage unit, each of the at least one charge storage unit includes: a plurality of charge storage devices connected in parallel through corresponding switches controlled by a context-aware controller; and a switch that controllably couples the plurality of charge storage devices to and decouples them from a power supply voltage.
[0057] In a fifth example according to the fourth example, the power network further comprises: a network-level switch controllably coupling and decoupling the at least one charge storage unit to and from the supply voltage.
[0058] In a sixth example based on the first example, the power network includes: at least one charge storage unit, each of the at least one charge storage unit includes: a charge storage device; and a switch coupled to the charge storage device, wherein the context-aware controller adjusts the capacitance of the power network or a charging parameter associated with the power network by generating a control signal to connect and disconnect the switch to a power source for a length of time corresponding to a power demand associated with an activity level.
[0059] As shown herein, in a seventh example, a context-aware power network includes: at least one charge storage unit; and a context-based controller including at least one input receiving a context of circuit operation; and at least one control signal output coupled to the at least one charge storage unit, wherein the context-based controller determines an activity level of the circuit based on the context of circuit operation received via the at least one input, and generates at least one control signal to be output via the at least one control signal output based on a power demand corresponding to the determined activity level. The context-based controller may operate as described in any implementation herein.
[0060] In an eighth example according to the seventh example, the power network includes: at least one charge storage unit, each of the at least one charge storage unit includes: a plurality of charge storage devices connected in parallel through corresponding switches controlled by the context-aware controller.
[0061] In a ninth example according to the eighth example, the power network further comprises: a network-level switch controllably coupling and decoupling the at least one charge storage unit to and from the supply voltage.
[0062] In a tenth example according to the seventh example, the power network includes: at least one charge storage unit, each of the at least one charge storage unit includes: a plurality of charge storage devices connected in parallel through corresponding switches controlled by a context-aware controller; and a switch that controllably couples the plurality of charge storage devices to and decouples them from a power supply voltage.
[0063] In an eleventh example according to the tenth example, the power network further includes: a network-level switch controllably coupling and decoupling the at least one charge storage unit to and from the supply voltage.
[0064] In a twelfth example according to the seventh example, an electric power network includes: at least one charge storage unit, each of the at least one charge storage unit includes: a charge storage device; and a switch coupled to the charge storage device, wherein the context-aware controller adjusts the capacitance of the electric power network or a charging parameter associated with the electric power network by generating a control signal to connect and disconnect the switch to a power source for a length of time corresponding to a power demand associated with an activity level.
[0065] The functions, methods, and processes described herein may be implemented, at least in part, by one or more hardware modules (or logic components). For example, a hardware module may include, but is not limited to, an application specific integrated circuit (ASIC) chip, a field programmable gate array (FPGA), a system on a chip (SoC) system, a complex programmable logic device (CPLD), and other known or later developed programmable logic devices. When a hardware module is activated, the hardware module executes the functions, methods, and processes contained in the hardware module.
[0066] Although the subject matter has been described in language specific to structural features and / or actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Instead, the specific features and actions described above are disclosed as examples of implementing the claims, and other equivalent features and actions are intended to be within the scope of the claims.
Claims
1. A system for context-aware power network, comprising: Circuit; a power network coupled to the circuit to provide a core voltage to the circuit; as well as A context-based controller, the context-based controller: monitoring a supply voltage level of a power supply, monitoring a core voltage level of the core voltage, and monitoring activity of the circuit to derive an activity level of the circuit, wherein the context-based controller monitors the activity of the circuit by monitoring a sequence of command or control signals from a command bus to the circuit or by monitoring a sequence of data from a data bus of the circuit; and Based on the activity level of the circuit, a capacitance of the power network or a charging parameter associated with the power network is adjusted to correspond to a power demand associated with the activity level.
2. The system according to claim 1, wherein: The scenario-based controller derives the activity level from the power supply voltage level based on a voltage drop or gradient across the power supply.
3. The system according to claim 1, wherein: The context-based controller derives the activity level from the core voltage level based on a voltage drop or gradient across the core voltage.
4. The system according to claim 1, wherein: The context-based controller derives the activity level from the supply voltage level and the core voltage level based on a voltage difference between the supply voltage level and the core voltage level.
5. The system according to claim 1, wherein: The scenario-based controller also includes: an input for receiving an activity level indicator; and based on the activity level indicator received via the input for receiving the activity level indicator, adjusting the capacitance of the power network or a charging parameter associated with the power network to correspond to a power demand associated with the activity level corresponding to the activity level indicator.
6. The system according to claim 1, wherein: The scenario-based controller adjusts the capacitance of the power network or a charging parameter associated with the power network by generating control signals to connect and disconnect appropriate switches of the power network.
7. The system according to claim 1, wherein: The context-based controller also adjusts a clock of the circuit based on an activity level of the circuit.
8. The system according to claim 7, wherein: The context-based controller gates the clock of the circuit based on an activity level of the circuit.
9. The system according to claim 7, wherein: The context-based controller adjusts an operating frequency of the circuit based on an activity level of the circuit.
10. A situation-aware power network, comprising: at least one charge storage unit; as well as a context-based controller, the context-based controller comprising at least one input receiving a context of circuit operation, wherein the context-based controller monitors activity of the circuit by monitoring a sequence of commands or control signals from a command bus to the circuit or by monitoring a sequence of data from a data bus of the circuit; and at least one control signal output coupled to the at least one charge storage unit, wherein the scenario-based controller determines an activity level of the circuit based on a scenario of operation of the circuit received via the at least one input, and generates at least one control signal for output via the at least one control signal output based on a power demand corresponding to the determined activity level.
11. The network according to claim 10, wherein: The context-based controller determines the activity level of the circuit by monitoring a supply voltage level of a power supply and monitoring a core voltage level of a core voltage provided to the circuit to derive the activity level of the circuit.
12. The network according to claim 10, wherein: The context-based controller determines the activity level of the circuit by monitoring the activity of the circuit based on the sequence of the command or control signals or from the sequence of the data to derive the activity level of the circuit.
13. The network according to claim 10, wherein: The context-based controller also adjusts a clock of the circuit based on a power demand corresponding to the determined activity level.
14. The network according to claim 13, wherein: The context-based controller includes a clock gating circuit that gates the clock of the circuit based on the activity level of the circuit.
15. The network according to claim 14, wherein: The clock gating circuit includes at least one sequential element and a combinational element.
16. The network of claim 13, wherein: The context-based controller adjusts an operating frequency of the circuit based on the activity level of the circuit.
17. The network according to claim 16, wherein: The context-based controller includes a clock selector that selects one of a plurality of clocks to adjust an operating frequency of the circuit based on the activity level of the circuit.
