Abstract:
An integrated circuit compensates for circuit aging by measuring the aging with an aging sensor and controlling a supply voltage based on the measured aging. The operating environment for the aging sensor can be set to reduce impacts of non-aging effects on the measured aging. For example, the operating environment can use a temperature inversion voltage. An initial aging measurement value which is the difference between an initial aged measurement and an initial unaged measurement can be stored on the integrated circuit. A core power reduction controller can use the measured aging and the stored initial aging measurement value to update a performance-sensor target value and then perform adaptive voltage scaling using the using the updated performance-sensor target value.
Abstract:
A method to prevent a malicious attack on CPU subsystem (CPUSS) hardware is described. The method includes auto-calibrating tunable delay elements of a dynamic variation monitor (DVM) using an auto-calibration value computed in response to each detected change of a clock frequency (Fclk)/supply voltage (Vdd) of the CPUSS hardware. The method also includes comparing the auto-calibration value with a threshold reference calibration value to determine whether the malicious attack is detected. The method further includes forcing a safe clock frequency (Fclk)/safe supply voltage (Vdd) to the CPUSS hardware when the malicious attack is detected.
Abstract:
A hardware system is disclosed for active-core-based performance boost. In an example aspect, the hardware system includes multiple cores and a power mode manager. Each core can be powered up if active or powered down if inactive. The power mode manager manages a power mode collection including an independent power mode collection and an active-core-dependent power mode collection. The power mode manager includes a software-accessible power mode manager and a hardware-reserved power mode manager. The software-accessible power mode manager provides a power-mode-triggering pathway to enable software to trigger activation of an independent power mode of the independent power mode collection. The hardware-reserved power mode manager excludes the software from being able to trigger activation of a dependent power mode of the active-core-dependent power mode collection and triggers activation of a dependent power mode of the active-core-dependent collection based on a number of active cores of the multiple cores.
Abstract:
Apparatuses and methods to adjust voltage for thermal mitigation are provided. The apparatus includes a circuit, a plurality of switches configured to provide power of a power domain to the circuit, a plurality of thermal sensors disposed at different locations about the circuit and configured to detect temperatures at the different locations, and a control circuit configured to determine that one of the detected temperatures at one of the locations exceeds a temperature threshold, and in response, adjust one or more of the plurality of switches in proximity with the one location to reduce power provided to the circuit. The method includes providing power of a power domain through a plurality of switches, detecting a temperature at a location exceeding a temperature threshold, and adjusting the plurality of switches in proximity with the location to reduce the power provided, in response to the detecting the temperature exceeding the temperature threshold.
Abstract:
In certain aspects, a system comprises a voltage-droop mitigation circuit configured to monitor voltage droop in a supply voltage supplied to a circuit, and to perform voltage-droop mitigation for the circuit if the monitored voltage droop is equal to or greater than a droop threshold. In one aspect, the system also includes a performance monitor configured to track a number of clock cycles over which the voltage-droop mitigation circuit performs the voltage-droop mitigation within a time duration, and to adjust the droop threshold based on the number of clock cycles. In another aspect, the system also includes a performance monitor configured to track a number of times that the voltage-droop mitigation circuit performs the voltage-droop mitigation within a time duration, and to adjust the droop threshold based on the number of times that the voltage-droop mitigation circuit performs the voltage-droop mitigation within the time duration.
Abstract:
A method including receiving an indication of a number of active processing units in a computer processor; in response to receiving the indication, determining an appropriate operating voltage margin for the computer processor; reducing an operating frequency of the active processing units in response to receiving the indication; adjusting a power supply to increase or decrease a voltage to the computer processor in accordance with the appropriate operating voltage margin; and increasing the operating frequency of the active processing units in response to an acknowledgment that the power supply has been adjusted.
Abstract:
The systems and method described herein provide efficient (e.g., low power and low area) techniques to track performance in numerous supply domains with heterogeneous circuits that are used in a large system-on-a-chip integrated circuit (SoCs). The heterogeneous circuits can include circuits made with different devices, different cell libraries, and different hard macros that are in different power supply domains. Performance measurements from performance sensors (or process-voltage-temperature (PVT) sensors) that are spread about the SoC are collected and processed to determine voltage levels for each of the supply domains. A single controller can receive can determine voltage levels for a whole SoC. The performance sensors are connected to the controller by a scan chain. The techniques are flexible and can be easily adapted for use in SoCs with different power supply domains and types of circuits.