Patent application number | Description | Published |
20090320031 | Power state-aware thread scheduling mechanism - A system filter is maintained to track which single-thread cores [or which multi-threaded logical CPUs] are in a low-latency power state. For at least one embodiment, low-latency power states include an active C | 12-24-2009 |
20100146513 | Software-based Thread Remapping for power Savings - On a multi-core processor that supports simultaneous multi-threading, the power state for each logical processor is tracked. Upon indication that a logical processor is ready to transition into a deep low power state, software remapping (e.g., thread-hopping) may be performed. Accordingly, if multiple logical processors, on different cores, are in a low-power state, they are re-mapped to same core and the core is then placed into a low power state. Other embodiments are described and claimed. | 06-10-2010 |
20100332856 | System and method for processor utilization adjustment to improve deep C-state use - In some embodiments, the invention involves modification of the processor utilization calculations that are used by operating system power management services to improve processor efficiency. An embodiment of the present invention is a system and method relating to power management policies under operating system control. In at least one embodiment, the present invention is intended to modify the processor utilization evaluation process so that C-state transition time and/or unhalted reference cycles are included in the calculation. Other embodiments are described and claimed. | 12-30-2010 |
20110161627 | MECHANISMS TO AVOID INEFFICIENT CORE HOPPING AND PROVIDE HARDWARE ASSISTED LOW-POWER STATE SELECTION - An apparatus and method is described herein for avoiding inefficient core hopping and providing hardware assisted power state selection. Future idle-activity of cores is predicted. If the residency of activity patterns for efficient core hop scenarios is predicted to be large enough, a core is determined to be efficient and allowed. However, if efficient activity patterns are not predicted to be resident for long enough—inefficient patterns are instead predicted to be resident for longer—then a core hop request is denied. As a result, designers may implement a policy for avoiding core hops that weighs the potential gain of the core hop, such as alleviation of a core hop condition, against a penalty for performing the core hop, such as a temporal penalty for the core hop. Separately, idle durations associated with hardware power states for cores may be predicted in hardware. Furthermore, accuracy of the idle duration prediction is determined. Upon receipt of a request for a core to enter a power state, a power management unit may select either the hardware predicted power state, if the accuracy is high enough, or utilize the requested power state, if the accuracy of the hardware prediction is not high enough. | 06-30-2011 |
20150135189 | SOFTWARE-BASED THREAD REMAPPING FOR POWER SAVINGS - On a multi-core processor that supports simultaneous multi-threading, the power state for each logical processor is tracked. Upon indication that a logical processor is ready to transition into a deep low power state, software remapping (e.g., thread-hopping) may be performed. Accordingly, if multiple logical processors, on different cores, are in a low-power state, they are re-mapped to same core and the core is then placed into a low power state. Other embodiments are described and claimed. | 05-14-2015 |
20150185797 | DYNAMIC POWER MEASUREMENT AND ESTIMATION TO IMPROVE MEMORY SUBSYSTEM POWER PERFORMANCE - Dynamic monitoring of current draw by a memory device or memory subsystem can enable a power management system to adjust a memory access performance parameter based on monitored power usage. The system can generate a power usage characterization for the memory device and/or memory subsystem based on monitoring current draw for a known pattern, and then subsequently use the power usage characterization to determine how to adjust the memory access performance parameter. | 07-02-2015 |
20150381237 | METHODS AND SYSTEMS FOR SERVER POWER LINE COMMUNICATION - A server system includes a common power bus, a power supply to provide direct current (DC) power through the common power bus, at least one node including a processor to receive the DC power through the common power bus, a transmitter capacitive coupled to the common power bus to transmit a power information signal from the power supply through the common power bus, and at least one receiver capacitive coupled to the common power bus to receive the power information signal transmitted by the transmitter and to provide the received power information signal to the at least one node. A plurality of buffers respectively coupled between the common power bus and each of the power supply and the at least one node provide path separation for high frequency and low frequency currents. | 12-31-2015 |
20160054774 | PROFILING A JOB POWER AND ENERGY CONSUMPTION FOR A DATA PROCESSING SYSTEM - A process identifier for a job is collected. The job runs on a plurality of nodes. The job is identified using the process identifier. A node for the job is identified. An amount of power consumed by the node to run the job is determined. | 02-25-2016 |
20160054780 | POWER AWARE JOB SCHEDULER AND MANAGER FOR A DATA PROCESSING SYSTEM - An indication of a mode for a job is received. An available power for the job is determined based on the mode. A first power for the job is allocated based on the available power. A first frequency for the job is determined based on the available power. The first power is adjusted based on the available power. | 02-25-2016 |