Processor of energy sensing network and processing method thereof

Abstract

The invention discloses a processor of an energy sensing network and a processing method thereof in the technical filed of network application, the processor comprises a co-processor module, a plurality of micro-engine modules, a thread monitoring module, a performance monitoring module, a voltage regulator module and a voltage decision-making module, wherein the micro-engine modules are connected with the thread monitoring module to transmit thread state information; the thread monitoring module is connected with the voltage decision-making module to transmit the thread state information; the voltage decision-making module is connected with the co-processor module to transmit voltage decision-making information; the co-processor module is connected with the voltage regulator module to transmit voltage regulator commands; the voltage regulator module is connected with the micro-engine modules to transmit physical signals for regulating voltage; the micro-engine modules are connected with the performance monitoring module to transmit system performance information; and the performance monitoring module is connected with the co-processor module to transmit the signal regulating the voltage to be maximum. In the invention, periodic statistics is performed on the task running condition of each core, and the purposes of low power consumption and high performance are achieved through dynamically changing the voltage of each core.

Description

technical field [0001] The present invention relates to a device and method in the field of network technology, in particular to a processor of an energy-aware network based on a task pipeline and a processing method thereof. Background technique [0002] Low power consumption has always been one of the most important issues in processor design and manufacture. Power consumption limits the further improvement of performance. Network processors are integrated in a large number of racks in application scenarios. High power consumption further causes system stability problems, and the corresponding system cooling components will consume more power. In addition, with the rise of low-carbon economy in recent years, more electricity consumption means more carbon emissions on the upstream electricity supplier side. In the past, in the design of processors, performance improvement was the first goal, that is, performance perception. Now, with energy as a reference, the network pro...

AI Technical Summary

Problems solved by technology

Its shortcoming is: consider the situation is simple, not flexible enough, it assumes that each core handles the same task, there is no correlation between different cores, this method cannot handle the situation of heterogeneity between cores (that is, different cores process different tasks ); did not fully consider the characteristics of the task pipeline in the network processor, there is a data dependency between several cores of the same task pipeline, and the shutdown of one core will affect other cores in the pipeline; it will affect the idle core The practice is to turn it off, resulting in inflexible handling

Images