6 results about "Throughput degradation" patented technology

The application provides an AI intelligent calculation platform reasoning resource scheduling method and system, relates to the technical field of computer resource management and task scheduling, and obtains task load information of a to-be-reasoned task, memory allocation records of an AI intelligent calculation platform in a preset time window, and a retransmission event timing sequence of a bus data link layer; obtains the discrete entropy of each computing node through a Shannon entropy formula according to the memory allocation records; counts the retransmission times of the retransmission event timing sequence in a sliding time window to generate a communication congestion index of each computing node; determines a weight coefficient according to the task load information, constructs a weighted bipartite graph, and determines a matching edge set that satisfies the most matching edges and the smallest sum of edge weights in the weighted bipartite graph by using a Hungarian algorithm, so as to be used for resource scheduling of the to-be-reasoned task. The problem of throughput decline caused by serious memory fragmentation and communication interference in the large model concurrent reasoning scene is solved.

The invention is suitable for the technical field of computing power server cluster operation and maintenance and GPU resource scheduling, and provides an intelligent operation and maintenance monitoring method and platform for a computing power server cluster, and the method comprises the steps: determining a turning threshold value corresponding to the short-period fluctuation degree of a candidate GPU with a specified mode, and obtaining a turning threshold value corresponding to the short-period fluctuation degree of the candidate GPU; and acquiring the short-period fluctuation degree of the latest historical task sample corresponding to the candidate GPU, generating a correction factor according to the deviation degree of the latest short-period fluctuation degree relative to the turning threshold, and performing down-regulation correction on the suitability value of the candidate GPU by using the correction factor. According to the method, excessive intervention on a normal GPU is avoided, and meanwhile, the risk of throughput degradation in the middle and later periods caused by implicit operation fluctuation accumulation when a computing task which is relatively stable in video memory occupation requirement, relatively long in execution time and sensitive to continuous computing power stability is processed is effectively reduced; and the stability, the accuracy and the overall operation efficiency of the scheduling decision of the computing power server cluster are improved.

Disclosed is a wireless communication base station apparatus whereby it is possible to prevent degradation of throughput of LTE terminals, even when LTE terminals and LTE+ terminals are present together. In this apparatus, a setting section (105) sets in each subframe a resource block in which is arranged a reference signal that is employed solely by LTE+ terminals, based on the pattern of arrangement of reference signals employed solely by LTE+ terminals. In the case of symbols that are mapped to antennas (110-1) to (110-4), an arrangement section (106) arranges the characteristic cell reference signals employed by both LTE terminals and LTE+ terminals in all of the resource blocks in a single frame. In contrast, in the case of the symbols that are mapped to the antennas (110-5) to (110-8), the arrangement section (106) arranges in some of the resource blocks, that are set in accordance with the setting results input from a setting section (105), the characteristic cell reference signals that are employed solely by the LTE+ terminals.

A system includes one or more movers configured to move along a track of an industrial automation system and a computing device. The computing device receives data associated with the movers as the movers move around the track. The data includes a first amount of time for at least one of the movers to move between two stations disposed along the track, a second amount of time for the at least one mover to settle at a station, a third amount of time for a respective robot at the station to prepare to perform a process, and a fourth amount of time for the respective robot to perform the process. The computing device identifies throughput degradation based on the first amount of time, the second amount of time, the third amount of time, or the fourth amount of time, or any combination thereof.

The application belongs to the field of communication, and particularly relates to a temperature regulation method and device, network equipment and a readable storage medium. The method comprises the following steps: determining target power consumption; reducing the transmitting power of the network equipment based on the target power consumption; determining the throughput of the network equipment and a plurality of terminal devices based on the reduced transmitting power of the network equipment; and if the throughput of the plurality of terminal devices does not decrease, stopping the determination of the target power consumption when the temperature of the network equipment is less than or equal to a preset temperature threshold. That is, the embodiment of the application reduces the transmitting power of the network equipment based on the target power consumption, and determines the throughput of the plurality of terminal devices; if the throughput of the plurality of terminal devices does not decrease, the determination of the target power consumption is stopped, so that the situation that the throughput of all the terminal devices decreases can be avoided while the temperature of the network equipment is regulated by reducing the transmitting power.