Quanta Computer Incorporated

Li-Tsung Chen

Yen-Ping Tung

Ching-Chih Shih

A method for testing a closed cooling loop for leaks. The method includes causing air to be pumped into the cooling loop. The method further includes identifying an initial internal temperature reading and an initial internal pressure reading of the cooling loop. In response to a predetermined amount of time passing since the air was pumped into the cooling loop, the method includes identifying an updated internal temperature reading and an updated internal pressure reading of the cooling loop. Thereafter, (i) the initial internal temperature, (ii) the initial internal pressure readings, (iii) the updated internal temperature, and (iv) the updated internal pressure readings, are used to determine if the cooling loop has a leak. Furthermore, in response to determining that the cooling loop has a leak, a warning is issued.

Leak auto detection for a rack liquid cooling system

A rack server system and an auto-addressing method thereof are disclosed. The rack server system comprises a plurality of backplanes and a rack management controller (RMC). The backplanes comprise a plurality of inter-integrated circuit (I2C) switches. The RMC comprises an I2C channel connected to the backplanes. When the RMC initializes the backplanes, the RMC controls a plurality of reset signals to be an enable level to reset the I2C switches, and automatically addresses a plurality of different I2C device addresses to the I2C switches. The RMC changes the reset signals to be a disable level from the enable level after the RMC addressed the I2C switches.

Rack server system and auto-addressing method thereof

In some implementations, a system administrator can specify a maximum amount of power for a data center or a rack of servers. If the data center or rack consumes more than a threshold amount of power, the data center or rack can be put into power capping mode. Once in power capping mode, servers in the data center or rack can be power capped. The amount of power at which a server is capped can be dynamically determined based on the amount of discretionary power available to distribute among the power capped servers in the rack or data center and the amount of power that the server consumes relative to other servers in the data center or rack. Once power consumption in the data center or rack falls below a threshold level, power capping for the data center or rack can be disabled.

Dynamic power capping of a subset of servers when a power consumption threshold is reached and allotting an amount of discretionary power to the servers that have power capping enabled

A mis-configuration detection method for a blade system. A blade system comprising a plurality of blade servers, a first Input/Output (I/O) module, a second I/O module, and a management module is provided, each blade server comprising a daughter card (DC). The management module determines a first candidate I/O type according to a DC type and activates the blade servers, or executes DC mis-configuration operations, determines a second candidate I/O type according to the DC type and activates the first I/O module, or executes an I/O mis-configuration operation, and activates the second I/O module according to the type of the first I/O module, or executes I/O mis-configuration operations.

Growing community of inventors

Tao Yuan Shien, Taiwan

Li-Tsung Chen