Core Scientific, Inc.

Core Scientific Operating Company

Lawrence Kom

Thomas Middleton Rutledge Fuller

Thomas Fuller

Devon Baldwin

Kamil Lazarowich

Jacob Taylor

Richard Dunne

System and methods for managing large numbers of network-enabled devices such as computing devices in a facility such as a data center are disclosed. A management application selects and instructs one of the devices to enter a low power state. The power draw from outlets in a power distribution unit supplying the devices are monitored, and an outlet that is detected as dropping below a predetermined threshold of power usage is then mapped to the selected device. This process is repeated for other devices. This permits the power outlet and device associate to be automatically verified to ensure accurate device location identification and power-based billing for the devices in a facility.

Identifying computing devices in a data center

A method of computer operation may include (a) providing a first computer and a second computer; (b) operating the second computer; (c) modifying settings of the second computer; (d) monitoring, via the first computer, changes to a plurality of operational parameters of the second computer after the settings are modified to identify a plurality of covariant parameters of the plurality of operational parameters; and/or (e) controlling, via the first computer, operation of the second computers according to the identified plurality of covariant parameters to increase computational efficiency of the second computer.

Method and system of optimizing the performance of a second computer by a first computer

Systems and methods for managing computing devices in a data center based on collected environmental data are disclosed. Temperature data from the computing devices is periodically gathered along with environmental data from sensors inside and outside the data center. The collected data is used to create a correlation model between the outside environmental data, the inside environmental data, and the device environmental status. Weather data for the geographic location of the data center may also be incorporated into the correlation model, which may be used to predict future heat spikes and trigger preventative measures such as adjusting device fans, operating frequencies, reducing or shifting workloads, adjusting vents, and engaging cooling units.

Environment-based tuning for computing devices

A method of operating computers may include connecting a computer with a proxy computer; obtaining, via the proxy computer, a list of pools compatible with the computer; selecting a pool from the list of pools; obtaining a header for the selected pool; providing, via the proxy computer, the header to the computer; authenticating the computer with a pool computer; performing compute tasks, via the computer, corresponding to the selected pool; and/or submitting output information from the compute tasks to the pool computer.

System and method for operating computers

An adjustable cooling duct for improved facility (e.g., data center) cooling is disclosed. In one embodiment, the cooling duct comprises a cylindrical flexible membrane (e.g., fabric) having a first annular end and a second annular end, and a cincture mechanism connected around the perimeter of the cylindrical flexible membrane between the first annular end and the second annular end. The cincture mechanism may be motorized and that may be controlled by a management computer based on pressure sensor readings. The cincture mechanisms may be automated for coordinated control of multiple ducts in parallel by the management computer based on environmental sensor readings to improve facility cooling.

Adjustable ducts for cooling computing devices

Systems and methods for managing airflow for cooling computing devices (e.g., in a data center) are disclosed. To prevent damage to computing devices after power outages in low temperatures, a subset of the computing devices may be preheated, started, and operated in an exhaust air recirculation mode to raise the temperatures of the other computing devices. Preheating may for example use heaters or warm air temporarily diverted from external sources such as nearby buildings. Thermal mass (e.g., phase change material) may be positioned near the computing devices to capture heat that is later released in the event of a dramatic temperature drop such a cold weather power outage.

Thermal management for container-based data centers

A facility (e.g., data center) for more efficient cooling of computing devices is disclosed. In one embodiment, the data center comprises a set of external walls, a ceiling connected to the external walls and having circular exhaust openings, and one or more pods positioned within the external walls and below the ceiling. A set of flexible tubes are used to connect the pod exhaust openings to the ceiling openings. Actuating closures may be used (e.g., at one end of each of the tubes) to control airflow, and air pressure sensors may be placed near or inside one or more of the tubes. A controller may be configured to receive air pressure data from the air pressure sensors and adjust the actuating closures to maintain a desired air pressure level. In some examples, the actuating closure may have a normally closed position when power is not applied.

Data center cooling

A cooling device for improving the cooling of computing devices in a data center is disclosed. The cooling device comprises an air deflector that is configured to be attached to a hot aisle air barrier at an angle to direct computing device exhaust air toward the nearest hot aisle exhaust vent. The device may comprise a mounting mechanism for fixing the air deflector in position, such as eyelets or clips that attach to the computing device's exhaust fan grille. The air deflector may be rotatable so that the angle between the hot aisle air barrier and the air deflector may be adjusted. The cooling device may further comprise a visual air flow indicator to assist in positioning of the air deflector and airflow monitoring.

Air deflector for cooling computing devices

A rack and system for cooling computing devices such as in a data center is disclosed. In one embodiment, the rack supports a plurality of computing devices and comprises a plurality of planar annular sector or trapezoidal shelves, each having one or more positions for holding computing devices. A number of vertical supports may hold the shelves in a vertically spaced arrangement, and an air barrier may be affixed to the rack and have openings that permit exhaust air to flow from the computing devices through the air barrier into a hot aisle. The rack may be configured to be connected to other similar racks to form a vertical annular cylinder, which may be tapered, and the computing devices may be installed to discharge exhaust at an angle relative to a centerline of the vertical annular cylinder to improve airflow.

Rack for cooling computing devices in a cylindrical configuration

Systems and methods for cooling large numbers of computing devices in a data center are disclosed. The computing devices have cooling fans that can have their fan speed set by management instructions. The devices are mounted in racks in a two-dimensional array and connected via network switches with ports corresponding to their location in the racks. The computer devices are oriented so that their cooling fans all exhaust waste heat to one side of the rack. Instructions are sent to the computing devices to set one or more attributes such as fan speed, operating frequency and voltage according to one ore more patterns. The patterns can be linear or nonlinear and can be shifted, rotated, or changed over time periodically or based on temperature or other performance measurements from the computing devices.

System and method for cooling computing devices

A method for measuring airflow for a plurality of computing devices may include characterizing the fan performance of a selected computing device of the plurality of computing devices to provide characterized RPM information. Each of the computing devices may include a cooling fan configured to exhaust heat into a hot aisle. Each cooling fan may include a rotor and a tachometer. The method may also include connecting the computing devices via a network, distributing computing workloads to the plurality of computing devices, and/or performing work on the computing workloads on the plurality of computing devices. Additionally, the method may include reading RPM information from the selected computing device's fan's tachometer and/or comparing the read RPM information with the characterized RPM information to determine a backpressure value for the selected computing device. The power to the selected computing device's fan may be reduced prior to reading the RPM information.

Measuring airflow for computing devices

Systems and methods for cooling computing devices in a data center and measuring airflow are disclosed. The computing devices have cooling fans that can have their fan speed set by management instructions. The devices are mounted in racks between cold aisles and hot aisles and are connected via network switches. The computing devices are oriented so that their cooling fans operate to exhaust waste heat to one side of the rack. Measurement instructions are sent to selected computing devices to cause them to go into a lower power state that reduces or stops computational work and reduces electrical power to the cooling fan so that the fan does not spin based on the electromotive force and can instead spin in an opposite direction based on ambient airflow caused by the difference in air pressure from the hot aisle to the cold aisle.

Growing community of inventors

Redmond, WA, United States of America

Lawrence Kom