Abstract: A graphics processing unit (GPU) server having a GPU host head with one or more host graphics processing units (GPUs). The GPU server further has a GPU system with a plurality of system GPUs that are separate from the host GPUs, and that are configured to rapidly accelerate creation of images for output to a display device. The GPU server also has a mounting assembly that integrates the GPU host head and the GPU system into a single GPU server unit. The GPU host head is independently movable relative to the GPU system.

Abstract: A graphics processing unit (GPU) server having a GPU host head with one or more host graphics processing units (GPUs). The GPU server further has a GPU system with a plurality of system GPUs that are separate from the host GPUs, and that are configured to rapidly accelerate creation of images for output to a display device. The GPU server also has a mounting assembly that integrates the GPU host head and the GPU system into a single GPU server unit. The GPU host head is independently movable relative to the GPU system.

Abstract: A graphics processing unit (GPU) server having a GPU host head with one or more host graphics processing units (GPUs). The GPU server further has a GPU system with a plurality of system GPUs that are separate from the host GPUs, and that are configured to rapidly accelerate creation of images for output to a display device. The GPU server also has a mounting assembly that integrates the GPU host head and the GPU system into a single GPU server unit. The GPU host head is independently movable relative to the GPU system.

Abstract: A system includes a first printed circuit board (PCB), a temperature sensor, a switching circuit provided on the first PCB, and a controller. The temperature sensor is configured to measure temperature of at least an area of the first PCB. The controller is configured to trigger the switching circuit to turn off power to the first PCB, based at least in part on the temperature sensor detecting a temperature above a temperature threshold. The system is able to disrupt power much faster than conventional methods of power protection which may have a blind spot to certain areas of the first PCB, since these methods rely on power disruption when a maximum power is sensed.

Abstract: Described herein is a server that includes a server chassis having a front window for providing access to an interior receptacle of the server chassis. A docking station is located in the interior receptacle. The server also includes a PCIe module coupled to the docking station by a coupling element. The PCIe module protrudes externally from the interior receptacle through the front window in the docked position. The PCIe module can be moved to a plurality of undocked positions while remaining coupled to the docking station.

Abstract: Described herein is a server that includes a server chassis having a front window for providing access to an interior receptacle of the server chassis. A docking station is located in the interior receptacle. The server also includes a PCIe module coupled to the docking station by a coupling element. The PCIe module protrudes externally from the interior receptacle through the front window in the docked position. The PCIe module can be moved to a plurality of undocked positions while remaining coupled to the docking station.

Abstract: An air baffle for optimizing thermal performance of memory components provided in a chassis is disclosed. The air baffle has a body and one or more venting units provided on the body. The body is configured to be removably coupled to the chassis. The body covers the memory components when coupled to the chassis. The one or more venting units direct air flowing through the chassis. Each of the venting units includes vent openings and a corresponding number of adjustable venting plates. The vent openings are each aligned with the memory components when the body is coupled to the chassis. Each of the adjustable venting plates have an open position or a closed position. A respective venting plate of the adjustable venting plates in the open position allows airflow through a respective vent opening of the vent openings. The respective venting plate of the adjustable venting plates in the closed position blocks airflow through the respective vent opening of the vent openings.

Abstract: An air baffle for optimizing thermal performance of memory components provided in a chassis is disclosed. The air baffle has a body and one or more venting units provided on the body. The body is configured to be removably coupled to the chassis. The body covers the memory components when coupled to the chassis. The one or more venting units direct air flowing through the chassis. Each of the venting units includes vent openings and a corresponding number of adjustable venting plates. The vent openings are each aligned with the memory components when the body is coupled to the chassis. Each of the adjustable venting plates have an open position or a closed position. A respective venting plate of the adjustable venting plates in the open position allows airflow through a respective vent opening of the vent openings. The respective venting plate of the adjustable venting plates in the closed position blocks airflow through the respective vent opening of the vent openings.

Abstract: A graphics processing unit (GPU) server having a GPU host head with one or more host graphics processing units (GPUs). The GPU server further has a GPU system with a plurality of system GPUs that are separate from the host GPUs, and that are configured to rapidly accelerate creation of images for output to a display device. The GPU server also has a mounting assembly that integrates the GPU host head and the GPU system into a single GPU server unit. The GPU host head is independently movable relative to the GPU system.

Abstract: The present disclosure provides for a partition in a chassis, which can transition between at least two positions to receive more than one configuration of modules. The partition can include a first end and a second end, each of which has resting positions adjacent to the chassis body. The partition can further include a middle portion, which is in a raised position above the floor of the chassis body, but can bend flexibly towards the floor. In a first configuration of the chassis, a single, larger module can slide over the partition. In a second configuration of the chassis, two smaller modules can slide into adjacent positions in the chassis body, separated by the partition.

Abstract: The present disclosure is directed to a removable cable arm bracket for attachment to an inner rail of a computer system. The cable arm bracket includes a spring sheet that can be biased for securing and removing the cable arm bracket from the inner rail. The spring sheet secures a projection from the inner rail within an aperture of the cable arm bracket to secure the cable arm bracket to the inner rail. Biasing of the spring sheet allows the projection to be removed from the aperture.

Abstract: The present disclosure provides a chassis assembly which can include a chassis body, at least one sled, and at least one rear bracket. The chassis body can have at least one pair of rail brackets along opposing sides of the chassis body. The at least one sled can hold computer components and be sized according to the computer components. The at least one rear bracket can couple with the at least one sled.

Abstract: A server includes a housing, a sled, a cable sleeve mounted to the sled, and an obstruction in the housing. The housing defines a sled channel and a cable channel. The sled can be inserted into the sled channel. A cable is positioned in the cable sleeve. The obstruction is aligned with the cable channel. Prior to the sled being inserted into the housing, the cable sleeve is in a first position aligned with the obstruction and the cable channel. As the sled is inserted into the housing, contact between the cable sleeve and the obstruction moves the cable sleeve to a second position. In the second position, the cable sleeve is not aligned with the obstruction and is positioned in the sled channel. The cable sleeve is biased towards the first position. As the cable sleeve travels past the obstruction, the cable sleeve moves back to the first position.

Abstract: A system and method for dynamic reconfiguration of at least one peripheral bus switch of a system includes a management controller that detects whether a server system is connected to each peripheral bus slot of the system. The management controller selects a peripheral bus switch topology for the at least one peripheral bus switch, based on the detecting. The management controller sets each port of the at least one peripheral bus switch to either an upstream port configuration or a downstream port configuration, based on the peripheral bus switch topology.

Abstract: A server device is provided. The server device includes a chassis with a containing space, first and second openings. The first and second openings are located at two opposite ends of the containing space. The chassis includes a motherboard module in the containing space; a power-supply module disposed in the containing space, electrically connected to the motherboard module, and capable of plugging in and out from the chassis via the first opening, wherein the power-supply module comprises a first power port located in the first opening; a power port connector comprising a second power port located in the second opening, and configured to selectively connect to the first power port located in the first opening; and a storage array module, electrically connected to the motherboard module and the power-supply module.

Abstract: A sled is provided to facilitate connection of the electrical components housed therein to a backplane located in a chassis. The sled includes a housing with a front portion and a rear portion connected by a base portion. The base portion includes a plurality of electronic unit modules. The sled also includes a connector laterally slidable with respect to the front portion of the housing. The connector is slidably configured to connect the electronic unit modules to a backplane located in the server rack.

Abstract: A sled is provided to facilitate connection of the electrical components housed therein to a backplane located in a chassis. The sled includes a housing with a front portion and a rear portion connected by a base portion. The base portion includes a plurality of electronic unit modules. The sled also includes a connector laterally slidable with respect to the front portion of the housing. The connector is slidably configured to connect the electronic unit modules to a backplane located in the server rack.

Abstract: A server device is provided. The server device includes a chassis with a containing space, first and second openings. The first and second openings are located at two opposite ends of the containing space. The chassis includes a motherboard module in the containing space; a power-supply module disposed in the containing space, electrically connected to the motherboard module, and capable of plugging in and out from the chassis via the first opening, wherein the power-supply module comprises a first power port located in the first opening; a power port connector comprising a second power port located in the second opening, and configured to selectively connect to the first power port located in the first opening; and a storage array module, electrically connected to the motherboard module and the power-supply module.

Abstract: A system and method for dynamic reconfiguration of at least one peripheral bus switch of a system includes a management controller that detects whether a server system is connected to each peripheral bus slot of the system. The management controller selects a peripheral bus switch topology for the at least one peripheral bus switch, based on the detecting. The management controller sets each port of the at least one peripheral bus switch to either an upstream port configuration or a downstream port configuration, based on the peripheral bus switch topology.

Abstract: A blade server system with a management bus and method for managing the same. The blade server system includes a connection board and a management module. The connection board is used for modular interconnection, including communication paths for conducting signals including a management bus signal group and a first bus signal group. The management module is used for management of the blade server system using signals including the management bus signal group and the first bus signal group. If a module is detected, the management module selects the detected module through the management bus and acquires module configuration information of the detected module through the first bus signal group. Distribution of power from a power source to the blade server system is determined according to system configuration information including the module configuration information of the module so that the power source is prevented from being overloaded.