Abstract: A jig is provided for a graphics processing unit (GPU) system. The jig includes a front component operable to couple with the GPU system. The front component including a top panel operable to enclose a GPU board of the GPU system, and a plurality of guide tubes extending from the top panel. The guide tubes are operable to be aligned with a plurality of screws to couple the GPU board with a GPU tray. The guide tubes are operable to receive and guide a tool to gain access to the plurality of screws.

Abstract: A jig includes a lifting component and a sliding component. The lifting component is operable to receive a graphics processing unit (GPU) tray. The lifting component includes one or more grooves operable to receive the GPU tray. A sliding component is operable to receive a GPU board when the sliding component is in a reception configuration. The sliding component is operable to transition from the reception configuration to a deployment configuration where the GPU board is received in the GPU tray.

Abstract: A method and apparatus for determining spatial characteristics of three-dimensional objects is described. In an exemplary embodiment, the device receives a point cloud representation of a three-dimensional surface structure of a plurality of objects. The device may further generate a set of bins to represent the three-dimensional surface structure based on the point cloud representation, each bin corresponding to a spatial occupancy related to the point cloud representation, each bin including a respective type indicating a spatial relationship of the surface structures and a corresponding spatial occupancy of the bin. In addition, the device may encode the set of bins using a convolutional neural network. The device may further determine a classification for the spatial characteristic of the surface structures based on the convolutional neural network with the encoded set of bins.

Abstract: Systems and methods are provided for determining a confidence level on contact pairs identified through traditional contact detection algorithms for an assembly of bodies. A plurality of testing tolerances are identified. For each of the identified testing tolerances, a determination is made as to whether a distance between any two faces of the bodies in the assembly is less than that identified testing tolerance and when this happens a count of positive detections is increased. A ratio of the count of positive detections to the total number of times the detections are run tolerances is determined (called repeatability of a face pair), where the confidence level is based on the determined repeatability ratio.