Abstract: Embodiments of the present disclosure include systems and methods to facilitate top down delete or un-synchronization on delete of a synchronization client to a cloud-based platform. The disclosed technology innovates the mechanism the synchronization client processes local deletes so that instead of synchronizing the delete of each individual item across to the cloud-based platform, it can accumulate the individual item deletes into a top-level folder delete and that top-level folder delete can be what is communicated across to the cloud-based platform. Embodiments of the present disclosure also include systems and methods for depicting item synchronization between a local and cloud based platform/service (e.g. the Box service). In some embodiments the system can identify items that cannot be synchronized between the local and cloud based platform/service (e.g., cloud based collaboration or storage platform/service).

Abstract: Systems and methods for managing organizational structures are provided. The method of managing organizational structures includes visually representing an organizational structure with movable and definable nodes provided in a hierarchical format, which is generated in a user interface of a computing system. The method further includes mapping, within a database, relationships of the movable and definable nodes and information therein based on visual connections within the hierarchical format. The method further includes updating the database with newly mapped relationships between the movable and definable nodes as the movable and definable nodes are moved into different locations with respect to each other within the hierarchical format.

Abstract: The present disclosure describes methods, systems, and computer program products for optimizing single-value database read operations. One computer-implemented method includes initializing, by a computer, a single value cache associated with a created database table where a column of the database table contains a single or unique data value, verifying that the created database table is in a useable state, performing an operation on the database table, and retrieving the data value from the single value cache without locking the column containing the single or unique data value.

Abstract: Provided is a high-performance implementation of the k-means clustering algorithm on a graphics processing unit (GPU), which leverages a set of GPU kernels with complimentary strengths for datasets of various dimensions and for different numbers of clusters. The concepts of non-dominated GPU kernels and efficient strategies to select high-throughput kernels that match the arguments of the clustering problem with the underlying GPU hardware for maximum speedup are provided.