Contributed a targeted feature enhancement to the tensorflow/tensorflow repository by implementing support for U32 and U64 data types in the SPMD partitioner’s gather index functionality. This work, developed in C++ with a focus on data structure design and parallel computing, expanded the partitioner’s ability to handle 32- and 64-bit indices directly, reducing the need for type casting in indexing operations. The update improved data-path flexibility and streamlined pipelines for workloads dependent on varied index sizes. Emphasis was placed on robust, maintainable code, laying the foundation for future extensibility in type support within distributed computation workflows. No bugs were addressed.