Worked on the ROCm/hipTensor repository to enhance performance benchmarking by refining the configuration of benchmark test ranges. Focused on tuning the 'Ranges' parameter, the work reduced upper bounds across multiple scenarios, resulting in more targeted and efficient performance testing. This adjustment improved the clarity of performance signals and established a stronger baseline for future optimizations. The implementation was delivered through a dedicated commit, emphasizing careful configuration management and benchmarking practices. Utilized C++ and YAML to modify test parameters and manage configurations, applying performance testing expertise to streamline benchmarking workflows and support ongoing code quality improvements within the project.