Developed and integrated a new FP8 GEMM client example for the StreamHPC/rocm-libraries repository, focusing on demonstrating matrix multiplication using the FP8 data type. The work involved implementing the core functionality in C++ and configuring the build process with CMake, ensuring seamless integration into the existing codebase. The example provided end-to-end capabilities, from building and running the client to collecting and reporting key performance metrics such as TFlops and GB/s. This addition enabled streamlined benchmarking and evaluation of FP8 performance within the ROCm stack, supporting high-performance GPU computing workflows and facilitating further development in FP8 tensor operations.