Developed a backend throughput cost modeling capability for the xdslproject/xdsl repository, focusing on integrating LLVM-MCA to enable data-driven performance analysis. Designed and implemented a cost model interface that estimates block throughput by leveraging external tools, specifically LLVM-MCA, and provided a concrete MCABlockThroughputCostModel to support backend decision making. The work included comprehensive tests to validate the accuracy of performance analysis for assembly-like code blocks, laying the groundwork for more precise cost estimation in scheduling. Utilized Python and applied skills in backend development, compiler design, and performance analysis to deliver a robust foundation for future optimization efforts within the project.