The roofline analysis model is a visually intuitive performance model used to understand hardware performance limitations as well as potential benefits of optimizations for science and engineering applications. Intel Advisor has provided a useful roofline analysis feature since its version 2017 update 2, but it has insufficient scalability as well as limited compatibility with other compilers and chip architectures. As an alternative, I have employed hardware performance counters for a scalable and flexible roofline pe rformance analysis for modern HPC applications. First, I present how to reliably measure a computational intensity for roofline analysis. Several numerical studies have been performed for validation via manually derived reference. Second, I discuss roofline analysis results of several HPC benchmarks and sparse linear algebra libraries on a Cray XE system (i.e., Blue Waters). In addition, I present an example of roofline-based performance projection for a future system.