Understanding which memory locations are actually shared across threads is fundamental for maintaining, debugging,and testing multithreaded software.However, this information is difficult to obtain in a developer-friendly form: static analyses often report conservative candidate sets, while race detectors typically expose pair-oriented bug reports rather than a compact view of the program’s sharing structure. This paper presents DtTsa,an open-source synchronization-aware dynamic thread-sharing analysis tool for C/C++ programs. DtTsa combines LLVM-based instrumentation, runtime tracing of memory and synchronization events, and offline aggregation to produce source-level sharing reports from concrete executions.Rather than exposing raw traces, DtTsa summarizes shared access sites with their functions, access kinds, read-write interactions, and observed synchronization contexts. An empirical evaluation on three benchmark suites shows that DtTsa produces concise and stable execution-grounded reports, covers many race-relevant access sites, and incurs practical overhead. The artifact is available at https://doi.org/10.5281/zenodo.20438303, and the video is available at https://youtu.be/r3uho1OcJ6U.