{"ID":2897334,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2507.04647","arxiv_id":"2507.04647","title":"RAPTOR: Practical Numerical Profiling of Scientific Applications","abstract":"The proliferation of low-precision units in modern high-performance architectures increasingly burdens domain scientists. Historically, the choice in HPC was easy: can we get away with 32 bit floating-point operations and lower bandwidth requirements, or is FP64 necessary? Driven by Artificial Intelligence, vendors introduce novel low-precision units for vector and tensor operations, and FP64 capabilities stagnate or are reduced. This forces scientists to re-evaluate their codes, but a trivial search-and-replace approach to go from FP64 to FP16 will not suffice. We introduce RAPTOR: a numerical profiling tool to guide scientists in their search for code regions where precision lowering is feasible. Using LLVM, we transparently replace high-precision computations using low-precision units, or emulate a user-defined precision. RAPTOR is a novel, feature-rich approach -- with focus on ease of use -- to change, profile, and reason about numerical requirements and instabilities, which we demonstrate with four real-world multi-physics Flash-X applications.","short_abstract":"The proliferation of low-precision units in modern high-performance architectures increasingly burdens domain scientists. Historically, the choice in HPC was easy: can we get away with 32 bit floating-point operations and lower bandwidth requirements, or is FP64 necessary? Driven by Artificial Intelligence, vendors int...","url_abs":"https://arxiv.org/abs/2507.04647","url_pdf":"https://arxiv.org/pdf/2507.04647v2","authors":"[\"Faveo Hoerold\",\"Ivan R. Ivanov\",\"Akash Dhruv\",\"William S. Moses\",\"Anshu Dubey\",\"Mohamed Wahib\",\"Jens Domke\"]","published":"2025-07-07T04:00:18Z","proceeding":"cs.DC","tasks":"[\"cs.DC\",\"math.NA\"]","methods":"[]","has_code":false}