... + CTAD + NTTP = 🤩

on YouTube ↗47m

" summary

Combines variadic templates, CTAD, and NTTP to build a compile-time declarative model description used to auto-generate JSON and to validate runtime tensor extents for thermodynamic simulations. Implements domain intervals and extent-constraint resolution (matrix-multiply example) in roughly 50 lines for domains, folds and pack expansions for serialization, and a constraint-table algorithm that rejects mismatched sizes; tests pass and code compiles on clang‑18, gcc‑14 and MSVC. Notes: large NTTP arrays can inflate compile time unless passed by reference.

AI-written summary · about

" chapters (8)

1. 0:00Introduction and agenda
2. 1:17Variadic templates and parameter packs
3. 8:17Fold expressions in C++17
4. 15:40Class Template Argument Deduction
5. 22:59Non-type template parameters
6. 31:29Real-world thermodynamic simulation
7. 40:49Results and limitations
8. 43:19Q&A session

" related talks

• 28mC2TACO: Lifting Tensor Code to TACOElizabeth Polgreen, Michael F. P. O'Boyle, Jackson Woodruff · GPCE 2023
• 28mFunctional Sparse Tensor CompilationAmir Shaikhha, Shideh Hashemian · FProPer 2024
• 15mVTC: DNN Compilation with Virtual Tensors for Data Movement EliminationXin Xu, Charith Mendis, Vikram S. Adve, Muyan Hu, Ahan Gupta, Jiachen Yuan, Vima Gupta, Taeksang Kim, Janardhan Kulkarni, Ofer Dekel · 20th USENIX Symposium on Operating Systems Design and Implementation (OSDI '26)
• 32mCompilation of Dynamic Sparse Tensor AlgebraSaman Amarasinghe, Stephen Chou · OOPSLA 2022
• 20mCompiling Structured Tensor AlgebraMahdi Ghorbani, Amir Shaikhha, Mathieu Huot, Shideh Hashemian · OOPSLA 2023
• 33mTeIL: a type-safe imperative Tensor Intermediate LanguageNorman A. Rink, Jeronimo Castrillon · ARRAY 2019
• 14mCompressed and Parallelized Structured Tensor AlgebraMahdi Ghorbani, Emilien Bauer, Tobias Grosser, Amir Shaikhha · OOPSLA 2025
• 13mAccelerating Complex-Valued Tensors With Torch.compileHameer Abbasi · PyTorch Conference Europe 2026