Rethinking the Way We Do Templates in C++

on YouTube ↗01h 30m

" summary

Library implements a physical-units system that keeps debuggable, human-readable types by deriving user-facing strong types from heavy template instantiations via a downcast CRTP trick, replaces class-template-heavy ratios with non-type template parameters (string and ratio values) and uses concepts instead of SFINAE to constrain interfaces. Benchmarks with metabench show NTTP/value-based designs cut compile time from tens of seconds to a few seconds in hotspot tests and concepts compile faster than complex SFINAE on modern compilers.

AI-written summary · about

" chapters (12)

1. 0:00Start
2. 1:44The User Experience
3. 2:32What the Physical Units Library Does
4. 27:17Donecast Implementation
5. 32:09Non-Type Template Parameters
6. 36:16Template Argument Equivalent
7. 43:08What Are Concepts Class Function
8. 55:16Benefits of Using Concepts
9. 1:02:07Parameter Kinds
10. 1:08:36Performance
11. 1:18:03Performance Outcome
12. 1:20:52What Is the Performance of Concepts

" related talks

• 01h 37mRethinking the Way We Do Templates in C++ even moreMateusz Pusz · C++ on Sea 2020
• 01h 01mRethinking the Way We Do Templates in C++Mateusz Pusz · CppCon 2019
• 01h 10mThe C++ Type System Is Your FriendHubert Matthews · ACCU 2017
• 57mIntroduction to C++ Template MetaprogrammingSasha Goldshtein · NDC Oslo 2016
• 47mReducing Template Compilation Overhead, Using C++11, 14, 17, and 20.Jorg Brown · CppCon 2019
• 01h 00mModern Template Metaprogramming: A Compendium, Part IWalter E. Brown · CppCon 2014
• 01h 21mWhat Has C++20 Ever Done For Templates?Hendrik Niemeyer · ACCU 2022
• 01h 23mThe Power of C++ Templates With mp-units: Lessons Learned & a New Library DesignMateusz Pusz · C++ on Sea 2023