Bringing C++ 17 Parallel Algorithms to a standard library near you
summary
Intel donates a near-complete C++17 parallel algorithms implementation and integration work adapts it into libstdc++/libc++ by cleaning macros/ADL, restructuring tests to CMake/LIT, and moving tests into shared suites; current backend depends on TBB but plan ships an experimental in GCC8 build and replaces the backend by GCC9–10; a 15,000-line patch is under review and license transfer to the LLVM Foundation is pending; an Executors-based backend prototype (PSTL + p0443/p1019) demonstrates feasibility for executor targets.
AI-written summary · about
chapters (27)
- 0:00Introduction
- 0:43Questions
- 1:03Parallel Algorithms
- 1:34Execution Policies
- 3:28Implementation Status
- 3:39GCC
- 3:50Parallel Mode
- 4:18Red Hat
- 4:34Intel
- 6:08uglification macros
- 6:38C reserves
- 7:23Argument Dependent Lookup
- 8:46Marshalls
- 9:04Microsoft
- 10:09Testing
- 12:32Intel donation
- 13:34GCC integration
- 14:37unofficial bits
- 15:50Intel API
- 16:39SG1 proposal
- 19:10Less optimal implementation
- 21:13Timeline for integration
- 23:07Application responsibility
- 24:42Initializing PBV scheduler
- 26:28Who is this for
- 27:58Is this enough
- 29:13Heterogeneous computations
related talks
57mHow To Implement the C++ Standard Library - (Part 1 of 2)
54mC++17 Features (part 2 of 2)
59mHow to Implement the C++ Standard Library (Part 2): An Excursion into libc++
58mC++17 in Practice - The good and the ugly
01h 00mC++17, part 1: The Language Features
45mC++ is a Metacompiler
01h 01mMoving to C++20 for Fun, Beauty... and Results!
01h 02mLatest & Greatest in Visual Studio for C++ developers