Programming with less effort in C++...
summary
Claims that using C++11/14 features measurably cuts coding effort by computing Halstead-style metrics with a Clang-based toolchain and a GitHub metric-calculator. Examples: a naive matrix multiply shows effort ~9,000, a small cache-aware tweak ~12,000, while hand-written SIMD jumps to ~700,000 effort for only ~4× speedup; OpenMP yields large gains on many cores (up to ~40× on 48 cores). Tooling uses Clang tooling and Halstead effort metric, is available on GitHub, and works but needs careful AST handling.
AI-written summary · about
chapters (25)
- 0:00Introduction and motivation
- 2:57Overview of software metrics
- 3:40Understanding Halstead metrics
- 3:59Metrics in C++ context
- 6:20Code analysis methodology
- 6:52Practical case study
- 10:02Defining operators and operands
- 10:52Parallel code examples
- 16:40Performance vs effort results
- 17:04C++ language features and effort
- 23:50C++11 features analysis
- 24:00Advanced static analysis
- 30:20Tooling demonstration
- 30:46Future research directions
- 32:26Feedback and Q&A
- 32:50Language comparisons
- 33:01Development environment impact
- 35:25Metric drawbacks
- 36:16Real-world applicability
- 36:36Final wrap-up
- 40:20Complexity and effort
- 42:45Metric verification
- 45:50Language evolution
- 57:06Reflecting on complexity
- 58:03End of recording
related talks
49mWriting better code with help from Qt and the compiler
56mWrite Fast Code Like a Native
44mFastware
01h 21mEmpirically Measuring, & Reducing, C++’s Accidental Complexity
01h 00mFastware
01h 12mComparison of C++ Performance Optimization Techniques for C++ Programmers
23mBeating the Benchmarks: Co-evolving the ISA and Development Tools
42mAdventures in Efficiency and Performance