SG14 (the GameDev & low latency ISO C++ working group)
summary
SG14 pushes concrete C++ fixes for low‑latency games: proposals include a standardized ring buffer (paper 59) and cache‑friendly flat_map/flat_set, while uninitialized memory algorithms already land in the standard. Claims and tools are specific: rendering vs simulation budgets (33 ms frame, ~100M cycles at 3 GHz, ~5k cycles per soldier), custom allocators and EASTL succeed where std::allocator and exceptions cost (a past recode causes a ~12% perf drop). Work targets SIMD, execution policies and heterogeneous APIs like bulk_invoke.
AI-written summary · about
chapters (16)
- 0:00Intro
- 0:12Summary
- 6:02Some tricks
- 7:43Further constraints
- 9:43Don't pay for what you don't use
- 10:23Exception costs
- 14:52RTTI
- 17:38The Standard Library
- 23:37Memory constraints
- 29:00Function calls
- 31:19Library extensions
- 32:05Ring
- 37:44Flat map and set
- 39:02Uninitialised memory algorithms
- 41:42Parallelism
- 46:42Heterogeneous computing
related talks
01h 01mThings Happening in SG14…
01h 04mConcurrency TS2: Improved C++ Concurrency and Lock-free Programming
01h 16mPaying for Lunch: C++ in the ManyCore Age
01h 31mISO C++ Standards Committee Panel Discussion
01h 05mThe Most Important Optimizations to Apply in Your C++ Programs
55mC++ 20 Innovations: High-Performance Cross-Platform Architecture in C++
55mC++ in the World of Embedded Systems
54mC++26: an overview