runtime features / performance

✓ 2 passing ○ 16 planned

Lessons

1 ○ Planned
array by reference — OWED: large arrays pass by reference, not by value.
2 ✓ Passing
001 profile metatype
3 ○ Planned
002 profile release — OWED: profile meta-type captures timing under --profile=release.
4 ✓ Passing
003 profiler loop
5 ○ Planned
006 rings vs channels — OWED: MPMC ring producer/consumer compiles as idiomatic Koru concurrency.
6 ○ Planned
007 simple loop — OWED: label loops (#/@) compile and run the 10M-count performance shape.
7 ○ Planned
008 event taps overhead — Event tap code injection not yet implemented
8 ○ Planned
009 multi consumer async — OWED: multi-consumer async (semantic events vs raw Zig threads) compiles.
9 ○ Planned
010 threading library — Feature: Threading library support
10 ○ Planned
011 loop optimization basic — OWED: checker-event + label-loop lowers to a native loop.
11 ○ Planned
012 loop optimization nested — Feature: Nested loop optimization
12 ○ Planned
013 loop optimization work — Feature: Loop optimization with work distribution
13 ○ Planned
generic ring type — OWED: MpmcRing in an event shape (generic ring type).
14 ○ Planned
generic ring actual — OWED: real MpmcRing(T, N) in an event shape.
15 ○ Planned
ring with taps — OWED: ring-buffer consumer with taps.
16 ○ Planned
ring with when filter — OWED: ring-buffer consumer with when-clause filtering.
17 ○ Planned
when at callsite — OWED: when-clause sits at the call site (pinned emit pattern).
18 ○ Planned
ring with multitaps — OWED: ring buffer with multiple taps on the same consumer.
19 ⏭ Skipped
dead code elimination — Phase 4: Dead code elimination not yet implemented
20 ⏭ Skipped
handler caching — Phase 5: Handler caching and specialization not yet implemented