zp-page-allocator-small — page-allocator-small
Claim
page_allocator 64B alloc+free vs process GPA (smp/c in ReleaseFast) under Zig 0.17.0 -Ofast on this Linux machine shows a clear before/after wall-clock gap.
How measured
- Harness:
packages/zperf/examples/bench/evidence_fill.zig(rulezp-page-allocator-smallcases) - Build:
zig build-exe -Ofast· optimize_mode=fast - Timing:
Io.Clockawake; warm=2; median of 9 rounds - before:
page_allocator_64B - after:
process_gpa_64B - ops=20000
- Raw:
zp-page-allocator-small.jsonl
Runtime (this evidence file)
| Field | Value |
|---|---|
| When | 2026-10-21 Asia/Shanghai (JSONL ts is UTC) |
| Zig | 0.17.0 · fast |
| Median before | ~137.1 ms |
| Median after | ~0.4018 ms |
| ratio before/after | ~341× |
Lint mapping
zperf-lint zp-page-allocator-small / page-allocator-small. Default severity in rules.zig is warning; see evidence-policy for keep/escalate. Evidence exists and agrees directionally.
Verdict
Agree ~341× — page allocator pays syscall/page per small block.
Limits
Large mmap-sized buffers can still use page_allocator intentionally.