zp-unbuffered-io — unbuffered-io

ClaimCopy Anchor Link

Writer empty buffer (&{{}}) vs 4KiB buffer to /dev/null under Zig 0.17.0 -Ofast on this Linux machine shows a clear before/after wall-clock gap.

How measuredCopy Anchor Link

• Harness: packages/zperf/examples/bench/evidence_fill.zig (rule zp-unbuffered-io cases)
• Build: zig build-exe -Ofast · optimize_mode=fast
• Timing: Io.Clock awake; warm=2; median of 9 rounds
• before: writer_empty_buf
• after: writer_4kib_buf
• ops=50000
• Raw: zp-unbuffered-io.jsonl

Runtime (this evidence file)Copy Anchor Link

Lint mappingCopy Anchor Link

zperf-lint zp-unbuffered-io / unbuffered-io. Default severity in rules.zig is off (the CLI does not report the rule); see evidence-policy for keep/escalate. Evidence exists and agrees directionally.

VerdictCopy Anchor Link

Agree ~49.0× wall; matches guide p19 (~50×).

LimitsCopy Anchor Link

One-shot writes / readVec-writeVec / tests skipped by lint.