Li Benchmarks

reduce_sum greenvalidity pass

Tier 1 · wall_time · N=100_000_000 · OS linux

passpass — perf claims allowed when ratio is green

Source: latest.csv:perf_present

Wall-clock row exists only — no explicit correctness signal yet.

Ratio vs best competitor: 1.001 (100.1% of julia speed) (table SOTA: li — Li leads; charts use julia = 1.0)
Best competitor value: 0.4287 s

No analytical verify rows in ingest yet. Run tier-1 with --verify so latest.csv exports verify_ulps / verify_within_1ulp.

Problem size: N=100_000_000 (family reduce_sum)
Category: micro
Pillar: numerics
Package: lic
Li / catalog oracle (mean ± σ): 0.4287 ± 0.014701 s (n=20) / 0.475 ± 0.038125 s (n=20) (cpp)
Ratio vs catalog oracle: 0.9025×
Best competitor: li (0.4287 s)
Li relative speed vs SOTA: 1.001 (1.0 = julia speed)
Validity: validity pass
Threshold: 1.2×
Source: lic/benchmarks/workloads/tier1_micro/reduce_sum

PH ids: PH-5b

Compare oracle: cpp

Relative speed vs best competitor (julia) — SOTA = 1.0, higher is better. Absolute s values are in the table below.

Language comparison
Lang	Mean ± σ	Runs	Unit	Variant	OS
li	0.4287 ± 0.014701	20	s	release	linux
cpp	0.475 ± 0.038125	20	s	release	linux
rust	0.438 ± 0.019584	20	s	release	linux
julia	0.4292 ± 0.017753	20	s	release	linux
Metric: wall_time (value = mean of timed runs)