Latency in this paper is measured under a fixed protocol designed to remove caching and warm-start confounds: Apple M4 Max CPU-only hardware, caches disabled, a warm-up phase followed by $200$ measured queries, and three repeats of the full measurement. The reported $\pm$ values are the mean within-repeat per-query standard deviation taken from the cited microbench JSONs, so the dispersion captures run-to-run variation within a repeat rather than across-repeat variance. Reporting latency under this CPU-only, cache-disabled, multi-repeat protocol makes the timings directly comparable across the paper's systems and benchmarks.