scale: 1000 records × 10000 tokens (10.0M tokens), pool capacity 10.0M, target ratio 0.3, gc keep 0.5
write: 1000 records / 10.0M tokens in 1.0s
db after write: 1.2 MiB, records dir 38.5 MiB
== reads ==
facets: median 0.44ms p95 0.45ms
list 20: median 0.59ms p95 0.62ms
catalog: median 1.05ms p95 1.23ms
search title term "topic 7": median 1.56ms p95 1.86ms
search ascii "alpha beta": median 2.09ms p95 2.54ms
search CJK "记忆": median 0.19ms p95 0.20ms
grep body "xxxxxxxx": median 5.23ms p95 5.31ms
get record: median 0.04ms p95 0.04ms
== compression ==
compressCold: moved 700 records, freed 7.00M tokens in 1.07s
hot 3.00M / cold 7.00M tokens; cold files 26.9 MiB; db 3.6 MiB
search after cold (card fields): median 6.72ms p95 6.95ms
catalog after cold: median 0.90ms p95 0.92ms
== cold gc ==
gcCold: removed 350 records, freed 3.50M cold tokens in 0.24s
db after gc: 3.6 MiB