Aren't the goals of t-digest a little bit different?
T-digest seeks to have a bounded size and an error proportional to q*(1-q), hence it gives up quantile accuracy in the middle of the distribution when under load. This algorithm seems to provide total bounded error without small but unbounded size.
6 comments
[ 2.9 ms ] story [ 25.8 ms ] thread---
BenchmarkMetrics/Add/streadway/quantile-8 5000000 358 ns/op
BenchmarkMetrics/Add/bmizerany/perks/quantile-8 5000000 291 ns/op
BenchmarkMetrics/Add/dgrisky/go-gk-8 5000000 363 ns/op
BenchmarkMetrics/Add/influxdata/tdigest-8 5000000 250 ns/op
BenchmarkMetrics/Add/axiom/quantiles-8 10000000 208 ns/op
---
I think its the fastest for insertion
Querying need finalization of state then its just pretty fast but will comment once i can get the API into a friendlier state :D
T-digest seeks to have a bounded size and an error proportional to q*(1-q), hence it gives up quantile accuracy in the middle of the distribution when under load. This algorithm seems to provide total bounded error without small but unbounded size.