package irmin-bench

You can search for identifiers within the package.

in-package search v0.2.0

Irmin benchmarking suite

Install

Dune Dependency

github.com Readme Changelog ISC License Edit opam file Versions (3)

Authors

T

Thomas Gazagnaire

Maintainers

T

thomas@gazagnaire.org

Sources

irmin-3.11.0.tbz

sha256=09996fbcc2c43e117a9bd8e9028c635e81cccb264d5e02d425ab8b06bbacdbdb

sha512=0391a6bf7b94a1edd50a3a8df9e58961739fa78d7d689d61f56bc87144483bad2ee539df595c33d9d52c29b3458da5dddf3a73b5eb85e49c4667c26d2cd46be1

doc/irmin-bench.traces/Irmin_traces/Trace_stat_summary_conf/index.html

Module `Irmin_traces.Trace_stat_summary_conf`Source

Summary configuration.

This file is NOT meant to be used from Tezos, as opposed to some other "trace_*" files.

Sourceval histo_bin_count : int

Sourceval curves_sample_count : int

Sourceval moving_average_half_life_ratio : float

Parameter to control the width of the moving average window.

This width is expressed as a fraction of the width of the final plot images, (i.e. the number of played blocks). This implies that the amount of smoothing remains visually the same, no matter curves_sample_count and the number of blocks in the stat trace.

Justification of the current value:

The 2nd commit of the full replay trace is a massive one, it contains ~1000x more operations than an average one, it takes ~10 half lives to fully get rid of it from the EMAs. With moving_average_half_life_ratio = 1. /. 80., that 2nd commit will only poluate 1 / 8 of the width of the smoothed curves.

Sourceval moving_average_relevance_threshold : float

See Exponential_moving_average