You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, criterion generates very nice plots of the distribution of runtimes for all of our benchmarks. This is the right level of detail for honing in on a specific benchmark, e.g., the distribution of times received from FBT probes on rx.
A glance view of processing times in and between families of packets would be useful. Each set of results processed by criterion outputs a list of box-whisker values (tukey.json) and other summary statistics (estimates.json), which makes this data easily available. We're already pulling in plotters as a library via criterion, which may be a useful starting point (as handy as Gnuplot is under other circumstances).
The text was updated successfully, but these errors were encountered:
Currently, criterion generates very nice plots of the distribution of runtimes for all of our benchmarks. This is the right level of detail for honing in on a specific benchmark, e.g., the distribution of times received from FBT probes on rx.
A glance view of processing times in and between families of packets would be useful. Each set of results processed by criterion outputs a list of box-whisker values (
tukey.json
) and other summary statistics (estimates.json
), which makes this data easily available. We're already pulling in plotters as a library via criterion, which may be a useful starting point (as handy as Gnuplot is under other circumstances).The text was updated successfully, but these errors were encountered: