Filter performance plot with number of complete events

[srimanob]

As additional point of view, I think we can filter number of complete events in performance plots, it should be great because we can see median/mean of specific group, i.e.

• Signal: events < 200,000
• Large
• Very large: MinBias or main backgroud

[franzoni]

hi Phat, can we discuss this ?
We'd like to understand better what you have in mined.
J & G

[srimanob]

@franzoni
It's my simple idea. We should not use performance plots in pMp to show the time it spends in "Submit-to-Done" step if request sizes are very different. So we need to weight, or specific the size of sample which I like more.

For example, I want to focus on signal samples and the size of requests is less than 200,000 events.