On 06/02/25 12:36, Sebastian Andrzej Siewior wrote: > On 2025-02-06 12:24:42 [-0300], Rafael Folco wrote: >>> so core 3 has min 33ns and max 129ns while 1us on average? >> Yes, because avg rounds up. >> There are some other potential fixes coming like: >> - stop rounding up avg >> - splitting bucket 001 (us) which is bigger (0 - 1.99999) >> - adding the max bucket to the end (do not limit to 32us) >> >> ... but these are out of scope for this change. > > I'm not totally against floating point numbers here but you should add > the bullet numbers to your patch description. Also you might lose > precision over time if you deal with floating point. I will add more context to the patch description. > > However what you just showed me looks like the statistics part is > working right. I mean max lower than average. At least it is above min. This patch just reduces this distortion with Avg=1.000 with Max=0.153. Still not ideal though, but better than Avg=1.000 and Max=0. The idea was to propose another change for Avg, like Avg=0.087 instead. I can include it here if that's the case. Thanks, Folco > >> Folco > > Sebastian >
