188th ROOT Parallelism, Performance and Programming Model Meeting

Thursday 25 Sept 2025, 16:00 → 17:30 Europe/Zurich

32/S-C22 (CERN)

Marta Czurylo (CERN), Vincenzo Eduardo Padulano (CERN)

Notes: 
histv7: Current status and directions

slide 1:
Jack: leaf of the dependency tree? -- Jonas: histogram package doesn't depend on anything else
Jacob -- this should be a complete package 

histogram as a data structure (no drawing, no fitting -- not inside this at least, as a ROOT histogram package - yes, it should all be included)

slide 4:
Josh: can underlflow and overflow bins be disabled? -- Jonas: yes 
Josh: Will you be able to specify the weight like in boost? -- yes, we will have the same thing but it will always be the last argument. 
Danilo: filling in with an array of values? - not enabled - what about automatic pythonisation? - to be seen
Stephan: vectorization while computing vector indexes -- Jonas: is this really relevant? show me microbenchmarks for this (maybe Josh has an example for this). 

slide 6:
Jacob: inconsistency in errors within ROOT should be avoided 
invalid argument - std version should be kept - maybe this should be changed in RNTuple

slide 8:
Jacob: RHistBufferedFiller - wasn't it going to be discarded? - after talking with stephan, we decided to keep it but in a very limited scope: 1D sequential filling only, Stephan: maybe we should call differently? RHistAutoAxis or similar?
Jakc: What about Fill Random? -- not part of the data structure itself. Jack: for prototyping? or as clean as possible for the real use cases. 

16:00 → 17:00 histv7: Current Status and Directions

Speaker: Jonas Hahnfeld (CERN & Goethe University Frankfurt)

20250925_Hahnfeld_histv7.pdf