Hi,
what is the status of ROOT <-> HDF5 interface?
Thanks.
Regards. Valeriy
Please read tips for efficient and successful posting and posting code
ROOT Version: Not Provided
Platform: Not Provided
Compiler: Not Provided
Hi,
what is the status of ROOT <-> HDF5 interface?
Thanks.
Regards. Valeriy
Please read tips for efficient and successful posting and posting code
ROOT Version: Not Provided
Platform: Not Provided
Compiler: Not Provided
I am not sure about that. Maybe @axel will know more.
There are no plans regarding HDF5 from our side. HDF5 provides its own reading API; you can then write data into TTree without problems. The other way around might also work but what for - HDF5 is not adequate for HEP’s data structures. For non-HEP data, see above
Hi Axel,
now I’m working in JINR (SPD detector at future NICA collider).
I’m back to particle physics after >15 years absence.
So, I missed a lot of things which have been happened in ROOT.
Because of triggerless the data rate in our experiment is huge and compatible with LHC experiment.
We desperately need the massive parallel data reading/writing like it’s done in HDF5 (check the doc)…
Do you know if it is possible the same with modern ROOT (RDataFrame?)?
Thanks.
Regards. Valeriy
FTWar!
Hi Axel,
I found very interesting talk from Philippe dated 2018
“HEP-CCE: Scalable IO for Energy and Intensity Frontier Experiments”
HDF5’s growing popularity among Machine Learning communities has brought it some attention. A recent update enabled parallel IO operations on HPCs (using MPI)
The discussion focused on the need for parallel IO capabilities inside ROOT and how those might perform compared to HDF5. Does the community need only one? The general consensus indicated support inside ROOT for HDF5 would be beneficial, particularly for experiments to provide custom outputs for their
analyzers. Partial implementations exist with various level of support and could be an opportunity for a funded project to consolidate and productionize into a complete implementation.
I wonder what is the current status of this activity?
Any benchmarks "modern ROOT I/O (RNtuple, RDataFrame) vs PHDF5?
Thanks.
Regards. Valeriy
++
FTW!
From Rene I found out “The ROOT project is run by Axel Naumann”
Congrats!
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.