after working with C-based root for more then 10 years I have come to love pyroot recently. However the relative slowness compared to a compiled macros still annoys me sometimes. Searching for solutions I stumbled upon the pypyroot project which sounds very interesting. However, I migrated my code to root 6 by now and I was not able to find something on pypyroot in root 6 apart from “With root 6 everything will be much easier”.
Is there any new development that I missed or is pypyroot for root 6 simply not ready yet?
well, there you put your finger on the chicken & egg problem. Getting ATLAS to allow me to even support PyROOT has always been a struggle, but at least there I’ve been able to point to a significant group of ATLAS users. That said, with the current crunch, I’m down to “rainy Sunday afternoons” only, and those, too, are severely lacking here in CA, given that we are experiencing an epic drought.
But for Python, current PyROOT users all find performance acceptable. If they don’t, they move on, just like you. As such, there is no need or group of people that I can point at (although multi-threading may become a strong motivation at some far away point in the future), and so nothing gets done on that front.
Of course, there are (again) strong predictions that this will be an El Nino year, bringing torrential rains to CA, maybe even on Sundays. But that’s no way to plan a project.
Plenty … we had a bumper year for rain! I also had a GSoC Summer Student last year and ROOT6 has been stripped and stuffed into a pip to allow cppyy on top of Cling in PyPy. See here:
However, as said, allowing me to work on this was a struggle and that has remained. So, I’ve thrown in the towel some time ago. Mgmt insisted there is simply no use of python in HEP let alone high-performance python. It is what it is and these days I’m only peripherally involved with HEP and instead work in HPC systems research.