Frequently asked questions

This document lists a number of questions that have been asked by the community. If you think you have a relevant question to add, please don't hesitate to contact docadmin and we will be happy to add it.

WILL BE EXPANDED - WORK IN PROGRESS

Task writing

Having multiple process functions with different subscriptions leads to code duplication which can be error prone. Can this be avoided?

Process functions are basic task construction elements that, by definition, tell the framework about data subscriptions. Therefore, to vary data subscriptions, multiple process functions have to be defined. However, the content of the process function itself can be a single template-based function or method so that the processing code doesn't necessarily need to be duplicated. For an example, please check cascadeanalysis.cxx or correlations.cxx.

Hyperloop

I haven't changed anything but sometimes my wagon test gives a memory warning and sometimes not, and also the estimated resources change somewhat.

We base the measurement of used memory and CPU on a 3 minute O2 execution with your task. Depending on how the different processes of O2 interact the memory consumption can slightly vary. If your task is just at the limit, this can result sometimes in a warning and sometimes not. Similarly, the estimated CPU is calculated from the short test and then scaled to the entire dataset. Already slight changes in execution time can have a big impact and therefore the numbers are not the same on each test. The best is to not operate close to the memory / CPU limit and so you do not get any warnings. You can inspect the history of all your tests in the wagon modal by clicking on the tab ‘test statistics'.

Why do different wagon tests on the same data process a different number of input files?

The wagon test runs for about 3 minutes and is then stopped. Therefore, if you have a fast wagon, you will process more input files than if you have a slow wagon.