Performance Evaluation of the ATLAS IBL Calibration

The Insertable-B-Layer has recently been commissioned at the ATLAS Experiment, adding 12~million channels to the existing Pixel Detector. The front-end chips (FE-I4) are connected to newly designed readout hardware situated in a VME crate. In order to take data under uniform conditions, one needs to periodically tune the detector in short breaks between data-taking sessions to accommodate for radiation damage and aging effects. Tuning involves a variety of components, ranging from high-level steering and analysis software (PixLib) running on commodity hardware, to embedded components situated inside the VME crate that feature only a minimal or no operating system at all. Understanding the interactions between these components is key in debugging and optimizing the tuning procedures to become more efficient. We therefore implement an instrumentation framework aimed at all major components. It features a uniform interface to the user and is able to take instrumentation data with microsecond-precision. A central server application is used to gather the instrumentation data of a tuning session. It processes the data and saves it into a SQLite database for later analysis.