Crystallography

Since many years 2D CCD detectors have been used in crystallographic experiments. Finer angular slicing of the data often results in higher signal to noise levels, especially in the case of pixel array detectors. Thus, data sets may contain more than 1800 frames with 32 MB each. For the time of the exposure, data rates in excess of 150 MB/s have to be dealt with. ‘Luckily’ sample mounting and alignment, even if carried out by a robot, takes place on the scale of several minutes. Regarding expedient evaluation of the measured data, (protein-) crystallography is already in decent shape since several programs are available for integrating and calculating first data quality indicators within minutes, provided that sufficient computing power is available. In this case the most difficult task, namely the implementation of the evaluation software for parallel computing (e.g. SMP), has already been carried out by experts from the community. However, data file formats are mostly proprietary with limited and mostly no automatic means for saving the experiment metadata that need to be recorded, e.g. for a synchrotron radiation experiment. Also, the handling of the huge amount of data is a big issue for this community. Therefore, the structures to be realized by HDRI will certainly streamline data handling processes. Crystallography experiments, therefore, are an ideal test case for the first implementation of the envisaged data flow structures.