v1.6
- In the pipeline, all merging-related steps are now consistently performed with 32 bit float precision
- Strong reduction of peak system memory usage and cumulative memory allocation
- Slight reduction of processing times
- Several bug fixes and minor improvements