Hi everyone,
I would appreciate some guidance on optimizing the SNAPHU Phase unwrapping process. I am trying to process a whole Sentinel-1 interferogram and the phase unwrapping process is the bottleneck at this stage. I am processing my data on a Windows Server with 256 Gb of Ram and 28 cores. In the Snaphu-Export step I am specifying MCF as initial method with number of tiles rows and columns as 30 and 10 respectively. This means the tiles are between 1000 and 3000 pixels across. The row and column overlap have been set to 100. I have set the number of processors to 16.
The phase unwrapping step takes over 11 hours to complete, but checking the Performance metrics revealed that only 3% of the CPU and 4% of the memory is being used by Snaphu, at least in the “optimizer for secondary network” phase.
Could anyone provide some guidance on how this Snaphu process can be optimized if possible? I would value any guidance in this regard.
Yours sincerely,
Jeanine