I noticed the following unusual behavior when using gpt in SNAP 6.0 Beta.
I developed a graph using SNAP 5 for Sentinel-1 processing. I could ran it without problems via gpt batch processing and it delivered the desired results (intensity Gamma0 values). Then I migrated to SNAP 6.0 by doing a clean new installation.
When I'm running the same graph via gpt under the new version, suddenly two output products per band (Gamma0_VH & Gamma0_VH_2/ Gamma0_VV & Gamma0_VV_3) are created. The biggest problem is that both products differ in pixel values.
I already tried different input products, but the error stays the same.
Here is a link to a dropbox folder containing the processed files (out of frustration everything including the .dim file, so I have only the .img and .hdf files left) and also the xml file and the corresponding bat file. The scene that was processed in this case is S1B_IW_GRDH_1SDV_20170711T170628_20170711T170653_006443_00B539_29A5
Maybe someone is experiencing similar issues?
Right now, I switched back to Version 5, since I'm in great need of the batch processing function via gpt.
Anyway, great new features (the AOI monitoring tool is really neat!)