Few CPUs used for SAR grey level co-occurrence matrix texture analisys in Linux

I’m using the SNAP Texture Features Analysis (GLCM) tool for Radarsat-2 VHR , and I have noticed that such tool for Windows 10 OS take advantage of all the available 4 cores while in our Linux machine does not (only 2 or 3 of 16 cores). As this texture analysis is very high CPU demanding it will be great if I can processes the SAR images using the 16 cores (32 threads ) of our Linux workstation with 64Gb mem RAM.
Is there any way to improve this issue?
However, other tools as the Speckle Filter or the Terrain Correction are well cpu parallel performed also in Linux.
Thanks in advance,