MintPy memory requirements for Sentinel-1 Burst time-series processing

[Valentina Macchiarulo]

Hello everyone,

I am currently processing a Sentinel-1 Burst InSAR stack (generated using ASF HyP3 On-Demand products) with MintPy and would appreciate some advice regarding hardware requirements and memory optimisation.

During the `load_data` and `network_inversion` steps, I have observed very high memory usage. On our Linux virtual machine, MintPy consumed almost all available resources (~63 GB RAM and 8 GB swap), causing the system to become unresponsive.

I have reviewed the documentation but have not found any guidance on recommended hardware specifications for MintPy processing. I understand that resource requirements depend on factors such as stack size, spatial extent, and processing settings, but I would be interested to hear about the experience of other users.

Specifically:

* What hardware specifications (RAM, CPU, storage) do you typically use for MintPy processing?
* What would you consider a reasonable amount of RAM for a small- to medium-sized Sentinel-1 time-series analysis?
* Are there particular MintPy configuration options that can significantly reduce memory consumption?

Any recommendations, examples, or lessons learned would be greatly appreciated.

Thank you,
Valentina

[Vittorio Bisaglia]

Hi Valentina, I haven't had any problems with the isce2+mintpy workflow (I don't use ASF HyP3) when setting AOI of a few dozen kilometers. I had the RAM overuse issue with the isce2+StaMPS workflow during the processing phase in MathLab with very large AOI.

Perhaps the problem also stems from using a virtual machine...

Vittorio

--
You received this message because you are subscribed to the Google Groups "MintPy" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mintpy+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/mintpy/894ece26-bf5f-4b24-91fe-80997c1df1e6n%40googlegroups.com.

--

Studio  Geologia Applicata - Geomatica

Geologo Vittorio Bisaglia

Albo dei Geologi del Veneto n° 468

Via Alessio, 19 – 35121 Padova (Italy)

C.F. BSG VTR 65D13 G224M - P.IVA 02700530286

mobile: +39 368 7217303 

E-mail: vittorio...@gmail.com

P.E.C.: vittorio...@pec.epap.it

[Asmit Roy Burman]

Hello Valentina,

I haven't used MintPy for almost a year, so I'm not aware of any major changes that may have been introduced recently. However, I can say that in the past I processed Sentinel-1 time series with MintPy on systems with only 16 GB of RAM (Windows, WSL Linux, and Linux) without any issues. You could try adjusting the following parameter in the configuration file:

########## computing resource configuration
mintpy.compute.maxMemory = auto #[float > 0.0], auto for 4, max memory to allocate in GB

Perhaps setting an explicit memory limit here may help resolve the problem.

I hope this suggestion is useful. I occasionally follow the research coming from your group, and I appreciate the work you are doing.

Good luck with your processing, and I hope you find a solution soon.

Best regards,
Roy