Machine for satellite data analysis

We asked you to discuss a machine for a wide range of satellite image analysis using Sentinel-2 satellite data.
The purpose is to smoothly perform analysis at the 100 million km square level using an analysis program created by myself in Python.
As a result of conducting a trial analysis on a Mac with an M1 chip, it took 412 hours to complete, so we have heard a request to reduce this to about 24 hours.

The proposed configuration is a machine with 3 CPU specifications equipped with 2rd generation Xeon scalable.
Relative performance comparison is difficult because the architecture is different between the M1 chip of ARM architecture and the x86 CPU used in most machines, but this configuration is budgeted as a computing machine within the budget. It is a calculation machine that can be used internally and to some extent for general purposes.
Assuming that you can fully demonstrate the CPU performance, it has a performance advantage over the comparison Mac.

If you are using your own program, you need to check the optimization status on the Python code side.At least for single-threaded calculations, it is possible that the results will not be much different from those of the Mac to be compared, so it may be necessary to make processing changes such as code parallelization along with the introduction of the machine.We apologize for the inconvenience, but please be aware in advance.

【Main Specifications】