2D Incompressible Navier-Stocks Data Dimension Description

[FengMingquan-sjtu]

Hi, I downloaded the "ns_incom_inhom_2d_512-0.h5" from the link and tried using the data in my experiments. The data consists of four items, whose names and dimensions are:

Key=force, shape=(4, 512, 512, 2)
Key=particles, shape=(4, 1000, 512, 512, 1)
Key=t, shape=(4, 1000)
Key=velocity, shape=(4, 1000, 512, 512, 2)

I have several questions about them:

1. what's the meaning of the first dimension of all data? E.g. t.shape=(4, 1000), I guess the '1000' means 1000 time steps, but the '4' is unclear.
2. what do the 'velocity' and 'particles' mean? I guess the 'velocity' means the average velocity of the fluid in the grid, and the 'particles' means the average number of particles in the grid. Is it correct?

Thank you!

[finger-monkey]

Hi, do you know how to draw pdebench_examples.PNG? This is the visualization picture shown by the author on the code homepage

[Hy23333]

Hi there, thanks for the giant efforts from the team.
I also share the same question with @FengMingquan-sjtu on the meaning of '4' in the shape of the data.
Meanwhile I have got some bit confused in the paper description of the sample number and the resolution of incompressible NS data. Here I quote the descriptions as:

For the incompressible Navier-Stokes equations, the file contains 1000 samples, with 256 × 256 spatial cells, 1000 time steps, and 2 variables.

And for compressible NS data is:

For the 2D compressible Navier-Stokes equations, the file contains 1000 samples, with 512 × 512 spatial cells, 21 time steps, and 4 variables.

But it seems 275 files (guess samples) uploaded [ns_incom_inhom_2d_512-274.h5](https://darus.uni-stuttgart.de/file.xhtml?fileId=166299&version=8.0), not sure they are just a part of 1000 samples or some samples are equally merged in a file... Also, resolution should be 256 × 256, but got 512 × 512?

It would be of much help if the seemingly conflicts get explained.

Many thanks.