Each weak lensing map is constructed from three simulation boxes covering different redshift ranges, with different box sizes and resolutions. The priors over which cosmological parameters are samples are:
Fix cosmological parameters:
Stored on HPSS at: /home/b/biwei/HSC/fastpm_box1536
This box uses the FastPM quasi N-body code with
Stored on HPSS at: /home/b/biwei/HSC/fastpm_box704
This box uses the FastPM code with
Stored on HPSS at: /home/b/biwei/HSC/MP-Gadget
This simulation uses MP-Gadget and has
The layout of the FastPM files is as so:
fastpm_box1536/
├── <Cosmology index>
│ ├── Om_<Omega_m>_S8_<S_8>_<scale_factor>
│ │ ├── 1
│ │ │ ├── Position
│ │ │ ├── Velocity # (only recorded for the first step)
│ │ │ ├── ...
│ │ ├── Header
│ │ │ ├── ...
The output structure of the MP-Gadget simulations is explained in the documentation here.
There are 101 cosmologies in total. Each particle has an associated position and velocity, and is captured at 19 timessteps so is described by 6x19=114 Float32 numbers. The total number of particles in each box differs, and so the total size in each box is:
- 1536 Mpc/h box:
$3 \times 32~{\rm Bits} \times (1536^3~{\rm particles}) * (101~{\rm cosmologies}) * (16~{\rm Snapshots}) = 83.4~{\rm TB}$ - 704 Mpc/h box:
$3 \times 32~{\rm Bits} \times (2816^3~{\rm particles}) * (101~{\rm cosmologies}) * (19~{\rm Snapshots}) = 1.2~{\rm TB}$ - 320 Mpc/h box:
$3 \times 32~{\rm Bits} \times (960^3~{\rm particles}) * (101~{\rm cosmologies}) = 0.1~{\rm TB}$