Data Volume
L5004320A/B runs
These are pure DMonly simulations carried out with GADGET4.

Full particle dumps for redshifts: About 2.8 TB per dump, yielding 28 TB for the 10 full dumps.

Group catalogues: Summed over the output dumps, we have ~2.0 x 10^10 subhalos and ~1.7 x 10^10 FOF halos, consuming 2.3 TB storage.

Progenitor and descendant information linking subhalos in time amounts to 1.1 TB.

Merger trees weigh in with 2.8 TB, and contain about 145.5 x 10^6 trees.

All snapshots with the previously mostbound particles make up in total 3.7 TB. The fraction of particles with this property grows monotonously in time, and reaches ~3.2 x 10^9 (4.0% of the total particle number) by z=0.

The lightcones contain a very significant amount of data, summing up to 392 TB in total.
Cone Number Lightcone Type Redshift Range Distance Particle Number Storage Cone 0 full sky z = 0  0.4 ~1090 Mpc/h 3.43 x 10^12 64.6 TB Cone 1 one octant z = 0  1.5 ~3050 Mpc/h 9.44 x 10^12 177.7 TB Cone 2 pencil beam, 10 deg^2 z = 0  5.0 ~5390 Mpc/h 1.02 x 10^12 19.2 TB Cone 3 disk, 15.0 Mpc/h thick z = 0  2.0 ~3600 Mpc/h 3.94 x 10^11 7.41 TB Cone 4 full sky, mostbound only z = 0  5.0 ~5390 Mpc/h 6.54 x 10^12 123.1 TB 
The weaklensing mass maps amount to 1.2 TB and use Nside = 12288.
Each of the two 4320^3 runs has hence created around 430 TB, with the full particle octant to z=1.5 and the full sky mostbound lightcone contributing 70% of this storage.