Conflicts:
	src/hydro/Gadget2/hydro_iact.h
	src/vector.h

Conflicts:
	src/kernel_hydro.h

Use intrinsics for vector operations as arithmetic overloading is not supported by the Intel compiler yet for AVX-512 vectors.

Conflicts:
	src/hydro/Gadget2/hydro_iact.h

Use intrinsics for vector operations as arithmetic overloading is not supported by the Intel compiler yet. Updated masking operations with new macros so that AVX-512 masking is supported.

Conflicts:
	src/kernel_hydro.h

Conflicts:
	src/vector.h

Use intrinsics for vector operations as arithmetic overloading is not supported by the Intel compiler yet.

Revert "Add checks for particles that are at same position, this isn't allowed, and update the cell h_max to that of the particles after the initial ghost tasks have ran. This is needed before regridding if the h values of the particles in the IC are set to be smaller than reality (ends up looking like rebuilding failed as the particles are too far outside their cells)."

This reverts commit 3d12f382.

Add checks for particles that are at same position, this isn't allowed, and update the cell h_max to that of the particles after the initial ghost tasks have ran. This is needed before regridding if the h values of the particles in the IC are set to be smaller than reality (ends up looking like rebuilding failed as the particles are too far outside their cells).

Use the basename for the HDF5 files as these are relative to the container file and could include directory information (we assume XMF file is written into same directory as the snapshots).

Updated the default values of space_subsize_pair space_subsize_self to the optimal values found in a parameter search using EAGLE_25 on 16 cosma-5 cores.

Removed the now unused space_maxcount global variable. Documented the new YAML parameters in the example parameter file.

Also use errno to get the detailed error about the failure

This would have revealed the last MPI bug a but quicker