Major bug fix version: DO NOT USE previous versions!

Thanks to Hyunbae Park, normalization error in previous versions was found. This version just fixes this major bug but nothing else from v1.1.5.

• Before, raylrnd was used as raylrnd(1, ...).

• Now, raylrnd is used as raylrnd(1/sqrt(2), ...).

• Action needed: do NOT use random seeds named "subgaussseed*.matbin" that were generated by older versions.

• The confusion in old versions came from a typo in Coles & Lucchin: Eq. 14.8.3 in 1st Edition (equivalently Eq. 13.7.3 in the 2nd Edition) is wrongfully written. $\delta_{k}^{2}$ in two denominators should be replaced by $\delta_{k}^{2}$, and thus the change above.

A bit more memory optimization

** Any version 1.x will be mainly about memory optimization.

• Now about 32xNcell_p^3 byte memory is saved.

• New flag "memory_save" should be set to "true" in params_patch.m to use memory-saving feature.

• There still exists bottleneck when calling interpn though; consequently you still need about ~ 140 Gbyte memory available for Ncell_p~1024. Set up a big swap memory if the system memory falls short. Next version is to resolve this memory bottleneck.

update - alleviate memory problem / octave "unifrnd" problem

This is a minor update, but resolves two issues:

1. Memory limit for high-resolution initial condition had existed. The function "interp2", which is used when generating particle data, consumes too much memory and may crash "bccomics". Now this problem is avoided by calling "interp2" iteratively. (Thanks to Hyunbae Park for experimeting)

2. "unifrnd" function to generate uniform random phases (0~2pi) is separated from the native Octave core from Octave v4.4. So test its functionality and if the Octave core is lacking the functionality, let bccomics use the provided "unifrnd.m".

First release

No bugs seem to exist in v0.9, and after some tests the first official release here.

Minor updates from pre-release version (v0.9):

• bccomics.m generates 'eptherm' in addition, the thermal specific energy of baryon particles, with SPH simulators (e.g. Gadget) in mind.
• dependence of Matlab on Toolboxes alleviated by providing working functions under mfiles_for_matlab.
• analysis snippet (analyze_plot_IC.m) added.