News for hackers. Stuff that matters.

A comment on 200 Lines of Python beats $50M supercomputer – Navier-Stokes at Re=10⁸ [pdf].

Re:200 Lines of Python beats $50M supercomputer – Navier-Sto... by jcamlin on Friday November 28, 2025 @11:56AM (#46080382)
The secret: T³ periodic domain + Fourier orthogonality = the weak solution is just a curve in coefficient space. Not a field. A curve. Temporal lifting samples that curve densely where it bends hardest. Standard DNS crashes from CFL blowup. This doesn't - because we're not fighting the geometry, we're riding it. Results:

Taylor-Green Re=10⁵: BKM=37.1, stable through full vortex-stretching cascade Kolmogorov Re=10⁸: 0.07% dissipation error 128³ grid → 411³ effective resolution (spectral super-resolution from temporal oversampling) Hardware: 8GB laptop, no GPU

No artificial dissipation. No hyperviscosity. Unmodified Navier-Stokes unlike all other DNS. GitHub link to code and data in the paper (linked)

[ Parent | on HN ]
Re:200 Lines of Python beats $50M supercomputer – Navier-Sto... by bediger4000 on Friday November 28, 2025 @12:18PM (#46080614)
What's the significance of those particular Reynold's numbers? 10^8 seems high for incompressible flow, but maybe I misread something.
[ Parent | on HN ]
Re:200 Lines of Python beats $50M supercomputer – Navier-Sto... by doctorpangloss on Friday November 28, 2025 @12:15PM (#46080583)
> because we're not fighting the geometry, we're riding it

Don’t let ChatGPT write summaries for you (please edit the comment to be real)

[ Parent | on HN ]