You might also like "Fluid Engine Development" if this stuff is your cup of tea....

_qb0p · on Jan 30, 2022

..and if it's really REALLY your cup of tea, join my research lab and do a PhD. :-)

https://cs.uwaterloo.ca/~c2batty/

</shameless plug>

yesenadam · on Jan 31, 2022

No need for shame!

"He did not wait for students to come to him; he made inquiry for promising youthful collectors, and when he heard of one, he wrote, inviting and urging him to come. Thus there is hardly one now of the American naturalists of my generation whom Agassiz did not train." - William James, Louis Agassiz, 1896.

https://archive.org/details/memoriesandstud02jamegoog/page/n...

kenrose · on Jan 31, 2022

+1

For context, Chris (batty above) was one of Robert’s best grad students.

If fluid simulation is your thing, Chris is one of the best people in the world to do a PhD with.

He’s also a very nice person. :)

nextaccountic · on Jan 31, 2022

I want cross-platform deterministic fluid simulations (at least in platforms with IEEE 2008 float semantics). What I mean is, I want to run the same software on different architectures and platforms and get the same exact answer. Is that feasible?

Also, are some techniques more amenable to cross-platform determinism while others are more challenging?

Note that Rapier can do this for mechanical simulations aimed at gamedev https://rapier.rs/docs/user_guides/rust/determinism/

Moreover: can I add threads or use SIMD or run on GPUs and still be deterministic? (Rapier can't do that: it runs on CPU, and on the cross-platform deterministic mode it disables threads and SIMD)

I'm not sure if those are considered worthwhile research lines, but this is what I care most.