Dennis Gustafsson – Parallelizing the physics solver – BSC 2025

Channel:Better Software Conference · Length:1:07:08 · Published:2025-07-18 · Views:78098

Watch on YouTube

Summary

Dennis Gustafsson, creator of Teardown, presents how he parallelized the physics solver for an upcoming game, followed by a Q&A. The talk was given at the Better Software Conference 2025.

Why HN recommends it

HN points to it as a practical example of solving a tricky multithreading problem, including using a spinlock to speed up a real-time physics solver.

The following is an interesting talk where the author used a spinlock to speed up a real-time physics solver significantly. — HN, 2026

Software performance & optimization · Game programming & engines

Subtitles (excerpt)

All right, welcome back everyone and uh thank you Casey for the previous talk. Amazing detail and amazing amount of research you must have done on that. This talk is hopefully going to be a little bit shorter so we get lunch. Um so Sam reached out to me a couple of weeks ago and asked if I wanted to present something and I came to think about this work I did about half a year ago on optimizing and parallelizing the constraint solver for our upcoming game. And as it turned out, uh, this talk isn't going to be that much about the constraint solver or how to paralyze it. It actually is going to be more about the low latency thread pools and job system that I had to write to make it all run efficiently on modern hardware. So, but I hope you you'll enjoy it. Anyway, you're going to see some physics movies at least. Uh, so just some background on me. I'm Dennis Gustafson and I am the original creator of the game Tearown, which is a Voxil uh sandbox heist game that came out in 2020. I was also the only programmer on the project up until the initial release and now we're a bit more. Um it it features a um custom u software rate tracing voxal rate tracing engine which is pretty special and also a a

On Hacker News

Not about multiplayer but still interesting. Dennis Gustafsson – Parallelizing the physics solver – BSC 2025 [VIDEO] — Hikikomori, 2026-03
As always: use standard libraries, profile, then write your own if the data indicates it's necessary. The following is an interesting talk where the author used a spinlock to speed up a real-time physics solver significantly. Dennis Gustafsson – Parallelizing the physics solver – BSC 2025 [VIDEO] — fasterik, 2026-01
From my understanding this talk describes how he implemented a solution for a similar problem [VIDEO] I'm saying if you're not writing multi-threaded code everyday, use a library. It can use atomics/locks but you shouldn't use it directly. If the library is designed well it'd be impossible to deadlock. — levodelellis, 2025-11

Related videos