@pervognsen @xoofx You may like the presentation section on https://logicoma.io/squishy as well :)
Staff Software Engineer at Unity. Demoscene: emoon / TBL. Has way too many spare time projects.
@pervognsen @xoofx You may like the presentation section on https://logicoma.io/squishy as well :)
“jemalloc Postmortem” https://jasone.github.io/2025/06/12/jemalloc-postmortem/
@rovarma @sschoener 100% agree. It is *much* more important to think: "What is the user trying to active" instead of blindly add the requested feature(s).
@ne7 Just read it... very sad news :(
"How we decreased GitLab repo backup times from 48 hours to 41 minutes"
O(N²) strikes again. Double nested for loop.
OpenMPT 1.32.01.00 released: https://openmpt.org/openmpt-1-32-01-00-released #openmpt
@runevision Some people really like https://github.com/jj-vcs/jj
@sschoener I guess so yes :)
Metal Gear for the MegaDrive is out now. Go grab it and get you're retro stealth on!
@sschoener I actually had no idea this was a thing :D (I know about 'Pantamera', but not there being a new song each year)
@range_marten @h0ffman nice nice :)
RustWeek 2025 presentations will start today. There is a schedule over at https://rustweek.org/schedule/tuesday
Also given that Rust 1.0 was released 10 years ago it means I started to fiddle around with Rust 10 years ago :) (I started as soon it turned 1.0)
I think the first presentation "10 Years or Rust: Why?" Will be interesting.
“Mipmap selection in too much detail” https://pema.dev/2025/05/09/mipmaps-too-much-detail/
Anyone use Intel's 386 processor from 1985, the first 32-bit x86? To improve performance, it preloaded instructions into a 16-byte prefetch queue (highlighted in red). Why does such a small queue take up so much of the chip? Let's look under the microscope at its complicated circuitry... 1/N
@chainq Yeah, it's a really cool system and while I didn't plan out my layout it's good to have at least *some* idea how you want to structure stuff :)
@chainq If you have a 3D printer I can suggest you check this out :) https://handsonkatie.com/underware-2-0-the-made-to-measure-collection/
This now ends up being a move to vector register and a cmp instead of a load. Whatever this is faster depends on the code around it of course, but I still think it's a neat trick.