I need to create a compilation video of all the times my kids complain about load times in games for when people say “performance doesn’t really matter.” They also told me, “Dad, if you ever make a game it will be fast, right?” Hell yes it will!

Any recommended resources for learning OpenGL?

I still get excited about new CPU designs but I’ve realized that software design is what really needs to be improved. We already have obscene abundance when it comes to compute power. Currently learning OpenGL so I can write my own UI for our tool.

I have found this article to be gold. Embracing handles instead of pointers has saved me from many foot-guns. Almost every time I use a pointer, I regret it and revert to handles. floooh.github.io/2018/06/17/…

I’m at the point where I believe Premature Abstraction is a far worse and pernicious blight than Premature Optimization.

I have to say, using @rfleury RadDBG with Odin and Sublime Text is just a dream. What an incredible tool!

New update for our tool, ReliaSim, adds outlining, which I think is a playful touch. What you are seeing is me editing parameters and the entire 1-year simulation being re-run in the background as I make changes. I then ask it to run 100 replications, which is 100 years, and it completes in 6 seconds. Other tools take 12 hours.

Here's a peek at our upcoming tool, ReliaSim. It's a Discrete Rate Simulation tool for high-volume manufacturing. This is a 1-year simulation running in 815ms (~110K simulation steps). Industry standard tools take 15 minutes. @odinlang gave us the freedom to optimize the performance. Simulations can run in parallel, so I can simulate 100 years in 6 seconds on my 12-core machine. That would take days with other tools. We are now focused on making a beautiful UI, but the performance already speaks for itself. At the end, you see me replaying the events. I can instantly jump to any moment in time and review the state of the system.

I think @cmuratori saying, "Functions, beautiful at any size," has to be one of my favorite new quotes.

You know you're working for some technical leadership when you're discussing taking a computation from 160 CPU cycles down to 30, and they get excited about it 🤣

Currently, I am down the rabbit hole of interpolation and approximation. I find it funny that I have spent far more time learning Linear Algebra and Calculus after college than I did in it 🤣

I’ve now written a Discrete Rate Simulation Engine in both .NET and Odin and believe that anyone who says managed languages can be just as fast as low level ones must not be taking advantage of the power control of memory layout gives you.

After working in Odin exclusively for over a year now, I now understand why C devs eschewed “modern” languages with GC. Once you learn how to manage memory, it just seems like a trivial problem. A GC is an extreme solution to a problem that’s easily managed with a few Arenas.

Working with @odinlang has made me allergic to GC. Once you see how easy manual memory management can be, you never want to go back. I'm also seeing an over 10x performance uplift moving away from a managed runtime. You get addicted to providing lightning-fast UX.