Build 3075 - VETR V2

 

S24 Borderlands Style :)

S24 Experimental Build Shadows, Shaders, and the Occasional Fireball

This build is technically optional, but it does fix a control‑XML bug and folds in several deeper changes that were overdue for a test flight. It’s not a headline release — more of a “quietly ambitious” one — but it pushes the client forward in ways I’m intrigued by.

The shadow system in this build comes with some elements from the next‑gen visual‑flair branch. It’s experimental, and includes a few KV‑specific alterations. If it behaves, excellent. If it misbehaves, at least it does so with enthusiasm.

I’ve reworked the Windlight settings logic. LL’s development branch improved things, but it still drifted slightly out of spec in edge cases. This build tightens that behaviour so lighting transitions and environmental overrides act more predictably.

The animation overrider core has been updated to match the Firestorm implementation.

VFX Overhaul — The Truly Experimental Part

This is where the fun begins.

The VFX system has been rebuilt to run on OpenCL compute, offloading the heavy lifting to the underused portion of your GPU — the same region typically tapped by DLSS or AI routines. On top of that, the logic now batches operations and uses OpenGL interop to sample render textures directly. No more CPU–GPU round‑robin.

The result:

• Around 50% better performance

• Frametimes scale with the number of layered effects

• Far less overhead

• A lot more headroom for future experiments

It’s experimental, yes, but it’s the kind of experiment that occasionally produces something brilliant.

Big Drops or Rapid Fire?

I’m still gauging whether people prefer:

• Larger, polished milestone releases

• Or the “I built this at 2AM, try it if you dare” experimental builds

Part of why I keep every major S24 release on SourceForge is because some setups simply can’t run anything past certain release points. That’s fine. I make no apologies for it — the TPV landscape is broad, and my approach isn’t meant for mass‑market appeal. Plenty of other clients serve that purpose.

At the same time, I’m not trying to be so niche that the viewer becomes a single‑use tool. The whole point of this project is to explore the “what ifs” while staying ahead of the official codebase.

Sometimes that means explosive crashes. Sometimes it means absurd stability. Both outcomes are part of the charm.

This client is a playground for speculative ideas, GPU experiments, and the occasional accidental masterpiece. Whether you’re here for the stability, the chaos, or the curiosity, I appreciate you.

Much love, KL