[Regression] Mobile Vibrant Visuals: Terrain stutter in 1.26

On Android devices (VIVO V2452A), Version 1.21.100.20 runs flawlessly, even while screen recording. However, upgrading to 1.26 and enabling Vibrant Visuals causes severe terrain stutter and complete engine deadlocks during surface exploration. This is not a device memory shortage, but a fatal thread synchronization failure in the RenderDragon pipeline when processing chunk generation. This is a critical regression caused by Vibrant Visuals itself, not a hardware limitation.

Steps to Reproduce:

1. Set simulation distance to 2 chunks, FOV to 30, enable Vibrant Visuals (all other effects disabled).

2. Load the world with Seed: -6778246788154565662

3. Test two distinct approaches along the route: Start (0, 107, 0) -> Midpoint (396, 72, 101) -> End (818, 63, 188, Village).

Observed Result:

Chunk Generation Deadlock (The Core Issue):

· Village Area (Entity/Block Entity Load): Attempting to enter the village area completely deadlocks the engine. The profiling chart reveals 0ms frame time and blank gaps, because the engine completely stopped submitting frames entirely. The minimum FPS drops to 0.0, freezing the game for several seconds.

· Visual Proof: This deadlock persists even after reducing resolution to 720x1600, lowering render distance to 2 chunks, and disabling all other visual effects. The game still freezes at 0 FPS despite minimal GPU load.

In-Hand Item Rendering Stutter (Secondary Issue):

· First-person: Picking up items causes severe latency and a complete freeze (0 FPS).

· Third-person: Picking up items only causes minor frame drops.

· Note: Enabling "Hide Hand" eliminates this pickup stutter, confirming a RenderDragon pipeline bug with in-hand rendering, though it does NOT fix the chunk generation deadlock.

Expected Result:

Smooth chunk generation, stable thread scheduling, and no rendering freezes when picking up items. The game should handle sudden chunk loading without deadlocking at 0 FPS, matching the performance of Version 1.21.100.20.

Additional Context & Request for Assistance:

Engine Scheduling Analysis (The Core Proof):

A direct comparison on the same device reveals the engine's thread scheduling has severely regressed, independent of hardware memory limits:

· Correction on Metrics: Please focus on Frame Time (Load) and Min FPS, not Average FPS.
Comparing average FPS is meaningless here because the testing conditions were completely different (1.21 was tested at higher graphics settings, 1.26 at minimum 720p/2-chunk settings).

The core evidence is the load spike and minimum frame rate:

· 1.21.100.20 (Baseline): Maximum Frame Time is only 178ms. Minimum FPS is 49. It handles the load and recovers instantly.
· 1.26.52 Hotfix (Regression): It suffers a massive 731ms Frame Time spike at the River, and a hard 0.0 FPS deadlock at the Village. The Render Thread stalls (Render: 107ms, waitForPreviousFrame: 103ms).

Note on the "0 FPS without a visible peak": If external profiling tools fail to register a load spike during a 0 FPS freeze, it is because the engine completely stopped submitting frames. However, the 731ms spike is concrete proof that the engine is already suffocating under load before it ultimately deadlocks.

Conclusion: This is a fatal thread synchronization failure in the 1.26.52 Vibrant Visuals pipeline, not a mere performance bottleneck. Please review the attached charts.

Request & Note:

This issue is not caused by a lack of physical RAM (the OS reports 4343MB free). It is a fatal engine-level deadlock in the 1.26 Vibrant Visuals pipeline. We kindly request QA/Devs to verify this thread synchronization failure on your internal devices using the provided seed. If high-end devices mask the issue due to faster storage/CPU, testing on a low-end Android device or limiting CPU resources is recommended to reproduce the synchronization deadlock.

Attachments:

· (Attached: Video comparison, 0 FPS debug overlay, 731ms load chart).