End of June Test

Some minor improvements have been made.

Mobile Renderer

• Support for the Mobile render method is being suspended.
  • Certain problems arose when trying to use rendering techniques which were compatible with the Mobile render method, mostly performance.
  • Support for it may return in the future, but for the time being, no further work will be done to support it in parallel with Forward+.

Drop Shadows

• Drop Shadows are no longer processed on the main thread. The implementation of Drop Shadows is as optimized as can be made.
  • Performance for Drop Shadows should be better when running the game on a system with multiple CPU cores.
• Light sources can be marked as being ignored for shadow calculations, either for being invisible or too far away.
  • Light sources can also be similarly marked as being on standby, bypassing the blacklist.

Miscellaneous

• Optimized fire to use fewer resources.
• Alterations to greenery and additional plant life visuals.
  •  Thanks, Nick!
• Internal fix for incorrect use of UIDs over file paths.
• Fix for calculating the true speed of moving characters.
  • Rixin and Aria benefit from this. A mild performance improvement was gained for using engine code over GDScript.
• Skeletons' animation playback is conditionally throttled.
  • When a good number of skeletons exist concurrently, the animations can be throttled.
    • Skeletons' animations are throttled more when further away from the player.
• Minor animation fixes for Rixin and Aria.
• The depth-based bloom effect can be turned off, saving a good amount of performance.
• The depth buffer is disabled, and is currently a candidate for removal.
  • Images in Godot are incredibly expensive to create, to the point that odd-frame generation of a depth buffer is insufficient.

• A more sophisticated performance monitor, complete with toggles.
  • Pressing F2 will hide the performance statistics.
  • Pressing a combination of F2 and Shift/Ctrl/Alt will toggle various performance statistics.
    • CPU model and GPU model are present to inform and verify problem areas.
    • Memory is imparted in both Kilobytes and Megabytes.
    • Frames-Per-Second is displayed against its target. The target refresh rate can change if the pace keeper is used.
    • Effective Physics Ticks per-second, marked as PhT. Higher is better.
    • Effective Process Ticks per-second, marked as PrT. Higher is better.
    • PhT and PrT are both sampled for adjusting the refresh rate target when the pace keeper is used.

From glossydon:

A Linux version! Not so hard when you aren't juggling up to four different versions of the same project. 
 
I'm not nearly as contented with what I've accomplished, though I am very happy about what I was able to achieve through the optimizations I made recently. With depth-based effects and systems disabled, a noticeable performance increase can be observed. My minimum-specification computer now hits the 60 FPS mark fairly easily when not under stress and allowed to run at full-speed. When fighting skeletons, it can vary wildly from ~30 FPS to ~57 FPS, depending on the number.
My suspicion as to why the number varies so much is in the AnimationTree, and Godot's very single-threaded nature.
If the playback of animations were instead delegated to separate threads or cores, then I believe Ventury could very reasonably be played at full-speed on a relatively underpowered machine.
There's still an eerie charm to it, the min-spec system. Using the pace keeper, it can really come off like I'm playing a game for a system that is being stressed to its limits. Coupled with the new dithering, and the illusion of playing on an archaic console comes to fruition.
 
Ok, maybe the illusion isn't totally there because of per-pixel shading, but I really quite like where things are going.
 
I look forward to making more enemies, objects, refining the actions and interactions of characters and such, in the name of making a playable demo sometime this year.
 
My free time has been much larger, and I've been on standby for the professional gig. I'm uncertain how long this intermediary period will last, but I will try my best to maximize it.