AutoPeek v1.3.0 – “No-More-War-Crimes” Edition
TL;DR
A full rewrite of the movement-prediction core.
FPS jumps from ~10 → 40+, warp is finally stable again, and the codebase no longer violates the Geneva Convention.
✨ Highlights
| Category | What changed | Why you care |
|---|---|---|
| Performance | One global simulation per tick, cached & reused.Removed 4-5 redundant SimulateMovement() calls per CreateMove.Visual-scan loop trimmed to 15 binary-search iterations. | CPU time per frame down by ~75 %.Real-world test on mid-tier laptop: FPS 10 → 40-60 (full hit-box checks ON). |
| Stability | Re-synced prediction & server tick; fixed “warp” that desynced after heavy load. | No more rubber-band deaths or missed peeks. |
| Memory | Nuked monster tables; on-demand indexing instead of full-map pre-alloc.Freed ~20 MB per round. | Still chunky at high settings, but no longer a RAM-sinkhole. |
| Code cleanup | • Modularised simulation layer• Purged dead debug hooks• Added inline profiler markers | Devs can read it without crying; future profiling is one key-press away. |
| Settings | New default config shipped (autopeek_default.json) tuned for 60 FPS rigs.All legacy keys migrated automatically. | Plug-and-play; no more first-launch stutter while you hunt sliders. |
🐛 Fixed
-
Warp / perfect-tick routine now tracks fractional ticks and clamps error < 0.1 ms.
-
Binary-search visual routine no longer spawns its own full simulation each iteration.
-
Edge-peek radius visualizer respects
r_drawparticles 0to avoid GPU bottleneck.
⚠️ Known Issues
-
The green “circle-under-feet” overlay still costs ~2 ms per frame on potato GPUs.
-
Memory spikes if you crank iterations > 30 and enable full hit-box heat-map; best left to future patch.
🚀 Upgrading
-
Pull/clone the tag or use the
v1.3.0zip. -
Delete your old
autopeek.cfg; a fresh one is generated on first run. -
Optional:
profile 1in console to view the new inline markers.
Enjoy your silky-smooth 40 FPS peeks.
— Terminator & contributors