FiveM performance guide
Judge MLO performance with real tests.
Why 0.00 ms in resmon does not guarantee a smooth interior, and how to test streaming, textures, collision, LODs and conflicts.
- Separate client and server
- Repeatable tests
- Inspect streaming
- Find conflicts
01
Resmon does not measure the whole MLO
Server time, client FPS and streaming are different.
A map-only resource can show almost no script time and still hitch on clients. Models, textures, collision, portals and lighting stream to the player and mainly affect rendering, VRAM and the streaming pool.
Resmon matters for door, target, elevator or job scripts. The map itself also needs repeatable in-game tests on more than one hardware class.
“0.00 ms” is not a complete MLO performance specification.
02
The key test areas
Each answers a different question.
- Streaming and hitches
- Test first approach, fast driving and repeated entry. Late models or freezes suggest heavy assets or streaming pressure.
- Textures and VRAM
- Oversized textures consume VRAM. Test weaker GPUs and the combined city, not only one resource.
- Collision
- Overly complex collision costs performance. Test floors, stairs, doors, vehicles and projectiles.
- LODs, portals and occlusion
- Invisible geometry should not render unnecessarily. Test outside, rooms and sight lines.
- Scripts and integrations
- Measure doors, targets and continuous threads separately in resmon.
- Location conflicts
- Two maps at one location may override YMAP, collision and portals.
03
A repeatable performance test
Compare before and after under the same conditions.
Record a baseline
Test without the MLO at the same location and note FPS, hitches and client behavior.
Enable only the MLO
Run the map and mandatory dependencies without optional jobs first.
Test several routes
Approach on foot and at speed, visit every room and move between inside and outside.
Include a weaker client
A development PC can hide issues players with less VRAM will notice.
Test the full city
Enable the real resource stack to reveal cumulative streaming and texture pressure.
Record game build, hardware, active resources and scenario so results remain comparable.
04
Warning signs
Ask for detail when you see these.
- Only a resmon screenshot
- No texture or LOD information
- Demo only on high-end hardware
- No known location conflicts
- Unclear game-build requirements
- “Optimized” without criteria
05
MLO performance questions
Common misconceptions.
01
Is 0.00 ms automatically optimized?
No. Resmon mainly reports script time. Rendering, VRAM, streaming, collision and portals need separate tests.
02
Can an MLO lag the server?
The map mainly affects clients. Scripts, entities and integrations may also affect server time and networking.
03
Why do buildings flicker?
Often two maps overlap at the same location. Disable competing resources one at a time.
04
What texture size is allowed?
No single limit fits every asset. Count, format, screen area, mipmaps and total resource load matter.
05
How should I test before launch?
Use staging, multiple hardware classes and a fixed approach, interior, player-count and restart scenario.
Nobody Workz Studio
MLOs built for practical use
Browse our FiveM interiors and test them with your real server stack.