The performance of Blender decreases as the number of buffered frame increases #108097
Labels
No Label
Interest
Alembic
Interest
Animation & Rigging
Interest
Asset System
Interest
Audio
Interest
Automated Testing
Interest
Blender Asset Bundle
Interest
BlendFile
Interest
Collada
Interest
Compatibility
Interest
Compositing
Interest
Core
Interest
Cycles
Interest
Dependency Graph
Interest
Development Management
Interest
EEVEE
Interest
Freestyle
Interest
Geometry Nodes
Interest
Grease Pencil
Interest
ID Management
Interest
Images & Movies
Interest
Import Export
Interest
Line Art
Interest
Masking
Interest
Metal
Interest
Modeling
Interest
Modifiers
Interest
Motion Tracking
Interest
Nodes & Physics
Interest
OpenGL
Interest
Overlay
Interest
Overrides
Interest
Performance
Interest
Physics
Interest
Pipeline, Assets & IO
Interest
Platforms, Builds & Tests
Interest
Python API
Interest
Render & Cycles
Interest
Render Pipeline
Interest
Sculpt, Paint & Texture
Interest
Text Editor
Interest
Translations
Interest
Triaging
Interest
Undo
Interest
USD
Interest
User Interface
Interest
UV Editing
Interest
VFX & Video
Interest
Video Sequencer
Interest
Viewport & EEVEE
Interest
Virtual Reality
Interest
Vulkan
Interest
Wayland
Interest
Workbench
Interest: X11
Legacy
Asset Browser Project
Legacy
Blender 2.8 Project
Legacy
Milestone 1: Basic, Local Asset Browser
Legacy
OpenGL Error
Meta
Good First Issue
Meta
Papercut
Meta
Retrospective
Meta
Security
Module
Animation & Rigging
Module
Core
Module
Development Management
Module
Grease Pencil
Module
Modeling
Module
Nodes & Physics
Module
Pipeline, Assets & IO
Module
Platforms, Builds & Tests
Module
Python API
Module
Render & Cycles
Module
Sculpt, Paint & Texture
Module
Triaging
Module
User Interface
Module
VFX & Video
Module
Viewport & EEVEE
Platform
FreeBSD
Platform
Linux
Platform
macOS
Platform
Windows
Severity
High
Severity
Low
Severity
Normal
Severity
Unbreak Now!
Status
Archived
Status
Confirmed
Status
Duplicate
Status
Needs Info from Developers
Status
Needs Information from User
Status
Needs Triage
Status
Resolved
Type
Bug
Type
Design
Type
Known Issue
Type
Patch
Type
Report
Type
To Do
No Milestone
No project
No Assignees
5 Participants
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: blender/blender#108097
Loading…
Reference in New Issue
Block a user
No description provided.
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
System Information
Operating system: Windows-10-10.0.19041-SP0 64 Bits
Graphics card: NVIDIA GeForce RTX 3050 Ti Laptop GPU/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.61
Blender Version
Broken: version: 3.6.0 Beta, branch: blender-v3.6-release, commit date: 2023-05-17 23:33, hash:
a9358f5274b6
Worked: unknown
Short description of error
[Using machine translation]
I created the “game of life” that uses simulation zone. This simulation performed well at the beginning, but when I played it to about 4000 frames, the entire Blender got stuck and the frame rate of the simulation was also decrease to lower than 24fps. At this point the utilization of CPU and GPU was also decreasing. Strange.
Exact steps for others to reproduce the error
use the attached .blend file (need blender 3.6)
play animation
observe frame rate changes
Hey, thanks for report this. Can you check, if on 4000...frame your RAM is full and your hard disk is start working?
Blender used about 1.6GB RAM with almost no SSD used which should not be a issue. Maybe the bottleneck is memory bandwidth? I'm using 16G DDR4 3200Mhz dual channel memory. Performance almost linearly decreases with frame count.
I can confirm the issue. It has to do with the Cache option (the in memory cache) as can be seen below. Performance remains "slow" even if you replay the parts of the simulation that's been cached. Baking the simulation does not experience the problem.
Still happened in the latest build (hash:
e2a75814ac
)Some more information: It seems like the Timeline being opened has an effect on the situation.
A quick profile does show a problematic section of code inside
timeline_cache_draw_simulation_nodes
. It loops over each frame and 1) calls intoModifierSimulationCache::has_state_at_frame
which also loops over frames in the cache (n^2 behavior) and 2) creates many small rectangles to draw the actual cached line which can maybe be made more efficient as wellThanks for looking into this @deadpin.
@JacquesLucke @LukasStockner I guess
ModifierSimulationCache::states_at_frames_
should become a map orVectorSet
, or we should guarantee that it's sorted and use binary search here.Using binary search sounds good.
Fixed by
6a05e5161b
. Thanks for reporting!