Metal: Improve AMD EEVEE Performance #104743

Closed
Jason Fielder wants to merge 3 commits from Jason-Fielder/blender:MetalAMDPerformanceEEVEE_2 into main

When changing the target branch, be careful to rebase the branch in your fork to match. See documentation.
Member

Complex EEVEE nodegraphs, particularly those combining multiple principledBSDF shader nodes have a tendancy to require a large number of simultaneous live registers due to function call depth. In some instances, this causes substantial performance drop and corruption if the stack gets too large.

To mitigate this, splitting calls to closure_eval such that only a single individual closure is evaluated in each call reduces the number of live registers required. This is preferred over using compound closure evaluation functions which require a large amount of in-flight data.

Note that this is generally not more optimal, if the stack does not spill, as there is an increased instruction count. The specific trade-off depends on the exact architecture in question. Hence, this is limited to AMD GPUs.

Authored by Apple: Michael Parkin-White

Ref T96261

Complex EEVEE nodegraphs, particularly those combining multiple principledBSDF shader nodes have a tendancy to require a large number of simultaneous live registers due to function call depth. In some instances, this causes substantial performance drop and corruption if the stack gets too large. To mitigate this, splitting calls to closure_eval such that only a single individual closure is evaluated in each call reduces the number of live registers required. This is preferred over using compound closure evaluation functions which require a large amount of in-flight data. Note that this is generally not more optimal, if the stack does not spill, as there is an increased instruction count. The specific trade-off depends on the exact architecture in question. Hence, this is limited to AMD GPUs. Authored by Apple: Michael Parkin-White Ref T96261
Jason Fielder added 1 commit 2023-02-14 14:40:04 +01:00
184345e885 Metal: Improve AMD EEVEE Performance
Complex EEVEE nodegraphs, particularly those combining multiple principledBSDF shader nodes have a tendancy to require a large number of simultaneous live registers due to function call depth. In some instances, this causes substantial performance drop and corruption if the stack gets too large.

To mitigate this, splitting calls to closure_eval such that only a single individual closure is evaluated in each call reduces the number of live registers required. This is preferred over using compound closure evaluation functions which require a large amount of in-flight data.

Note that this is generally not more optimal, if the stack does not spill, as there is an increased instruction count. The specific trade-off depends on the exact architecture in question. Hence, this is limited to AMD GPUs.

Authored by Apple: Michael Parkin-White

Ref T96261
Jason Fielder requested review from Clément Foucault 2023-02-14 14:40:17 +01:00
Brecht Van Lommel added this to the EEVEE & Viewport project 2023-02-15 09:53:46 +01:00
Clément Foucault requested changes 2023-02-15 10:58:41 +01:00
@ -120,2 +120,4 @@
}
CLOSURE_EVAL_FUNCTION_DECLARE_1(GlossyBSDFAlways, Glossy)
Closure closure_eval_always(ClosureReflection reflection)

Change closure_eval_always to be an overload of closure_eval with an optional const bool parameter for ssr output: closure_eval(ClosureReflection reflection, const bool do_ssr_output)

Change `closure_eval_always` to be an overload of `closure_eval` with an optional const bool parameter for ssr output: `closure_eval(ClosureReflection reflection, const bool do_ssr_output)`
Jason Fielder added 2 commits 2023-02-16 12:19:23 +01:00
Clément Foucault approved these changes 2023-02-16 12:59:48 +01:00
Jason Fielder closed this pull request 2023-02-20 19:03:44 +01:00

Pull request closed

Sign in to join this conversation.
No reviewers
No Label
Interest
Alembic
Interest
Animation & Rigging
Interest
Asset Browser
Interest
Asset Browser Project Overview
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
EEVEE & Viewport
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
Virtual Reality
Interest
Vulkan
Interest
Wayland
Interest
Workbench
Interest: X11
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
EEVEE & Viewport
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
Platform
FreeBSD
Platform
Linux
Platform
macOS
Platform
Windows
Priority
High
Priority
Low
Priority
Normal
Priority
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 Assignees
2 Participants
Notifications
Due Date
The due date is invalid or out of range. Please use the format 'yyyy-mm-dd'.

No due date set.

Dependencies

No dependencies set.

Reference: blender/blender#104743
No description provided.