Different Render Results Between CPU and GPU #89351

Closed
opened 2021-06-22 06:47:42 +02:00 by Rick · 17 comments

System Information
Operating system: Windows-10-10.0.19041-SP0 64 Bits
Graphics card: NVIDIA GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 466.77

Blender Version
Broken: version: 2.93.0, branch: master, commit date: 2021-06-02 11:21, hash: 84da05a8b8
Worked: (newest version of Blender that worked as expected)

Short description of error
I've noticed when rendering an ocean scene in cycles, using optix with 1 GPU and a CPU, I get little discolored squares all around the visible portions of the ocean, and the only way to prevent that from happening is to render on either my CPU or GPU alone, switch to a different render device, or just make sure that the ocean isn't visible in the render. Upon writing out this letter, I've also noticed that Optix seems to produce an odd shadow across the object that Cuda does not produce. This shadow also vanishes when rendered in Optix with only my CPU. My graphics card drivers are up to date, and I also tried restarting my PC, but it didn't change anything. I can render with CUDA instead, it's not a problem, but figured I should report the bug anyway.

Exact steps for others to reproduce the error
Render the attached .blend file, with optix as the selected render device, using a single GPU and CPU.Bug Sample blend file.blend Optix CPU and GPU.png

Cuda CPU and GPU.png

Optix CPU only.png

Optix GPU only.png

**System Information** Operating system: Windows-10-10.0.19041-SP0 64 Bits Graphics card: NVIDIA GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 466.77 **Blender Version** Broken: version: 2.93.0, branch: master, commit date: 2021-06-02 11:21, hash: `84da05a8b8` Worked: (newest version of Blender that worked as expected) **Short description of error** I've noticed when rendering an ocean scene in cycles, using optix with 1 GPU and a CPU, I get little discolored squares all around the visible portions of the ocean, and the only way to prevent that from happening is to render on either my CPU or GPU alone, switch to a different render device, or just make sure that the ocean isn't visible in the render. Upon writing out this letter, I've also noticed that Optix seems to produce an odd shadow across the object that Cuda does not produce. This shadow also vanishes when rendered in Optix with only my CPU. My graphics card drivers are up to date, and I also tried restarting my PC, but it didn't change anything. I can render with CUDA instead, it's not a problem, but figured I should report the bug anyway. **Exact steps for others to reproduce the error** Render the attached .blend file, with optix as the selected render device, using a single GPU and CPU.[Bug Sample blend file.blend](https://archive.blender.org/developer/F10187380/Bug_Sample_blend_file.blend) ![Optix CPU and GPU.png](https://archive.blender.org/developer/F10187423/Optix_CPU_and_GPU.png) ![Cuda CPU and GPU.png](https://archive.blender.org/developer/F10187424/Cuda_CPU_and_GPU.png) ![Optix CPU only.png](https://archive.blender.org/developer/F10187425/Optix_CPU_only.png) ![Optix GPU only.png](https://archive.blender.org/developer/F10187426/Optix_GPU_only.png)
Author

Added subscriber: @Arcomade

Added subscriber: @Arcomade

#91570 was marked as duplicate of this issue

#91570 was marked as duplicate of this issue

Changed status from 'Needs Triage' to: 'Confirmed'

Changed status from 'Needs Triage' to: 'Confirmed'

Added subscriber: @Zandman

Added subscriber: @Zandman

Isn't it perfectly normal to get different results between rendering on CPU and GPU? They use different algorithms. In general GPU is faster, but CPU supports more rendering features and produces better results. I think I read somewhere that when you want consistent rendering results, you have to either render on GPU or CPU, but not mix both.

Now, there could definitely still be a bug here. Just wanted to add my two cents.

Isn't it perfectly normal to get different results between rendering on CPU and GPU? They use different algorithms. In general GPU is faster, but CPU supports more rendering features and produces better results. I think I read somewhere that when you want consistent rendering results, you have to either render on GPU or CPU, but not mix both. Now, there could definitely still be a bug here. Just wanted to add my two cents.
Member

Added subscriber: @Blendify

Added subscriber: @Blendify
Member

@Zandman CPU and GPU should give the same results.

@Zandman CPU and GPU should give the same results.

In #89351#1181105, @Blendify wrote:
@Zandman CPU and GPU should give the same results.

Google says otherwise:

But since this isn’t a discussion forum and it’s not my expertise, I don’t have anything more to add here :-)

> In #89351#1181105, @Blendify wrote: > @Zandman CPU and GPU should give the same results. Google says otherwise: * https://www.reddit.com/r/Amd/comments/65h4py/eli5_why_is_gpu_rendering_lower_quality_than_cpu/ * https://renderpool.net/blog/cpu-vs-gpu-rendering/ * https://blog.cherryservers.com/cpu-or-gpu-rendering-which-is-better-one But since this isn’t a discussion forum and it’s not my expertise, I don’t have anything more to add here :-)

Closed as duplicate of #85078

Closed as duplicate of #85078

What makes you think this is a duplicate of #85078? This issue is about mixing CPU and GPU, right. The other one is just using one variant.

What makes you think this is a duplicate of #85078? This issue is about mixing CPU and GPU, right. The other one is just using one variant.
Member

Added subscriber: @Alaska

Added subscriber: @Alaska
Member

Changed status from 'Duplicate' to: 'Needs Triage'

Changed status from 'Duplicate' to: 'Needs Triage'
Alaska reopened this issue 2021-09-18 08:48:34 +02:00
Member

In #89351#1182663, @Zandman wrote:
What makes you think this is a duplicate of #85078? This issue is about mixing CPU and GPU, right. The other one is just using one variant.

The reason why these reports were probably merged is because the report is talking about a issue with the GPU rendering differently from the CPU. In this specific case, that OptiX is rendering differently from the CPU.
#85078 (OptiX GPU rendering artifacts with large plane) is a bug report for the issue that's causing OptiX to render differently from the CPU in this case. So once #85078 (OptiX GPU rendering artifacts with large plane) is fixed then this specific bug report will also be fixed.

> In #89351#1182663, @Zandman wrote: > What makes you think this is a duplicate of #85078? This issue is about mixing CPU and GPU, right. The other one is just using one variant. The reason why these reports were probably merged is because the report is talking about a issue with the GPU rendering differently from the CPU. In this specific case, that OptiX is rendering differently from the CPU. #85078 (OptiX GPU rendering artifacts with large plane) is a bug report for the issue that's causing OptiX to render differently from the CPU in this case. So once #85078 (OptiX GPU rendering artifacts with large plane) is fixed then this specific bug report will also be fixed.

Closed as duplicate of #85078

Closed as duplicate of #85078

Added subscriber: @brecht

Added subscriber: @brecht

If the underlying issue is the same, reports should be merged.

If the underlying issue is the same, reports should be merged.

Added subscribers: @PabloCio, @ThomasDinges, @PeterBaintner

Added subscribers: @PabloCio, @ThomasDinges, @PeterBaintner
Sign in to join this conversation.
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 project
No Assignees
7 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#89351
No description provided.