Blender crashed with enabled HIP support #112608

Closed
opened 2023-09-20 01:43:50 +02:00 by Luya Tshimbalanga · 11 comments

System Information
Operating system: Fedora Linux 39 (Design Suite Prerelease)
Graphics card: AMD Radeon RX 6950 XT (navi21, LLVM 16.0.6, DRM 3.54, 6.5.3-300.fc39.x86_64)

ROCM HIP version 5.6.1

Blender Version
Broken: 3.6.2
Worked: n/a

Short description of error
Blender silently crashed when rendering an object with enabled HIP support on Cycle engine

Exact steps for others to reproduce the error
Any blend file including default start up. In this case, a Ryzen chipset demo attached below. Also see both system-info and blender_debug for details.

**System Information** Operating system: Fedora Linux 39 (Design Suite Prerelease) Graphics card: AMD Radeon RX 6950 XT (navi21, LLVM 16.0.6, DRM 3.54, 6.5.3-300.fc39.x86_64) ROCM HIP version 5.6.1 **Blender Version** Broken: 3.6.2 Worked: n/a **Short description of error** Blender silently crashed when rendering an object with enabled HIP support on Cycle engine **Exact steps for others to reproduce the error** Any blend file including default start up. In this case, a Ryzen chipset demo attached below. Also see both system-info and blender_debug for details.
Luya Tshimbalanga added the
Priority
Normal
Type
Report
Status
Needs Triage
labels 2023-09-20 01:43:51 +02:00
Member

Will add to the HIP bug list.

Will add to the HIP bug list.
Member

Hi, please also attach /tmp/RyzenGraphic_27.crash.txt

Hi, please also attach `/tmp/RyzenGraphic_27.crash.txt`
Pratik Borhade added
Status
Needs Information from User
and removed
Status
Needs Triage
labels 2023-09-20 04:32:38 +02:00

Here is the requested attached file. It lacks important information like traceback =(

Here is the requested attached file. It lacks important information like traceback =(
Member

Thanks. Launch with --debug-cycles / --debug-gpu argument, see if that generates additional information about crash.
Could you try drivers from official amd website?:

Thanks. Launch with `--debug-cycles / --debug-gpu` argument, see if that generates additional information about crash. Could you try drivers from official amd website?: - https://www.amd.com/en/support/graphics/amd-radeon-rx-6000-series/amd-radeon-rx-6900-series/amd-radeon-rx-6950-xt - https://docs.blender.org/manual/en/dev/troubleshooting/gpu/linux/amd.html

Additional info

blender --debug-cycles
Read prefs: "/home/luya/.config/blender/3.6/config/userpref.blend"
Read blend: "/home/luya/Documents/design stuff/blender/RyzenGraphic_27.blend"
WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0.
WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0.
WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0.
WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0.
WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0.
WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0.
WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0.
WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0.
I0921 00:14:21.590732  6912 device.cpp:36] HIPEW initialization succeeded
I0921 00:14:21.590787  6912 device.cpp:38] Found precompiled kernels
I0921 00:14:21.607435  6912 device.cpp:196] Device has compute preemption or is not used for display.
I0921 00:14:21.607455  6912 device.cpp:200] Added device "AMD Radeon RX 6950 XT" with id "HIP_AMD Radeon RX 6950 XT_0000:0b:00".
I0921 00:14:23.282105  7039 device.cpp:511] Mapped host memory limit set to 62,482,317,312 bytes. (58.19G)
I0921 00:14:23.282312  7039 device_impl.cpp:62] Using AVX2 CPU kernels.
Writing: /tmp/RyzenGraphic_27.crash.txt
Segmentation fault (core dumped)

With --debug-gpu

blender --debug-cycles  --debug-gpu
Read prefs: "/home/luya/.config/blender/3.6/config/userpref.blend"
INFO (gpu.debug): Notification : Successfully hooked OpenGL debug callback using OpenGL 4.3
INFO (gpu.debug): Notification : Successfully hooked OpenGL debug callback using OpenGL 4.3
Read blend: "/home/luya/Documents/design stuff/blender/RyzenGraphic_27.blend"
WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0.
WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0.
WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0.
WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0.
WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0.
WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0.
WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0.
WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0.
I0921 00:16:10.911640  7299 device.cpp:36] HIPEW initialization succeeded
I0921 00:16:10.911669  7299 device.cpp:38] Found precompiled kernels
I0921 00:16:10.926707  7299 device.cpp:196] Device has compute preemption or is not used for display.
I0921 00:16:10.926720  7299 device.cpp:200] Added device "AMD Radeon RX 6950 XT" with id "HIP_AMD Radeon RX 6950 XT_0000:0b:00".
INFO (gpu.debug): Notification : Successfully hooked OpenGL debug callback using OpenGL 4.3
I0921 00:16:13.190274  7463 device.cpp:511] Mapped host memory limit set to 62,482,317,312 bytes. (58.19G)
I0921 00:16:13.190510  7463 device_impl.cpp:62] Using AVX2 CPU kernels.
Writing: /tmp/RyzenGraphic_27.crash.txt
Segmentation fault (core dumped)

As the issue seems related to HIP, the listed driver on the lack lack ROCm support.

Additional info ``` blender --debug-cycles Read prefs: "/home/luya/.config/blender/3.6/config/userpref.blend" Read blend: "/home/luya/Documents/design stuff/blender/RyzenGraphic_27.blend" WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0. WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0. WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0. WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0. WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0. WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0. WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0. WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0. I0921 00:14:21.590732 6912 device.cpp:36] HIPEW initialization succeeded I0921 00:14:21.590787 6912 device.cpp:38] Found precompiled kernels I0921 00:14:21.607435 6912 device.cpp:196] Device has compute preemption or is not used for display. I0921 00:14:21.607455 6912 device.cpp:200] Added device "AMD Radeon RX 6950 XT" with id "HIP_AMD Radeon RX 6950 XT_0000:0b:00". I0921 00:14:23.282105 7039 device.cpp:511] Mapped host memory limit set to 62,482,317,312 bytes. (58.19G) I0921 00:14:23.282312 7039 device_impl.cpp:62] Using AVX2 CPU kernels. Writing: /tmp/RyzenGraphic_27.crash.txt Segmentation fault (core dumped) ``` With --debug-gpu ``` blender --debug-cycles --debug-gpu Read prefs: "/home/luya/.config/blender/3.6/config/userpref.blend" INFO (gpu.debug): Notification : Successfully hooked OpenGL debug callback using OpenGL 4.3 INFO (gpu.debug): Notification : Successfully hooked OpenGL debug callback using OpenGL 4.3 Read blend: "/home/luya/Documents/design stuff/blender/RyzenGraphic_27.blend" WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0. WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0. WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0. WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0. WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0. WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0. WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0. WARN (bke.customdata): source/blender/blenkernel/intern/customdata.cc:4525 CustomData_layer_ensure_data_exists: CustomDataLayer->data is NULL for type 0. I0921 00:16:10.911640 7299 device.cpp:36] HIPEW initialization succeeded I0921 00:16:10.911669 7299 device.cpp:38] Found precompiled kernels I0921 00:16:10.926707 7299 device.cpp:196] Device has compute preemption or is not used for display. I0921 00:16:10.926720 7299 device.cpp:200] Added device "AMD Radeon RX 6950 XT" with id "HIP_AMD Radeon RX 6950 XT_0000:0b:00". INFO (gpu.debug): Notification : Successfully hooked OpenGL debug callback using OpenGL 4.3 I0921 00:16:13.190274 7463 device.cpp:511] Mapped host memory limit set to 62,482,317,312 bytes. (58.19G) I0921 00:16:13.190510 7463 device_impl.cpp:62] Using AVX2 CPU kernels. Writing: /tmp/RyzenGraphic_27.crash.txt Segmentation fault (core dumped) ``` As the issue seems related to HIP, the listed driver on the lack lack ROCm support.
Pratik Borhade added
Status
Needs Triage
and removed
Status
Needs Information from User
labels 2023-09-21 13:18:55 +02:00
Member
cc @BrianSavery
Pratik Borhade added
Module
Render & Cycles
Status
Needs Info from Developers
and removed
Status
Needs Triage
labels 2023-09-26 05:22:03 +02:00

Does not appear to work for me either. Trying to render anything in Cycles results in OP's behaviour. The following results are with a completely unmodified default cube scene; simply entering the scene after the start-up pop-up and hitting F12.

  • Fedora 39
  • Linux kernel 6.7.9
  • GPU is RX 6700 XT
  • Tested with both X11 and Wayland. Same result in either case.
  • Blender 4.0.2
  • ROCm 5.7.1

Output of blender --debug-cycles:

Read prefs: "/home/maxwellj/.config/blender/4.0/config/userpref.blend"
WARN (ghost.wl.keyboard.depressed): intern/ghost/intern/GHOST_SystemWayland.cc:2011 keyboard_depressed_state_key_event: modifier (6) has negative keys held (-1)!
I0320 12:12:41.514802 65091 device.cpp:37] HIPEW initialization succeeded
I0320 12:12:41.514838 65091 device.cpp:39] Found precompiled kernels
I0320 12:12:41.532549 65091 device.cpp:197] Device has compute preemption or is not used for display.
I0320 12:12:41.532573 65091 device.cpp:201] Added device "AMD Radeon RX 6700 XT" with id "HIP_AMD Radeon RX 6700 XT_0000:08:00".
I0320 12:12:48.860416 65217 device.cpp:407] CPU render threads reduced from 16 to 15, to dedicate to GPU.
I0320 12:12:48.860512 65217 device.cpp:541] Mapped host memory limit set to 29,280,153,600 bytes. (27.27G)
I0320 12:12:48.860538 65217 device_impl.cpp:63] Using AVX2 CPU kernels.
I0320 12:12:48.861395 65217 device_impl.cpp:63] Using AVX2 CPU kernels.
Writing: /tmp/blender.crash.txt
zsh: segmentation fault (core dumped)  blender --debug-cycles

Contents of /tmp/blender.crash.txt:

# Blender 4.0.2, Commit date: 1970-01-01 00:00, Hash unknown
bpy.context.space_data.context = 'RENDER'  # Property
bpy.context.scene.render.engine = 'CYCLES'  # Property
bpy.context.scene.cycles.device = 'GPU'  # Property

# backtrace
blender(+0xef5c9c) [0x5556474aec9c]
blender(+0x9549eb) [0x555646f0d9eb]
/lib64/libc.so.6(+0x3e9a0) [0x7fb98ac5c9a0]

# Python backtrace
Does not appear to work for me either. Trying to render anything in Cycles results in OP's behaviour. The following results are with a completely unmodified default cube scene; simply entering the scene after the start-up pop-up and hitting F12. - Fedora 39 - Linux kernel 6.7.9 - GPU is RX 6700 XT - Tested with both X11 and Wayland. Same result in either case. - Blender 4.0.2 - ROCm 5.7.1 Output of `blender --debug-cycles`: ``` Read prefs: "/home/maxwellj/.config/blender/4.0/config/userpref.blend" WARN (ghost.wl.keyboard.depressed): intern/ghost/intern/GHOST_SystemWayland.cc:2011 keyboard_depressed_state_key_event: modifier (6) has negative keys held (-1)! I0320 12:12:41.514802 65091 device.cpp:37] HIPEW initialization succeeded I0320 12:12:41.514838 65091 device.cpp:39] Found precompiled kernels I0320 12:12:41.532549 65091 device.cpp:197] Device has compute preemption or is not used for display. I0320 12:12:41.532573 65091 device.cpp:201] Added device "AMD Radeon RX 6700 XT" with id "HIP_AMD Radeon RX 6700 XT_0000:08:00". I0320 12:12:48.860416 65217 device.cpp:407] CPU render threads reduced from 16 to 15, to dedicate to GPU. I0320 12:12:48.860512 65217 device.cpp:541] Mapped host memory limit set to 29,280,153,600 bytes. (27.27G) I0320 12:12:48.860538 65217 device_impl.cpp:63] Using AVX2 CPU kernels. I0320 12:12:48.861395 65217 device_impl.cpp:63] Using AVX2 CPU kernels. Writing: /tmp/blender.crash.txt zsh: segmentation fault (core dumped) blender --debug-cycles ``` Contents of `/tmp/blender.crash.txt`: ``` # Blender 4.0.2, Commit date: 1970-01-01 00:00, Hash unknown bpy.context.space_data.context = 'RENDER' # Property bpy.context.scene.render.engine = 'CYCLES' # Property bpy.context.scene.cycles.device = 'GPU' # Property # backtrace blender(+0xef5c9c) [0x5556474aec9c] blender(+0x9549eb) [0x555646f0d9eb] /lib64/libc.so.6(+0x3e9a0) [0x7fb98ac5c9a0] # Python backtrace ```

OS: Fedora 39
Kernel: 6.7.10-200.fc39.x86_64
tested with X11
GPU: AMD ATI Radeon RX 7900 XTX
Blender 4.0.2
Tested with both ROCm 5.7.1 and 5.6.1, I have the exact same backtrace as Maxwell-Jensen (in some cases reverting ROCm to 5.6.1 did help, sadly not here)
Tested with --debug-all flag
Tested by enabling GPU in options and trying to render viewport with cycles

OS: Fedora 39 Kernel: 6.7.10-200.fc39.x86_64 tested with X11 GPU: AMD ATI Radeon RX 7900 XTX Blender 4.0.2 Tested with both ROCm 5.7.1 and 5.6.1, I have the exact same backtrace as Maxwell-Jensen (in some cases reverting ROCm to 5.6.1 did help, sadly not here) Tested with `--debug-all` flag Tested by enabling GPU in options and trying to render viewport with cycles

OS: Fedora 40
Kernel: 6..8.7-300.fc40.x86_64
tested with X11
GPU: AMD ATI Radeon RX 7900 XTX
Blender 4.1.1
ROCM 6.0.2
Rendering with HIP works, viewport render works, tested with 7900 XTX, 7900 XTX + 7950X CPU, GPU + iGPU + CPU

OS: Fedora 40 Kernel: 6..8.7-300.fc40.x86_64 tested with X11 GPU: AMD ATI Radeon RX 7900 XTX Blender 4.1.1 ROCM 6.0.2 Rendering with HIP works, viewport render works, tested with 7900 XTX, 7900 XTX + 7950X CPU, GPU + iGPU + CPU

I can confirm that Blender now works after upgrading to Fedora 40. Same exact software specifications as @Halamix2.

I can confirm that Blender now works after upgrading to Fedora 40. Same exact software specifications as @Halamix2.
Member

Thanks. I think we can close this now

Thanks. I think we can close this now
Blender Bot added
Status
Archived
and removed
Status
Needs Info from Developers
labels 2024-04-26 08:27:28 +02:00
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
5 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#112608
No description provided.