FS#75097 - [blender] GPU-rendering using HIP now working in blender-3.2.0-3
Attached to Project:
Arch Linux
Opened by Martin Saraceno (tinux) - Friday, 17 June 2022, 07:45 GMT
Last edited by Toolybird (Toolybird) - Saturday, 16 July 2022, 23:13 GMT
Opened by Martin Saraceno (tinux) - Friday, 17 June 2022, 07:45 GMT
Last edited by Toolybird (Toolybird) - Saturday, 16 July 2022, 23:13 GMT
|
Details
Description:
GPU-rendering using HIP and a supported RDNA2 GPU (6900 XT in my case) does not work using the arch-packaged version. However, the pre-built version from blender.org does work. Additional info: * not-working arch package blender-3.2.0-3 * working version from blender.org: e05e1e369187 Steps to reproduce: * start blender in the console * set to HIP in the system settings * make changes to render on the (AMD) GPU * render on GPU console output: ``` blender Read prefs: /home/tinux/.config/blender/3.2/config/userpref.blend Traceback (most recent call last): File "/usr/share/blender/3.2/scripts/modules/addon_utils.py", line 335, in enable mod = __import__(module_name) ModuleNotFoundError: No module named 'io_import_obj' Read blend: /home/tinux/Nextcloud/blender/GHz-tdl/GHz-tdl.blend Compiling HIP kernel ... hipcc -Wno-parentheses-equality -Wno-unused-value --hipcc-func-supp -O3 -ffast-math --amdgpu-target=gfx1030 -I /usr/share/blender/3.2/scripts/addons/cycles/source --genco /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/hip/kernel.cpp -o "/home/tinux/.cache/cycles/kernels/cycles_kernel_gfx1030_BA3FAEAF2213AAED34D26DDD69E1C387" In file included from /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/hip/kernel.cpp:12: /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:55:1: error: unknown type name '__noinline__' ccl_device_noinline T kernel_tex_image_interp_bicubic(ccl_global const TextureInfo &info, ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/hip/compat.h:33:40: note: expanded from macro 'ccl_device_noinline' #define ccl_device_noinline __device__ __noinline__ ^ In file included from /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/hip/kernel.cpp:12: /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:55:21: warning: variable templates are a C++14 extension [-Wc++14-extensions] ccl_device_noinline T kernel_tex_image_interp_bicubic(ccl_global const TextureInfo &info, ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:55:22: error: expected ';' at end of declaration ccl_device_noinline T kernel_tex_image_interp_bicubic(ccl_global const TextureInfo &info, ^ ; /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:55:23: error: C++ requires a type specifier for all declarations ccl_device_noinline T kernel_tex_image_interp_bicubic(ccl_global const TextureInfo &info, ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:69:15: error: no matching function for call to 'cubic_g0' float g0x = cubic_g0(fx); ^~~~~~~~ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:34:18: note: candidate function not viable: call to __device__ function from __host__ function ccl_device float cubic_g0(float a) ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:70:15: error: no matching function for call to 'cubic_g1' float g1x = cubic_g1(fx); ^~~~~~~~ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:38:18: note: candidate function not viable: call to __device__ function from __host__ function ccl_device float cubic_g1(float a) ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:72:20: error: no matching function for call to 'cubic_h0' float x0 = (px + cubic_h0(fx) + 0.5f) / info.width; ^~~~~~~~ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:44:18: note: candidate function not viable: call to __device__ function from __host__ function ccl_device float cubic_h0(float a) ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:73:20: error: no matching function for call to 'cubic_h1' float x1 = (px + cubic_h1(fx) + 0.5f) / info.width; ^~~~~~~~ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:48:18: note: candidate function not viable: call to __device__ function from __host__ function ccl_device float cubic_h1(float a) ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:74:20: error: no matching function for call to 'cubic_h0' float y0 = (py + cubic_h0(fy) + 0.5f) / info.height; ^~~~~~~~ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:44:18: note: candidate function not viable: call to __device__ function from __host__ function ccl_device float cubic_h0(float a) ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:75:20: error: no matching function for call to 'cubic_h1' float y1 = (py + cubic_h1(fy) + 0.5f) / info.height; ^~~~~~~~ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:48:18: note: candidate function not viable: call to __device__ function from __host__ function ccl_device float cubic_h1(float a) ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:77:10: error: no matching function for call to 'cubic_g0' return cubic_g0(fy) * (g0x * ccl_gpu_tex_object_read_2D<T>(tex, x0, y0) + ^~~~~~~~ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:34:18: note: candidate function not viable: call to __device__ function from __host__ function ccl_device float cubic_g0(float a) ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:79:10: error: no matching function for call to 'cubic_g1' cubic_g1(fy) * (g0x * ccl_gpu_tex_object_read_2D<T>(tex, x0, y1) + ^~~~~~~~ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:38:18: note: candidate function not viable: call to __device__ function from __host__ function ccl_device float cubic_g1(float a) ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:85:1: error: unknown type name '__noinline__' ccl_device_noinline T ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/hip/compat.h:33:40: note: expanded from macro 'ccl_device_noinline' #define ccl_device_noinline __device__ __noinline__ ^ In file included from /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/hip/kernel.cpp:12: /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:85:21: warning: variable templates are a C++14 extension [-Wc++14-extensions] ccl_device_noinline T ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:85:22: error: expected ';' at end of declaration ccl_device_noinline T ^ ; /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:86:1: error: C++ requires a type specifier for all declarations kernel_tex_image_interp_tricubic(ccl_global const TextureInfo &info, float x, float y, float z) ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:101:15: error: no matching function for call to 'cubic_g0' float g0x = cubic_g0(fx); ^~~~~~~~ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:34:18: note: candidate function not viable: call to __device__ function from __host__ function ccl_device float cubic_g0(float a) ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:102:15: error: no matching function for call to 'cubic_g1' float g1x = cubic_g1(fx); ^~~~~~~~ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:38:18: note: candidate function not viable: call to __device__ function from __host__ function ccl_device float cubic_g1(float a) ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:103:15: error: no matching function for call to 'cubic_g0' float g0y = cubic_g0(fy); ^~~~~~~~ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:34:18: note: candidate function not viable: call to __device__ function from __host__ function ccl_device float cubic_g0(float a) ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:104:15: error: no matching function for call to 'cubic_g1' float g1y = cubic_g1(fy); ^~~~~~~~ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:38:18: note: candidate function not viable: call to __device__ function from __host__ function ccl_device float cubic_g1(float a) ^ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:105:15: error: no matching function for call to 'cubic_g0' float g0z = cubic_g0(fz); ^~~~~~~~ /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:34:18: note: candidate function not viable: call to __device__ function from __host__ function ccl_device float cubic_g0(float a) ^ fatal error: too many errors emitted, stopping now [-ferror-limit=] 2 warnings and 20 errors generated when compiling for gfx1030. Failed to execute compilation command, see console for details. Refer to the Cycles GPU rendering documentation for possible solutions: https://docs.blender.org/manual/en/latest/render/cycles/gpu_rendering.html Saved session recovery to '/tmp/quit.blend' Blender quit Error: Not freed memory blocks: 2, total unfreed memory 0.000168 MB ``` |
This task depends upon
Closed by Toolybird (Toolybird)
Saturday, 16 July 2022, 23:13 GMT
Reason for closing: Fixed
Additional comments about closing: @grinness - Current Blender Arch package (Version : 17:3.2.1-1) finds and uses ROCM (5.2.0 -- installed from AUR) and enables GPU cycles hip rendering
Saturday, 16 July 2022, 23:13 GMT
Reason for closing: Fixed
Additional comments about closing: @grinness - Current Blender Arch package (Version : 17:3.2.1-1) finds and uses ROCM (5.2.0 -- installed from AUR) and enables GPU cycles hip rendering
same problem here. Blender binary from blender.org (3.2.0) renders fine with GPU (rx6800) using HIP (rocm 5.1.3 compiled from AUR) and opensouce amdgpu kernel module (5.18.5-zen1-1-zen)
Running blender ARCH package(3.2.0-3), rendering using HIP break with following error on console:
022-06-21 09:03:45,597 INFO rpr.engine.init [139918294880256]: Core version: 2.2.13 build 0xbe24f121
2022-06-21 09:03:45,601 INFO rpr.engine.init [139918294880256]: RIF version: 1.7.1 build 0xfdc139ba
Error: Required device extension VK_KHR_acceleration_structure is not supported.
2022-06-21 09:03:45,657 WARNING rpr.engine.init [139918294880256]: ('Plugin is not registered', PosixPath('/home/marco/.config/blender/3.2/scripts/addons/rprblender/HybridPro.so'), '')
Compiling HIP kernel ...
hipcc -Wno-parentheses-equality -Wno-unused-value --hipcc-func-supp -O3 -ffast-math --amdgpu-target=gfx1030 -I /usr/share/blender/3.2/scripts/addons/cycles/source --genco /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/hip/kernel.cpp -o "/home/marco/.cache/cycles/kernels/cycles_kernel_gfx1030_BA3FAEAF2213AAED34D26DDD69E1C387"
In file included from /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/hip/kernel.cpp:12:
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:55:1: error: unknown type name '__noinline__'
ccl_device_noinline T kernel_tex_image_interp_bicubic(ccl_global const TextureInfo &info,
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/hip/compat.h:33:40: note: expanded from macro 'ccl_device_noinline'
#define ccl_device_noinline __device__ __noinline__
^
In file included from /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/hip/kernel.cpp:12:
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:55:21: warning: variable templates are a C++14 extension [-Wc++14-extensions]
ccl_device_noinline T kernel_tex_image_interp_bicubic(ccl_global const TextureInfo &info,
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:55:22: error: expected ';' at end of declaration
ccl_device_noinline T kernel_tex_image_interp_bicubic(ccl_global const TextureInfo &info,
^
;
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:55:23: error: C++ requires a type specifier for all declarations
ccl_device_noinline T kernel_tex_image_interp_bicubic(ccl_global const TextureInfo &info,
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:69:15: error: no matching function for call to 'cubic_g0'
float g0x = cubic_g0(fx);
^~~~~~~~
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:34:18: note: candidate function not viable: call to __device__ function from __host__ function
ccl_device float cubic_g0(float a)
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:70:15: error: no matching function for call to 'cubic_g1'
float g1x = cubic_g1(fx);
^~~~~~~~
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:38:18: note: candidate function not viable: call to __device__ function from __host__ function
ccl_device float cubic_g1(float a)
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:72:20: error: no matching function for call to 'cubic_h0'
float x0 = (px + cubic_h0(fx) + 0.5f) / info.width;
^~~~~~~~
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:44:18: note: candidate function not viable: call to __device__ function from __host__ function
ccl_device float cubic_h0(float a)
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:73:20: error: no matching function for call to 'cubic_h1'
float x1 = (px + cubic_h1(fx) + 0.5f) / info.width;
^~~~~~~~
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:48:18: note: candidate function not viable: call to __device__ function from __host__ function
ccl_device float cubic_h1(float a)
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:74:20: error: no matching function for call to 'cubic_h0'
float y0 = (py + cubic_h0(fy) + 0.5f) / info.height;
^~~~~~~~
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:44:18: note: candidate function not viable: call to __device__ function from __host__ function
ccl_device float cubic_h0(float a)
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:75:20: error: no matching function for call to 'cubic_h1'
float y1 = (py + cubic_h1(fy) + 0.5f) / info.height;
^~~~~~~~
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:48:18: note: candidate function not viable: call to __device__ function from __host__ function
ccl_device float cubic_h1(float a)
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:77:10: error: no matching function for call to 'cubic_g0'
return cubic_g0(fy) * (g0x * ccl_gpu_tex_object_read_2D<T>(tex, x0, y0) +
^~~~~~~~
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:34:18: note: candidate function not viable: call to __device__ function from __host__ function
ccl_device float cubic_g0(float a)
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:79:10: error: no matching function for call to 'cubic_g1'
cubic_g1(fy) * (g0x * ccl_gpu_tex_object_read_2D<T>(tex, x0, y1) +
^~~~~~~~
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:38:18: note: candidate function not viable: call to __device__ function from __host__ function
ccl_device float cubic_g1(float a)
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:85:1: error: unknown type name '__noinline__'
ccl_device_noinline T
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/hip/compat.h:33:40: note: expanded from macro 'ccl_device_noinline'
#define ccl_device_noinline __device__ __noinline__
^
In file included from /usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/hip/kernel.cpp:12:
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:85:21: warning: variable templates are a C++14 extension [-Wc++14-extensions]
ccl_device_noinline T
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:85:22: error: expected ';' at end of declaration
ccl_device_noinline T
^
;
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:86:1: error: C++ requires a type specifier for all declarations
kernel_tex_image_interp_tricubic(ccl_global const TextureInfo &info, float x, float y, float z)
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:101:15: error: no matching function for call to 'cubic_g0'
float g0x = cubic_g0(fx);
^~~~~~~~
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:34:18: note: candidate function not viable: call to __device__ function from __host__ function
ccl_device float cubic_g0(float a)
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:102:15: error: no matching function for call to 'cubic_g1'
float g1x = cubic_g1(fx);
^~~~~~~~
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:38:18: note: candidate function not viable: call to __device__ function from __host__ function
ccl_device float cubic_g1(float a)
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:103:15: error: no matching function for call to 'cubic_g0'
float g0y = cubic_g0(fy);
^~~~~~~~
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:34:18: note: candidate function not viable: call to __device__ function from __host__ function
ccl_device float cubic_g0(float a)
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:104:15: error: no matching function for call to 'cubic_g1'
float g1y = cubic_g1(fy);
^~~~~~~~
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:38:18: note: candidate function not viable: call to __device__ function from __host__ function
ccl_device float cubic_g1(float a)
^
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:105:15: error: no matching function for call to 'cubic_g0'
float g0z = cubic_g0(fz);
^~~~~~~~
/usr/share/blender/3.2/scripts/addons/cycles/source/kernel/device/gpu/image.h:34:18: note: candidate function not viable: call to __device__ function from __host__ function
ccl_device float cubic_g0(float a)
^
fatal error: too many errors emitted, stopping now [-ferror-limit=]
2 warnings and 20 errors generated when compiling for gfx1030.
Failed to execute compilation command, see console for details.
Refer to the Cycles GPU rendering documentation for possible solutions:
https://docs.blender.org/manual/en/latest/render/cycles/gpu_rendering.html
Error: Failed to execute compilation command, see console for details.
https://github.com/rocm-arch/rocm-arch
That being said, I am willing to help...
https://github.com/rocm-arch/rocm-arch/discussions/673
I had the error with the __noinline__ some time ago compiling ROCm from AUR:
'The problem was that HIP defines __noinline__ but his is also defined in gcc, causing above issue.'
https://github.com/rocm-arch/rocm-arch/issues/780
Since HIP kernels works with blender binary from blender.org (and ROCm compiled from AUR), I wonder if the issue is due to blender from Arch repo missing or having miss-configured libraries/conf files.
The errors pops up when blender from Arch repo is trying to compile the HIP kernel.
Thanks
Blender from Arch repos now works with cycles hip on GPU (amdgpu), using ROCM 5.2 from AUR
pacman -Q --info blender
Name : blender
Version : 17:3.2.1-1
pacman -Q --info linux-zen
Name : linux-zen
Version : 5.18.12.zen1-1
However the amdgpu kernel module systematically crash when enabling rendering with cycles hip on GPU in the viewport:
[ 2674.170431] [drm:amdgpu_dm_atomic_commit_tail [amdgpu]] *ERROR* Waiting for fences timed out!
[ 2679.300365] [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx_0.0.0 timeout, signaled seq=502589, emitted seq=502591
[ 2679.300490] [drm:amdgpu_job_timedout [amdgpu]] *ERROR* Process information: process blender pid 10728 thread blender:cs0 pid 10762
[ 2679.300601] amdgpu 0000:0c:00.0: amdgpu: GPU reset begin!
[ 2679.300615] amdgpu: Failed to suspend process 0x800c
[ 2679.729744] amdgpu 0000:0c:00.0: [drm:amdgpu_ring_test_helper [amdgpu]] *ERROR* ring kiq_2.1.0 test failed (-110)
[ 2679.729846] [drm:gfx_v10_0_hw_fini [amdgpu]] *ERROR* KGQ disable failed
[ 2680.003239] amdgpu 0000:0c:00.0: [drm:amdgpu_ring_test_helper [amdgpu]] *ERROR* ring kiq_2.1.0 test failed (-110)
[ 2680.003337] [drm:gfx_v10_0_hw_fini [amdgpu]] *ERROR* KCQ disable failed
[ 2680.276602] [drm:gfx_v10_0_hw_fini [amdgpu]] *ERROR* failed to halt cp gfx
[ 2680.306668] [drm] free PSP TMR buffer
[ 2680.353268] CPU: 4 PID: 10932 Comm: kworker/u64:4 Not tainted 5.18.12-zen1-1-zen #1 bcafb1d6be6b7621cf2780efccdf702c74e989be
[ 2680.353271] Hardware name: System manufacturer System Product Name/ROG STRIX X570-F GAMING, BIOS 4204 02/24/2022
[ 2680.353272] Workqueue: amdgpu-reset-dev drm_sched_job_timedout [gpu_sched]
[ 2680.353275] Call Trace:
[ 2680.353277] <TASK>
[ 2680.353277] dump_stack_lvl+0x48/0x5d
[ 2680.353281] amdgpu_do_asic_reset+0x2a/0x470 [amdgpu 10164cb52bed15f4d975fcf74169b4340a4d1152]
[ 2680.353417] amdgpu_device_gpu_recover_imp.cold+0x5f9/0x9c2 [amdgpu 10164cb52bed15f4d975fcf74169b4340a4d1152]
[ 2680.353540] amdgpu_job_timedout+0x18c/0x1c0 [amdgpu 10164cb52bed15f4d975fcf74169b4340a4d1152]
[ 2680.353661] drm_sched_job_timedout+0x76/0x100 [gpu_sched 8469bfffda8dc0462de87eb377c7411bd418f6b4]
[ 2680.353664] process_one_work+0x255/0x410
[ 2680.353666] worker_thread+0x55/0x4d0
[ 2680.353668] ? process_one_work+0x410/0x410
[ 2680.353669] kthread+0x13f/0x160
[ 2680.353670] ? kthread_complete_and_exit+0x20/0x20
[ 2680.353672] ret_from_fork+0x22/0x30
[ 2680.353675] </TASK>
[ 2680.353679] amdgpu 0000:0c:00.0: amdgpu: MODE1 reset
[ 2680.353682] amdgpu 0000:0c:00.0: amdgpu: GPU mode1 reset
[ 2680.353749] amdgpu 0000:0c:00.0: amdgpu: GPU smu mode1 reset
[ 2680.858748] amdgpu 0000:0c:00.0: amdgpu: GPU reset succeeded, trying to resume
[ 2680.858903] [drm] PCIE GART of 512M enabled (table at 0x0000008000300000).
[ 2680.858924] [drm] VRAM is lost due to GPU reset!
[ 2680.858927] [drm] PSP is resuming...
[ 2680.937452] [drm] reserve 0xa00000 from 0x83fe000000 for PSP TMR
[ 2681.061309] amdgpu 0000:0c:00.0: amdgpu: SECUREDISPLAY: securedisplay ta ucode is not available
[ 2681.061311] amdgpu 0000:0c:00.0: amdgpu: SMU is resuming...
[ 2681.135255] amdgpu 0000:0c:00.0: amdgpu: SMU is resumed successfully!
[ 2681.136631] [drm] DMUB hardware initialized: version=0x0202000F
[ 2681.281744] [drm] kiq ring mec 2 pipe 1 q 0
[ 2681.289755] [drm] VCN decode and encode initialized successfully(under DPG Mode).
[ 2681.289932] [drm] JPEG decode initialized successfully.
[ 2681.289945] amdgpu 0000:0c:00.0: amdgpu: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
[ 2681.289946] amdgpu 0000:0c:00.0: amdgpu: ring comp_1.0.0 uses VM inv eng 1 on hub 0
[ 2681.289947] amdgpu 0000:0c:00.0: amdgpu: ring comp_1.1.0 uses VM inv eng 4 on hub 0
[ 2681.289948] amdgpu 0000:0c:00.0: amdgpu: ring comp_1.2.0 uses VM inv eng 5 on hub 0
[ 2681.289948] amdgpu 0000:0c:00.0: amdgpu: ring comp_1.3.0 uses VM inv eng 6 on hub 0
[ 2681.289949] amdgpu 0000:0c:00.0: amdgpu: ring comp_1.0.1 uses VM inv eng 7 on hub 0
[ 2681.289949] amdgpu 0000:0c:00.0: amdgpu: ring comp_1.1.1 uses VM inv eng 8 on hub 0
[ 2681.289950] amdgpu 0000:0c:00.0: amdgpu: ring comp_1.2.1 uses VM inv eng 9 on hub 0
[ 2681.289950] amdgpu 0000:0c:00.0: amdgpu: ring comp_1.3.1 uses VM inv eng 10 on hub 0
[ 2681.289951] amdgpu 0000:0c:00.0: amdgpu: ring kiq_2.1.0 uses VM inv eng 11 on hub 0
[ 2681.289952] amdgpu 0000:0c:00.0: amdgpu: ring sdma0 uses VM inv eng 12 on hub 0
[ 2681.289952] amdgpu 0000:0c:00.0: amdgpu: ring sdma1 uses VM inv eng 13 on hub 0
[ 2681.289953] amdgpu 0000:0c:00.0: amdgpu: ring sdma2 uses VM inv eng 14 on hub 0
[ 2681.289953] amdgpu 0000:0c:00.0: amdgpu: ring sdma3 uses VM inv eng 15 on hub 0
[ 2681.289954] amdgpu 0000:0c:00.0: amdgpu: ring vcn_dec_0 uses VM inv eng 0 on hub 1
[ 2681.289955] amdgpu 0000:0c:00.0: amdgpu: ring vcn_enc_0.0 uses VM inv eng 1 on hub 1
[ 2681.289955] amdgpu 0000:0c:00.0: amdgpu: ring vcn_enc_0.1 uses VM inv eng 4 on hub 1
[ 2681.289956] amdgpu 0000:0c:00.0: amdgpu: ring vcn_dec_1 uses VM inv eng 5 on hub 1
[ 2681.289957] amdgpu 0000:0c:00.0: amdgpu: ring vcn_enc_1.0 uses VM inv eng 6 on hub 1
[ 2681.289957] amdgpu 0000:0c:00.0: amdgpu: ring vcn_enc_1.1 uses VM inv eng 7 on hub 1
[ 2681.289958] amdgpu 0000:0c:00.0: amdgpu: ring jpeg_dec uses VM inv eng 8 on hub 1
[ 2681.297823] amdgpu 0000:0c:00.0: amdgpu: recover vram bo from shadow start
[ 2681.297847] amdgpu 0000:0c:00.0: amdgpu: recover vram bo from shadow done
[ 2681.297848] [drm] Skip scheduling IBs!
[ 2681.297849] [drm] Skip scheduling IBs!
[ 2681.297863] amdgpu 0000:0c:00.0: amdgpu: GPU reset(2) succeeded!
[ 2681.297865] [drm] Skip scheduling IBs!
[ 2681.297871] [drm] Skip scheduling IBs!
[ 2681.297877] [drm] Skip scheduling IBs!
[ 2681.297881] [drm] Skip scheduling IBs!
[ 2681.297886] [drm] Skip scheduling IBs!
[ 2681.297888] [drm] Skip scheduling IBs!
[ 2681.297893] [drm] Skip scheduling IBs!
[ 2681.297897] [drm] Skip scheduling IBs!
[ 2681.297899] [drm] Skip scheduling IBs!
[ 2681.297901] [drm] Skip scheduling IBs!
[ 2681.297903] [drm] Skip scheduling IBs!
[ 2681.297908] [drm] Skip scheduling IBs!
[ 2681.297910] [drm] Skip scheduling IBs!
[ 2681.297912] [drm] Skip scheduling IBs!
[ 2681.298054] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2681.303650] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2681.310562] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2681.310923] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2681.311277] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2681.311399] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2681.311628] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2681.814347] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2682.815560] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2683.816749] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2691.310349] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2691.311211] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2691.312923] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2691.313119] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2691.313636] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2691.313760] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2691.313874] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2691.313990] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2691.314656] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2691.315144] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2701.329418] amdgpu_cs_ioctl: 6 callbacks suppressed
[ 2701.329420] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2701.329664] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2701.333369] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2701.333498] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2701.333873] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2701.334229] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2701.334341] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2701.334626] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2701.334816] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
[ 2701.366960] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!
attaching dmesg and journalctl relevant error messages
Thanks
the crash of amdgpu kernel module was due to installed AMD Radeon ProRender Blender plugin.
It seems the plugin brings in its own hip libraries, causing conflicts within blender.
Removed plugin -> no more crash
Now Archlinux Blender package correctly uses ROCM 5.2.0 from AUR and allows to use Cycles+HIP on GPU and CPU
Thanks