Blender 2.83 crashing and lagging when rendering the viewport (cycles). #77466

New Issue

Matheus · 2020-06-05T21:43:45+02:00

Matheus commented

2020-06-05 21:43:45 +02:00

System Information
Operating system: Arch Linux
Graphics card: GTX 1050 ti
Driver: nvidia-lts-1:440.82-19

Blender Version
Broken: Blender 2.83 LTS (Tesded with and same issue with blender from blender.org website, Snap package and from the Arch Community Repository)
Worked: There was no problems in 2.82a
Suspected culprit: d712f1f83a

Description of the error
The render view (the viewport typically when working with shaders) causes the interface to get extremely lag and sometimes totally stuck when rendering the viewport. Didn't find any issue rendering 250 frames of animation, but most of the times I'm working with the nodes shades setup and I change some nodes making the viewport start rendering, using cycles, it gets stuck for something like 10 seconds and then it can crash or just continue.

But I haven't been able to use blender for more than 5 minutes without a crash with this problem and lagging even when I simply change the view position on the viewport. I have tested many other files and they all have this same problem. Also happened when using CPU rendering instead of GPU.

I tried using the option debug-all but I don't think there is any hint there at all.

Shader name: default_empty
Number of SVM nodes: 3
Peak stack usage:    0
Time (in seconds):
Finalize:            0.000000
  Surface:           0.000002
  Bump:              0.000000
  Volume:            0.000001
  Displacement:      0.000001
Generate:            0.000004
Total:               0.000005
I0605 16:02:45.568426 1404890 svm.cpp:66] Compilation summary:
Shader name: shader
Number of SVM nodes: 5
Peak stack usage:    0
Time (in seconds):
Finalize:            0.000000
  Surface:           0.000004
  Bump:              0.000000
  Volume:            0.000000
  Displacement:      0.000001
Generate:            0.000005
Total:               0.000005
I0605 16:02:45.568414 1404906 svm.cpp:66] Compilation summary:
Shader name: shader
Number of SVM nodes: 5
Peak stack usage:    0
Time (in seconds):
Finalize:            0.000001
  Surface:           0.000003
  Bump:              0.000000
  Volume:            0.000002
  Displacement:      0.000001
Generate:            0.000006
Total:               0.000007
I0605 16:02:45.568498 1404889 constant_fold.cpp:46] Folding Base Height::Value to constant (0.2).
I0605 16:02:45.568521 1404889 constant_fold.cpp:46] Folding Cone Height::Value to constant (1.4).
I0605 16:02:45.568538 1404889 constant_fold.cpp:46] Folding Radius::Value to constant (0.45).
I0605 16:02:45.568555 1404889 constant_fold.cpp:46] Folding SideLengh::Value to constant (0.9).
I0605 16:02:45.568574 1404889 constant_fold.cpp:46] Folding punchin::Value to constant (-2.15).
I0605 16:02:45.568588 1404889 constant_fold.cpp:46] Folding Margin::Value to constant (3.8).
I0605 16:02:45.568603 1404889 constant_fold.cpp:46] Folding NoiseNoise::Value to constant (-0.495).
I0605 16:02:45.568615 1404889 constant_fold.cpp:46] Folding stripe::Value to constant (-0.04).
I0605 16:02:45.568629 1404889 constant_fold.cpp:46] Folding midlevel::Value to constant (0.08).
I0605 16:02:45.568642 1404889 constant_fold.cpp:46] Folding Displacement.001::Value to constant (1).
I0605 16:02:45.568656 1404889 constant_fold.cpp:46] Folding DirtStrenght::Value to constant (2.7).
I0605 16:02:45.568673 1404889 constant_fold.cpp:58] Folding Combine XYZ.001::Vector to constant (0.45, 0.45, 0).
I0605 16:02:45.568691 1404889 constant_fold.cpp:58] Folding Combine XYZ.001::Vector to constant (0.45, 0.45, 0).
I0605 16:02:45.568706 1404889 constant_fold.cpp:46] Folding Math.001::Value to constant (-0.9).
I0605 16:02:45.568719 1404889 constant_fold.cpp:46] Folding Math.004::Value to constant (4.7).
I0605 16:02:45.568735 1404889 constant_fold.cpp:58] Folding Vector Math.005::Vector to constant (0.2475, 0.2475, 0.2475).
I0605 16:02:45.568754 1404889 constant_fold.cpp:58] Folding Combine XYZ::Vector to constant (-0.9, -0.9, 0).
I0605 16:02:45.568773 1404889 constant_fold.cpp:46] Folding Math.001::Value to constant (-4.7).
I0605 16:02:45.568795 1404889 constant_fold.cpp:58] Folding Vector Math.005::Vector to constant (-0.45, -0.45, 0).
I0605 16:02:45.568812 1404889 constant_fold.cpp:58] Folding Combine XYZ::Vector to constant (-4.7, -4.7, 0).
I0605 16:02:45.568837 1404889 constant_fold.cpp:58] Folding Vector Math.005::Vector to constant (-4.25, -4.25, 0).
I0605 16:02:45.568871 1404889 constant_fold.cpp:113] Folding Math.002::Value to socket convert_color_to_float::value_float.
I0605 16:02:45.568920 1404889 constant_fold.cpp:113] Folding Mix Shader.001::Closure to socket Mix Shader::Closure.
I0605 16:02:45.569067 1404889 graph.cpp:675] Deduplicated 9 nodes.
I0605 16:02:45.570071 1404889 svm.cpp:66] Compilation summary:
Shader name: cone
Number of SVM nodes: 4000
Peak stack usage:    61
Time (in seconds):
Finalize:            0.000878
  Surface:           0.000478
  Bump:              0.000175
  Volume:            0.000035
  Displacement:      0.000051
Generate:            0.000739
Total:               0.001618
I0605 16:02:45.570232 1404906 device_cuda_impl.cpp:819] Buffer allocate: __svm_nodes, 64,640 bytes. (63.12K) in device memory
I0605 16:02:45.570369 1404906 device_cuda_impl.cpp:819] Buffer allocate: __shaders, 256 bytes. (256) in device memory
I0605 16:02:45.570436 1404906 svm.cpp:158] Shader manager updated 8 shaders in 0.0022831 seconds.
I0605 16:02:45.570477 1404906 object.cpp:676] Total 1 objects.
I0605 16:02:45.570567 1404906 device_cuda_impl.cpp:819] Buffer allocate: __objects, 176 bytes. (176) in device memory
I0605 16:02:45.570607 1404906 geometry.cpp:1231] Total 1 meshes.
I0605 16:02:45.576633 1404906 geometry.cpp:1307] Updating images used for true displacement.
I0605 16:02:45.576761 1404906 device_cuda_impl.cpp:819] Buffer allocate: __object_flag, 4 bytes. (4) in device memory
I0605 16:02:45.576845 1404906 device_cuda_impl.cpp:819] Buffer allocate: __object_volume_step, 4 bytes. (4) in device memory
UI_menutype_draw: opening menu "NODE_MT_editor_menus"
I0605 16:02:45.577569 1404906 device_cuda_impl.cpp:819] Buffer allocate: __tri_shader, 204,544 bytes. (199.75K) in device memory
I0605 16:02:45.577716 1404906 device_cuda_impl.cpp:819] Buffer allocate: __tri_vnormal, 499,728 bytes. (488.02K) in device memory
I0605 16:02:45.577925 1404906 device_cuda_impl.cpp:819] Buffer allocate: __tri_vindex, 818,176 bytes. (799.00K) in device memory
I0605 16:02:45.578192 1404906 device_cuda_impl.cpp:819] Buffer allocate: __tri_patch, 204,544 bytes. (199.75K) in device memory
I0605 16:02:45.578296 1404906 device_cuda_impl.cpp:819] Buffer allocate: __tri_patch_uv, 249,864 bytes. (244.01K) in device memory
I0605 16:02:45.578438 1404906 device_cuda_impl.cpp:819] Buffer allocate: __patches, 8,192 bytes. (8.00K) in device memory
I0605 16:02:45.579053 1404906 device_cuda_impl.cpp:819] Buffer allocate: __prim_tri_verts, 2,454,528 bytes. (2.34M) in device memory
I0605 16:02:45.579543 1404906 device_cuda_impl.cpp:819] Buffer allocate: __attributes_map, 96 bytes. (96) in device memory
I0605 16:02:45.579593 1404906 device_cuda_impl.cpp:819] Buffer allocate: __attributes_float3, 9,248 bytes. (9.03K) in device memory
I0605 16:02:45.580037 1404906 device_cuda_impl.cpp:819] Buffer allocate: displace_output, 495,104 bytes. (483.50K) in device memory
I0605 16:02:45.580083 1404906 device_cuda_impl.cpp:819] Buffer allocate: displace_input, 499,728 bytes. (488.02K) in device memory
UI_menutype_draw: opening menu "TIME_MT_editor_menus"
[1]    1404486 abort (core dumped)  ~/programs/blender-2.83.0-linux64/blender --debug-all

This is another crash:


(...)

wm_event_do_handlers: Handling event
wmEvent  type:220 / RET, val:2 / RELEASE,
         prev_type:154 / NUMPAD_4, prev_val:2 / RELEASE,
         shift:0, ctrl:0, alt:0, oskey:0, keymodifier:0, is_repeat:0,
         mouse:(1841,556), ascii:'', utf8:'', pointer:0x7f4ac0f5d3f8
UI_menutype_draw: opening menu "VIEW3D_MT_editor_menus"

wm_event_do_handlers: Handling event
wmEvent  type:11 / WHEELDOWNMOUSE, val:1 / PRESS,
         prev_type:220 / RET, prev_val:1 / PRESS,
         shift:0, ctrl:0, alt:0, oskey:0, keymodifier:0, is_repeat:0,
         mouse:(1271,660), ascii:'', utf8:'', pointer:0x7f4abfe5bdf8
I0605 03:34:47.383342 1303666 camera.cpp:511] Camera is outside of the volume.

wm_event_do_handlers: Handling event
wmEvent  type:11 / WHEELDOWNMOUSE, val:1 / PRESS,
         prev_type:220 / RET, prev_val:1 / PRESS,
         shift:0, ctrl:0, alt:0, oskey:0, keymodifier:0, is_repeat:0,
         mouse:(1271,660), ascii:'', utf8:'', pointer:0x7f4abfe5bdf8
I0605 03:34:47.507941 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory
I0605 03:34:47.521967 1297814 device_cuda_impl.cpp:819] Buffer allocate: RenderBuffers, 11,419,200 bytes. (10.89M) in device memory
I0605 03:34:47.522083 1303666 camera.cpp:511] Camera is outside of the volume.
I0605 03:34:47.522354 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory
I0605 03:34:47.536906 1303666 camera.cpp:511] Camera is outside of the volume.
I0605 03:34:47.537266 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory
I0605 03:34:47.560497 1303666 camera.cpp:511] Camera is outside of the volume.
I0605 03:34:47.560959 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory
I0605 03:34:47.614507 1303666 camera.cpp:511] Camera is outside of the volume.
I0605 03:34:47.615080 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory
I0605 03:34:47.756260 1303666 camera.cpp:511] Camera is outside of the volume.
I0605 03:34:47.757364 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory
I0605 03:34:47.902868 1303666 camera.cpp:511] Camera is outside of the volume.
I0605 03:34:47.903980 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory

wm_event_do_handlers: Handling event
wmEvent  type:10 / WHEELUPMOUSE, val:1 / PRESS,
         prev_type:220 / RET, prev_val:1 / PRESS,
         shift:0, ctrl:0, alt:0, oskey:0, keymodifier:0, is_repeat:0,
         mouse:(1124,699), ascii:'', utf8:'', pointer:0x7f4abfe593f8
I0605 03:34:48.050220 1297814 device_cuda_impl.cpp:819] Buffer allocate: RenderBuffers, 91,353,600 bytes. (87.12M) in device memory
I0605 03:34:48.050329 1303666 camera.cpp:511] Camera is outside of the volume.
I0605 03:34:48.052258 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory
I0605 03:34:48.066417 1303666 camera.cpp:511] Camera is outside of the volume.
I0605 03:34:48.067525 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory
I0605 03:34:48.092654 1303666 camera.cpp:511] Camera is outside of the volume.
I0605 03:34:48.093850 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory
I0605 03:34:48.151859 1303666 camera.cpp:511] Camera is outside of the volume.
I0605 03:34:48.153138 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory

wm_event_do_handlers: Handling event
wmEvent  type:2 / MIDDLEMOUSE, val:1 / PRESS,
         prev_type:220 / RET, prev_val:1 / PRESS,
         shift:0, ctrl:0, alt:0, oskey:0, keymodifier:0, is_repeat:0,
         mouse:(1144,703), ascii:'', utf8:'', pointer:0x7f4abfe59338
I0605 03:34:48.490563 1303666 camera.cpp:511] Camera is outside of the volume.
I0605 03:34:48.491914 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory
I0605 03:34:48.510012 1303666 camera.cpp:511] Camera is outside of the volume.
I0605 03:34:48.511392 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory
I0605 03:34:48.543624 1303666 camera.cpp:511] Camera is outside of the volume.
I0605 03:34:48.545732 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory
Skipping auto-save, modal operator running, retrying in ten seconds...

wm_event_do_handlers: Handling event
wmEvent  type:2 / MIDDLEMOUSE, val:2 / RELEASE,
         prev_type:220 / RET, prev_val:2 / RELEASE,
         shift:0, ctrl:0, alt:0, oskey:0, keymodifier:0, is_repeat:0,
         mouse:(1221,703), ascii:'', utf8:'', pointer:0x7f4abfe930b8
I0605 03:34:59.571421 1297814 device_cuda_impl.cpp:819] Buffer allocate: RenderBuffers, 11,419,200 bytes. (10.89M) in device memory
I0605 03:34:59.750665 1303666 camera.cpp:511] Camera is outside of the volume.
I0605 03:34:59.775646 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory
I0605 03:35:00.032346 1303666 camera.cpp:511] Camera is outside of the volume.
I0605 03:35:00.032780 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory
[1]    1297814 abort (core dumped)  ~/programs/blender-2.83.0-linux64/blender --debug-all

Those lines like this:

I0605 03:34:59.750665 1303666 camera.cpp:511] Camera is outside of the volume.
I0605 03:34:59.775646 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory

They always come when the interface gets stuck when the preview renderer is starting.

**System Information** Operating system: Arch Linux Graphics card: GTX 1050 ti Driver: nvidia-lts-1:440.82-19 **Blender Version** Broken: Blender 2.83 LTS (Tesded with and same issue with blender from blender.org website, Snap package and from the Arch Community Repository) Worked: There was no problems in 2.82a Suspected culprit: d712f1f83a **Description of the error** The render view (the viewport typically when working with shaders) causes the interface to get extremely lag and sometimes totally stuck when rendering the viewport. Didn't find any issue rendering 250 frames of animation, but most of the times I'm working with the nodes shades setup and I change some nodes making the viewport start rendering, using cycles, it gets stuck for something like 10 seconds and then it can crash or just continue. But I haven't been able to use blender for more than 5 minutes without a crash with this problem and lagging even when I simply change the view position on the viewport. I have tested many other files and they all have this same problem. Also happened when using CPU rendering instead of GPU. I tried using the option debug-all but I don't think there is any hint there at all. ``` Shader name: default_empty Number of SVM nodes: 3 Peak stack usage: 0 Time (in seconds): Finalize: 0.000000 Surface: 0.000002 Bump: 0.000000 Volume: 0.000001 Displacement: 0.000001 Generate: 0.000004 Total: 0.000005 I0605 16:02:45.568426 1404890 svm.cpp:66] Compilation summary: Shader name: shader Number of SVM nodes: 5 Peak stack usage: 0 Time (in seconds): Finalize: 0.000000 Surface: 0.000004 Bump: 0.000000 Volume: 0.000000 Displacement: 0.000001 Generate: 0.000005 Total: 0.000005 I0605 16:02:45.568414 1404906 svm.cpp:66] Compilation summary: Shader name: shader Number of SVM nodes: 5 Peak stack usage: 0 Time (in seconds): Finalize: 0.000001 Surface: 0.000003 Bump: 0.000000 Volume: 0.000002 Displacement: 0.000001 Generate: 0.000006 Total: 0.000007 I0605 16:02:45.568498 1404889 constant_fold.cpp:46] Folding Base Height::Value to constant (0.2). I0605 16:02:45.568521 1404889 constant_fold.cpp:46] Folding Cone Height::Value to constant (1.4). I0605 16:02:45.568538 1404889 constant_fold.cpp:46] Folding Radius::Value to constant (0.45). I0605 16:02:45.568555 1404889 constant_fold.cpp:46] Folding SideLengh::Value to constant (0.9). I0605 16:02:45.568574 1404889 constant_fold.cpp:46] Folding punchin::Value to constant (-2.15). I0605 16:02:45.568588 1404889 constant_fold.cpp:46] Folding Margin::Value to constant (3.8). I0605 16:02:45.568603 1404889 constant_fold.cpp:46] Folding NoiseNoise::Value to constant (-0.495). I0605 16:02:45.568615 1404889 constant_fold.cpp:46] Folding stripe::Value to constant (-0.04). I0605 16:02:45.568629 1404889 constant_fold.cpp:46] Folding midlevel::Value to constant (0.08). I0605 16:02:45.568642 1404889 constant_fold.cpp:46] Folding Displacement.001::Value to constant (1). I0605 16:02:45.568656 1404889 constant_fold.cpp:46] Folding DirtStrenght::Value to constant (2.7). I0605 16:02:45.568673 1404889 constant_fold.cpp:58] Folding Combine XYZ.001::Vector to constant (0.45, 0.45, 0). I0605 16:02:45.568691 1404889 constant_fold.cpp:58] Folding Combine XYZ.001::Vector to constant (0.45, 0.45, 0). I0605 16:02:45.568706 1404889 constant_fold.cpp:46] Folding Math.001::Value to constant (-0.9). I0605 16:02:45.568719 1404889 constant_fold.cpp:46] Folding Math.004::Value to constant (4.7). I0605 16:02:45.568735 1404889 constant_fold.cpp:58] Folding Vector Math.005::Vector to constant (0.2475, 0.2475, 0.2475). I0605 16:02:45.568754 1404889 constant_fold.cpp:58] Folding Combine XYZ::Vector to constant (-0.9, -0.9, 0). I0605 16:02:45.568773 1404889 constant_fold.cpp:46] Folding Math.001::Value to constant (-4.7). I0605 16:02:45.568795 1404889 constant_fold.cpp:58] Folding Vector Math.005::Vector to constant (-0.45, -0.45, 0). I0605 16:02:45.568812 1404889 constant_fold.cpp:58] Folding Combine XYZ::Vector to constant (-4.7, -4.7, 0). I0605 16:02:45.568837 1404889 constant_fold.cpp:58] Folding Vector Math.005::Vector to constant (-4.25, -4.25, 0). I0605 16:02:45.568871 1404889 constant_fold.cpp:113] Folding Math.002::Value to socket convert_color_to_float::value_float. I0605 16:02:45.568920 1404889 constant_fold.cpp:113] Folding Mix Shader.001::Closure to socket Mix Shader::Closure. I0605 16:02:45.569067 1404889 graph.cpp:675] Deduplicated 9 nodes. I0605 16:02:45.570071 1404889 svm.cpp:66] Compilation summary: Shader name: cone Number of SVM nodes: 4000 Peak stack usage: 61 Time (in seconds): Finalize: 0.000878 Surface: 0.000478 Bump: 0.000175 Volume: 0.000035 Displacement: 0.000051 Generate: 0.000739 Total: 0.001618 I0605 16:02:45.570232 1404906 device_cuda_impl.cpp:819] Buffer allocate: __svm_nodes, 64,640 bytes. (63.12K) in device memory I0605 16:02:45.570369 1404906 device_cuda_impl.cpp:819] Buffer allocate: __shaders, 256 bytes. (256) in device memory I0605 16:02:45.570436 1404906 svm.cpp:158] Shader manager updated 8 shaders in 0.0022831 seconds. I0605 16:02:45.570477 1404906 object.cpp:676] Total 1 objects. I0605 16:02:45.570567 1404906 device_cuda_impl.cpp:819] Buffer allocate: __objects, 176 bytes. (176) in device memory I0605 16:02:45.570607 1404906 geometry.cpp:1231] Total 1 meshes. I0605 16:02:45.576633 1404906 geometry.cpp:1307] Updating images used for true displacement. I0605 16:02:45.576761 1404906 device_cuda_impl.cpp:819] Buffer allocate: __object_flag, 4 bytes. (4) in device memory I0605 16:02:45.576845 1404906 device_cuda_impl.cpp:819] Buffer allocate: __object_volume_step, 4 bytes. (4) in device memory UI_menutype_draw: opening menu "NODE_MT_editor_menus" I0605 16:02:45.577569 1404906 device_cuda_impl.cpp:819] Buffer allocate: __tri_shader, 204,544 bytes. (199.75K) in device memory I0605 16:02:45.577716 1404906 device_cuda_impl.cpp:819] Buffer allocate: __tri_vnormal, 499,728 bytes. (488.02K) in device memory I0605 16:02:45.577925 1404906 device_cuda_impl.cpp:819] Buffer allocate: __tri_vindex, 818,176 bytes. (799.00K) in device memory I0605 16:02:45.578192 1404906 device_cuda_impl.cpp:819] Buffer allocate: __tri_patch, 204,544 bytes. (199.75K) in device memory I0605 16:02:45.578296 1404906 device_cuda_impl.cpp:819] Buffer allocate: __tri_patch_uv, 249,864 bytes. (244.01K) in device memory I0605 16:02:45.578438 1404906 device_cuda_impl.cpp:819] Buffer allocate: __patches, 8,192 bytes. (8.00K) in device memory I0605 16:02:45.579053 1404906 device_cuda_impl.cpp:819] Buffer allocate: __prim_tri_verts, 2,454,528 bytes. (2.34M) in device memory I0605 16:02:45.579543 1404906 device_cuda_impl.cpp:819] Buffer allocate: __attributes_map, 96 bytes. (96) in device memory I0605 16:02:45.579593 1404906 device_cuda_impl.cpp:819] Buffer allocate: __attributes_float3, 9,248 bytes. (9.03K) in device memory I0605 16:02:45.580037 1404906 device_cuda_impl.cpp:819] Buffer allocate: displace_output, 495,104 bytes. (483.50K) in device memory I0605 16:02:45.580083 1404906 device_cuda_impl.cpp:819] Buffer allocate: displace_input, 499,728 bytes. (488.02K) in device memory UI_menutype_draw: opening menu "TIME_MT_editor_menus" [1] 1404486 abort (core dumped) ~/programs/blender-2.83.0-linux64/blender --debug-all ``` This is another crash: ``` (...) wm_event_do_handlers: Handling event wmEvent type:220 / RET, val:2 / RELEASE, prev_type:154 / NUMPAD_4, prev_val:2 / RELEASE, shift:0, ctrl:0, alt:0, oskey:0, keymodifier:0, is_repeat:0, mouse:(1841,556), ascii:'', utf8:'', pointer:0x7f4ac0f5d3f8 UI_menutype_draw: opening menu "VIEW3D_MT_editor_menus" wm_event_do_handlers: Handling event wmEvent type:11 / WHEELDOWNMOUSE, val:1 / PRESS, prev_type:220 / RET, prev_val:1 / PRESS, shift:0, ctrl:0, alt:0, oskey:0, keymodifier:0, is_repeat:0, mouse:(1271,660), ascii:'', utf8:'', pointer:0x7f4abfe5bdf8 I0605 03:34:47.383342 1303666 camera.cpp:511] Camera is outside of the volume. wm_event_do_handlers: Handling event wmEvent type:11 / WHEELDOWNMOUSE, val:1 / PRESS, prev_type:220 / RET, prev_val:1 / PRESS, shift:0, ctrl:0, alt:0, oskey:0, keymodifier:0, is_repeat:0, mouse:(1271,660), ascii:'', utf8:'', pointer:0x7f4abfe5bdf8 I0605 03:34:47.507941 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory I0605 03:34:47.521967 1297814 device_cuda_impl.cpp:819] Buffer allocate: RenderBuffers, 11,419,200 bytes. (10.89M) in device memory I0605 03:34:47.522083 1303666 camera.cpp:511] Camera is outside of the volume. I0605 03:34:47.522354 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory I0605 03:34:47.536906 1303666 camera.cpp:511] Camera is outside of the volume. I0605 03:34:47.537266 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory I0605 03:34:47.560497 1303666 camera.cpp:511] Camera is outside of the volume. I0605 03:34:47.560959 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory I0605 03:34:47.614507 1303666 camera.cpp:511] Camera is outside of the volume. I0605 03:34:47.615080 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory I0605 03:34:47.756260 1303666 camera.cpp:511] Camera is outside of the volume. I0605 03:34:47.757364 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory I0605 03:34:47.902868 1303666 camera.cpp:511] Camera is outside of the volume. I0605 03:34:47.903980 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory wm_event_do_handlers: Handling event wmEvent type:10 / WHEELUPMOUSE, val:1 / PRESS, prev_type:220 / RET, prev_val:1 / PRESS, shift:0, ctrl:0, alt:0, oskey:0, keymodifier:0, is_repeat:0, mouse:(1124,699), ascii:'', utf8:'', pointer:0x7f4abfe593f8 I0605 03:34:48.050220 1297814 device_cuda_impl.cpp:819] Buffer allocate: RenderBuffers, 91,353,600 bytes. (87.12M) in device memory I0605 03:34:48.050329 1303666 camera.cpp:511] Camera is outside of the volume. I0605 03:34:48.052258 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory I0605 03:34:48.066417 1303666 camera.cpp:511] Camera is outside of the volume. I0605 03:34:48.067525 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory I0605 03:34:48.092654 1303666 camera.cpp:511] Camera is outside of the volume. I0605 03:34:48.093850 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory I0605 03:34:48.151859 1303666 camera.cpp:511] Camera is outside of the volume. I0605 03:34:48.153138 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory wm_event_do_handlers: Handling event wmEvent type:2 / MIDDLEMOUSE, val:1 / PRESS, prev_type:220 / RET, prev_val:1 / PRESS, shift:0, ctrl:0, alt:0, oskey:0, keymodifier:0, is_repeat:0, mouse:(1144,703), ascii:'', utf8:'', pointer:0x7f4abfe59338 I0605 03:34:48.490563 1303666 camera.cpp:511] Camera is outside of the volume. I0605 03:34:48.491914 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory I0605 03:34:48.510012 1303666 camera.cpp:511] Camera is outside of the volume. I0605 03:34:48.511392 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory I0605 03:34:48.543624 1303666 camera.cpp:511] Camera is outside of the volume. I0605 03:34:48.545732 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory Skipping auto-save, modal operator running, retrying in ten seconds... wm_event_do_handlers: Handling event wmEvent type:2 / MIDDLEMOUSE, val:2 / RELEASE, prev_type:220 / RET, prev_val:2 / RELEASE, shift:0, ctrl:0, alt:0, oskey:0, keymodifier:0, is_repeat:0, mouse:(1221,703), ascii:'', utf8:'', pointer:0x7f4abfe930b8 I0605 03:34:59.571421 1297814 device_cuda_impl.cpp:819] Buffer allocate: RenderBuffers, 11,419,200 bytes. (10.89M) in device memory I0605 03:34:59.750665 1303666 camera.cpp:511] Camera is outside of the volume. I0605 03:34:59.775646 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory I0605 03:35:00.032346 1303666 camera.cpp:511] Camera is outside of the volume. I0605 03:35:00.032780 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory [1] 1297814 abort (core dumped) ~/programs/blender-2.83.0-linux64/blender --debug-all ``` Those lines like this: ``` I0605 03:34:59.750665 1303666 camera.cpp:511] Camera is outside of the volume. I0605 03:34:59.775646 1303635 device_cuda_impl.cpp:819] Buffer allocate: work_tiles, 40 bytes. (40) in device memory ``` They always come when the interface gets stuck when the preview renderer is starting.

Matheus commented

2020-06-05 21:43:45 +02:00

Added subscriber: @mattffly

blender-admin commented

2020-06-05 21:43:45 +02:00

#81908 was marked as duplicate of this issue

blender-admin commented

2020-06-05 21:43:45 +02:00

#80073 was marked as duplicate of this issue

Germano Cavalcante commented

2020-06-08 15:51:20 +02:00

Added subscriber: @mano-wii

Germano Cavalcante commented

2020-06-08 15:51:20 +02:00

Are there any steps to reproduce the problem?

Matheus commented

2020-06-08 20:09:05 +02:00

In #77466#948548, @mano-wii wrote:
Are there any steps to reproduce the problem?

Yeah: be on linux (maybe specially on arch but i don't think so), use nvidia 440.82-19 (but I was having the same issue on another drivers as well). Start seeing any rendered scene with cycles and blender becomes super lagy to work with (which also happened with evee) and it will eventually crashes (only crashes with cycles).

There is not much to it, just conventional usage changing anything on the materials or even moving on the viewport. The only pattern I found is that it crashes or lags while rendering the viewport.

> In #77466#948548, @mano-wii wrote: > Are there any steps to reproduce the problem? Yeah: be on linux (maybe specially on arch but i don't think so), use nvidia 440.82-19 (but I was having the same issue on another drivers as well). Start seeing any rendered scene with cycles and blender becomes super lagy to work with (which also happened with evee) and it will eventually crashes (only crashes with cycles). There is not much to it, just conventional usage changing anything on the materials or even moving on the viewport. The only pattern I found is that it crashes or lags while rendering the viewport.

Matheus commented

2020-06-09 05:56:17 +02:00

Updates. I changed from the kernel-lts to kernel (5.6.1) and added a file like this:

$ cat /etc/modprobe.d/i915.conf
# enables firmware update of intel gpu
options i915 enable_guc=2 enable_psr=0

Since someone recommended me to add that option due to the error:

[J

un 9 00:27] i915 0000:00:02.0: GPU HANG: ecode 9:1:0x00000000, hang on rcs0
[  +0.001009] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ +25.990686] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ +18.011293] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ +26.344984] i915 0000:00:02.0: Resetting rcs0 for preemption time out
[  +0.000005] i915 0000:00:02.0: blender[21226] context reset due to GPU hang
[  +0.016295] i915 0000:00:02.0: GPU HANG: ecode 9:1:87f99eb9, in blender [21226]
[  +0.000003] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
[  +0.000000] Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new.
[  +0.000000] Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details.
[  +0.000001] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
[  +0.000000] The GPU crash dump is required to analyze GPU hangs, so please always attach it.
[  +0.000001] GPU crash dump saved to /sys/class/drm/card0/error
[  +0.890507] i915 0000:00:02.0: Resetting rcs0 for preemption time out
[  +0.000006] i915 0000:00:02.0: blender[21226] context reset due to GPU hang
[  +0.011065] i915 0000:00:02.0: GPU HANG: ecode 9:1:85dfbfff, in blender [21226]
[Jun 8 23:58] Asynchronous wait on fence 0000:00:02.0:kwin_x11[2636]:3ad4 timed out (hint:intel_atomic_commit_ready+0x0/0x50 [i915])
[  +1.734737] i915 0000:00:02.0: Resetting rcs0 for preemption time out
[  +0.000009] i915 0000:00:02.0: blender[21226] context reset due to GPU hang
[  +0.026699] audit: type=1701 audit(1591671483.743:362): auid=1000 uid=1000 gid=1000 ses=4 pid=21226 comm="blender" exe="/home/matheus/programs/blender-2.83.0-linux64/blender" sig=6 res=1

When blender is lagging and/or crashing. The result was that the whole Xorg session frost with the kernel (5.6) and the only solution was to kill Xorg, instead of just blender crashing. I switched back to the kernel-lts and blender 2.82 is working fine but 2.83 still crashes. I also kept the kernel option: "enable_psr=0".

Updates. I changed from the kernel-lts to kernel (5.6.1) and added a file like this: ``` $ cat /etc/modprobe.d/i915.conf # enables firmware update of intel gpu options i915 enable_guc=2 enable_psr=0 ``` Since someone recommended me to add that option due to the error: [J ``` un 9 00:27] i915 0000:00:02.0: GPU HANG: ecode 9:1:0x00000000, hang on rcs0 [ +0.001009] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0 [ +25.990686] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0 [ +18.011293] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0 [ +26.344984] i915 0000:00:02.0: Resetting rcs0 for preemption time out [ +0.000005] i915 0000:00:02.0: blender[21226] context reset due to GPU hang [ +0.016295] i915 0000:00:02.0: GPU HANG: ecode 9:1:87f99eb9, in blender [21226] [ +0.000003] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [ +0.000000] Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new. [ +0.000000] Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details. [ +0.000001] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [ +0.000000] The GPU crash dump is required to analyze GPU hangs, so please always attach it. [ +0.000001] GPU crash dump saved to /sys/class/drm/card0/error [ +0.890507] i915 0000:00:02.0: Resetting rcs0 for preemption time out [ +0.000006] i915 0000:00:02.0: blender[21226] context reset due to GPU hang [ +0.011065] i915 0000:00:02.0: GPU HANG: ecode 9:1:85dfbfff, in blender [21226] [Jun 8 23:58] Asynchronous wait on fence 0000:00:02.0:kwin_x11[2636]:3ad4 timed out (hint:intel_atomic_commit_ready+0x0/0x50 [i915]) [ +1.734737] i915 0000:00:02.0: Resetting rcs0 for preemption time out [ +0.000009] i915 0000:00:02.0: blender[21226] context reset due to GPU hang [ +0.026699] audit: type=1701 audit(1591671483.743:362): auid=1000 uid=1000 gid=1000 ses=4 pid=21226 comm="blender" exe="/home/matheus/programs/blender-2.83.0-linux64/blender" sig=6 res=1 ``` When blender is lagging and/or crashing. The result was that the whole Xorg session frost with the kernel (5.6) and the only solution was to kill Xorg, instead of just blender crashing. I switched back to the kernel-lts and blender 2.82 is working fine but 2.83 still crashes. I also kept the kernel option: "enable_psr=0".

Matheus commented

2020-06-10 02:30:41 +02:00

Crash dump

Jun 09 21:29:11 arch-matheus systemd-coredump[440018]: Process 438978 (blender) of user 1000 dumped core.

                                                       Stack trace of thread 438978:
                                                       - 0  0x00007f0a9d30e355 raise (libc.so.6 + 0x3c355)
                                                       - 1  0x00007f0a9d2f790b abort (libc.so.6 + 0x2590b)
                                                       - 2  0x00007f0a8170214c n/a (iris_dri.so + 0xfb14c)
                                                       - 3  0x00007f0a820017bb n/a (iris_dri.so + 0x9fa7bb)
                                                       - 4  0x00007f0a8173c12a n/a (iris_dri.so + 0x13512a)
                                                       - 5  0x0000000000ffef5c n/a (/home/matheus/programs/blender-2.83.0-linux64/blender + 0xbfef5c)

Crash dump ``` Jun 09 21:29:11 arch-matheus systemd-coredump[440018]: Process 438978 (blender) of user 1000 dumped core. Stack trace of thread 438978: - 0 0x00007f0a9d30e355 raise (libc.so.6 + 0x3c355) - 1 0x00007f0a9d2f790b abort (libc.so.6 + 0x2590b) - 2 0x00007f0a8170214c n/a (iris_dri.so + 0xfb14c) - 3 0x00007f0a820017bb n/a (iris_dri.so + 0x9fa7bb) - 4 0x00007f0a8173c12a n/a (iris_dri.so + 0x13512a) - 5 0x0000000000ffef5c n/a (/home/matheus/programs/blender-2.83.0-linux64/blender + 0xbfef5c) ```

Matheus commented

2020-06-10 02:53:41 +02:00

I just made a test running blender with prime-run (the arch linux package for running a application with the dedicated gpu) and blender 2.83 worked flawlessly. I could even use optix and it worked and my gpu is theoretically not supported. (CYCLES_OPTIX_TEST=1 ).

I guess the bug is specific to i915 so something on the preview render viewport, when it is updating. At least i have a good workaround now.

I just made a test running blender with prime-run (the arch linux package for running a application with the dedicated gpu) and blender 2.83 worked flawlessly. I could even use optix and it worked and my gpu is theoretically not supported. (CYCLES_OPTIX_TEST=1 ). I guess the bug is specific to i915 so something on the preview render viewport, when it is updating. At least i have a good workaround now.

Johan Hidding commented

2020-06-17 16:50:22 +02:00

Added subscriber: @jhidding

Johan Hidding commented

2020-06-17 16:50:22 +02:00

I have the same behaviour on Linux with an Intel integrated graphics card on Fedora 32.

$ uname -a
Linux ESLT0107 5.6.18-300.fc32.x86_64 #1 SMP Wed Jun 10 21:38:25 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux

$ lspci | grep VGA
00:02.0 VGA compatible controller: Intel Corporation UHD Graphics 620 (rev 07)

I have the same behaviour on Linux with an Intel integrated graphics card on Fedora 32. ``` $ uname -a Linux ESLT0107 5.6.18-300.fc32.x86_64 #1 SMP Wed Jun 10 21:38:25 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux $ lspci | grep VGA 00:02.0 VGA compatible controller: Intel Corporation UHD Graphics 620 (rev 07) ```

a commented

2020-06-18 13:57:00 +02:00

Added subscriber: @dniku

a commented

2020-06-18 13:57:01 +02:00

Similar symptoms on Dell XPS 15 9560 under Manjaro GNOME. I run Blender 2.83 on the integrated Intel GPU, and ≈5 minutes after startup the whole Xorg session freezes. journalctl contains the following:

Jun 17 21:30:13 thamiel kernel: Asynchronous wait on fence 0000:00:02.0:gnome-shell[1436]:63dee timed out (hint:intel_atomic_commit_ready+0x0/0x50 [i915])
Jun 17 21:30:13 thamiel kernel: Asynchronous wait on fence 0000:00:02.0:gnome-shell[1436]:63dee timed out (hint:intel_atomic_commit_ready+0x0/0x50 [i915])
Jun 17 21:30:14 thamiel kernel: i915 0000:00:02.0: Resetting rcs0 for preemption time out
Jun 17 21:30:14 thamiel kernel: i915 0000:00:02.0: blender[54049] context reset due to GPU hang
Jun 17 21:30:14 thamiel kernel: i915 0000:00:02.0: GPU HANG: ecode 9:1:87f99eb9, in blender [54049]
Jun 17 21:30:14 thamiel kernel: GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
Jun 17 21:30:14 thamiel kernel: Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new.
Jun 17 21:30:14 thamiel kernel: Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details.
Jun 17 21:30:14 thamiel kernel: drm/i915 developers can then reassign to the right component if it's not a kernel issue.
Jun 17 21:30:14 thamiel kernel: The GPU crash dump is required to analyze GPU hangs, so please always attach it.
Jun 17 21:30:14 thamiel kernel: GPU crash dump saved to /sys/class/drm/card1/error

(note ecode identical to the one provided by @mattffly)

Similar symptoms on Dell XPS 15 9560 under Manjaro GNOME. I run Blender 2.83 on the integrated Intel GPU, and ≈5 minutes after startup the whole Xorg session freezes. `journalctl` contains the following: ``` Jun 17 21:30:13 thamiel kernel: Asynchronous wait on fence 0000:00:02.0:gnome-shell[1436]:63dee timed out (hint:intel_atomic_commit_ready+0x0/0x50 [i915]) Jun 17 21:30:13 thamiel kernel: Asynchronous wait on fence 0000:00:02.0:gnome-shell[1436]:63dee timed out (hint:intel_atomic_commit_ready+0x0/0x50 [i915]) Jun 17 21:30:14 thamiel kernel: i915 0000:00:02.0: Resetting rcs0 for preemption time out Jun 17 21:30:14 thamiel kernel: i915 0000:00:02.0: blender[54049] context reset due to GPU hang Jun 17 21:30:14 thamiel kernel: i915 0000:00:02.0: GPU HANG: ecode 9:1:87f99eb9, in blender [54049] Jun 17 21:30:14 thamiel kernel: GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. Jun 17 21:30:14 thamiel kernel: Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new. Jun 17 21:30:14 thamiel kernel: Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details. Jun 17 21:30:14 thamiel kernel: drm/i915 developers can then reassign to the right component if it's not a kernel issue. Jun 17 21:30:14 thamiel kernel: The GPU crash dump is required to analyze GPU hangs, so please always attach it. Jun 17 21:30:14 thamiel kernel: GPU crash dump saved to /sys/class/drm/card1/error ``` (note `ecode` identical to the one provided by @mattffly)

Hussam Al-Tayeb commented

2020-06-19 21:23:31 +02:00

Added subscriber: @hussam

Hussam Al-Tayeb commented

2020-06-19 21:23:31 +02:00

I am using the official linux tarball.
2.83.0 hangs on my intel laptop but works well on my nvidia gpu desktop..

2.82a works well without hangs.

I am using the official linux tarball. 2.83.0 hangs on my intel laptop but works well on my nvidia gpu desktop.. 2.82a works well without hangs.

Rafael Domiciano commented

2020-06-24 04:49:18 +02:00

Added subscriber: @rafaeldomi

Hussam Al-Tayeb commented

2020-06-25 18:53:21 +02:00

This comment was removed by @hussam

*This comment was removed by @hussam*

luc commented

2020-07-06 21:00:34 +02:00

Added subscriber: @luc2

Erin M commented

2020-07-11 05:53:37 +02:00

Added subscriber: @Tuxgirl

Erin M commented

2020-07-11 05:53:37 +02:00

Same problem here. All related information that I have gathered is located here: https://gitlab.freedesktop.org/drm/intel/-/issues/2075

I'm not sure where the problem is located, but it has made Blender completely unusable for me.

Same problem here. All related information that I have gathered is located here: https://gitlab.freedesktop.org/drm/intel/-/issues/2075 I'm not sure where the problem is located, but it has made Blender completely unusable for me.

Kotra commented

2020-08-14 09:25:00 +02:00

Added subscriber: @kotrfa

Kotra commented

2020-08-14 09:25:00 +02:00

Same here. Out of nowhere, I get:

8437   │ Aug 14 09:19:07 think460s kernel: i915 0000:00:02.0: Resetting rcs0 for preemption time out
8438   │ Aug 14 09:19:07 think460s kernel: i915 0000:00:02.0: blender[1550210] context reset due to GPU hang
8439   │ Aug 14 09:19:07 think460s kernel: i915 0000:00:02.0: GPU HANG: ecode 9:1:85df9ebd, in blender [1550210]
8440   │ Aug 14 09:19:08 think460s redshift-gtk[1549083]: redshift-gtk: Fatal IO error 11 (Resource temporarily unavailable) on X server :0.

The last line shows how it simply kills the whole xorg, in this case demonstrated by a single app not able to communicate with it. I have to restart whole Xorg session.

Same here. Out of nowhere, I get: ``` 8437 │ Aug 14 09:19:07 think460s kernel: i915 0000:00:02.0: Resetting rcs0 for preemption time out 8438 │ Aug 14 09:19:07 think460s kernel: i915 0000:00:02.0: blender[1550210] context reset due to GPU hang 8439 │ Aug 14 09:19:07 think460s kernel: i915 0000:00:02.0: GPU HANG: ecode 9:1:85df9ebd, in blender [1550210] 8440 │ Aug 14 09:19:08 think460s redshift-gtk[1549083]: redshift-gtk: Fatal IO error 11 (Resource temporarily unavailable) on X server :0. ``` The last line shows how it simply kills the whole xorg, in this case demonstrated by a single app not able to communicate with it. I have to restart whole Xorg session.

Andrew Dionne commented

2020-08-17 10:01:38 +02:00

Added subscriber: @CepheidVariable

Andrew Dionne commented

2020-08-17 10:01:38 +02:00

Right, so as a bit of practice in learning new software development techniques, I took the time to use my PC to bisect the commit in which this particular problem shows up.

It shows up in commit d712f1f83a

The commit prior to it, d0ff3434cf, works.

I've tested even with the most recent commits, and the GPU hang still occurs. I am currently also using drm-tip, and have the i915.enable_psr=0 module parameter set.

I'm using an Intel UHD 620 in a Core i7 10710U.

dmesg output. In my case, it doesn't even properly recover. I have to kill xorg.

[92105.569870] i915 0000:00:02.0: [drm] Resetting rcs0 for preemption time out
[92105.569883] i915 0000:00:02.0: [drm] blender[203377] context reset due to GPU hang
[92105.602975] i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:87f99eb9, in blender [203377]
[92106.343263] i915 0000:00:02.0: [drm] Resetting rcs0 for preemption time out
[92106.343275] i915 0000:00:02.0: [drm] blender[203377] context reset due to GPU hang
[92106.354584] i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:85dfbfff, in blender [203377]
[92107.089851] i915 0000:00:02.0: [drm] Resetting rcs0 for preemption time out
[92107.089863] i915 0000:00:02.0: [drm] blender[203377] context reset due to GPU hang
[92107.099436] i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:85dfbfff, in blender [203377]
[94410.106126] i915 0000:00:02.0: [drm] Resetting rcs0 for preemption time out
[94410.106138] i915 0000:00:02.0: [drm] blender[208399] context reset due to GPU hang
[94410.120128] i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:87f99eb9, in blender [208399]

I doubt it's helpful much, but it's a start, and gives us somewhere to look?

Edit: Took a look at the code of the commit. I see no indications as to why it would cause a GPU hang. I'd argue it's definitely driver-side. I'll take a look later, when I got the energy to do so.

Right, so as a bit of practice in learning new software development techniques, I took the time to use my PC to bisect the commit in which this particular problem shows up. It shows up in commit d712f1f83af881be536ec0d183b7d3025c172684 The commit prior to it, d0ff3434cffa2e056e4f191ead21226f32ea8c15, works. I've tested even with the most recent commits, and the GPU hang still occurs. I am currently also using drm-tip, and have the i915.enable_psr=0 module parameter set. I'm using an Intel UHD 620 in a Core i7 10710U. dmesg output. In my case, it doesn't even properly recover. I have to kill xorg. ``` [92105.569870] i915 0000:00:02.0: [drm] Resetting rcs0 for preemption time out [92105.569883] i915 0000:00:02.0: [drm] blender[203377] context reset due to GPU hang [92105.602975] i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:87f99eb9, in blender [203377] [92106.343263] i915 0000:00:02.0: [drm] Resetting rcs0 for preemption time out [92106.343275] i915 0000:00:02.0: [drm] blender[203377] context reset due to GPU hang [92106.354584] i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:85dfbfff, in blender [203377] [92107.089851] i915 0000:00:02.0: [drm] Resetting rcs0 for preemption time out [92107.089863] i915 0000:00:02.0: [drm] blender[203377] context reset due to GPU hang [92107.099436] i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:85dfbfff, in blender [203377] [94410.106126] i915 0000:00:02.0: [drm] Resetting rcs0 for preemption time out [94410.106138] i915 0000:00:02.0: [drm] blender[208399] context reset due to GPU hang [94410.120128] i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:87f99eb9, in blender [208399] ``` I doubt it's helpful much, but it's a start, and gives us somewhere to look? Edit: Took a look at the code of the commit. I see no indications as to why it would cause a GPU hang. I'd argue it's definitely driver-side. I'll take a look later, when I got the energy to do so.

Timothy commented

2020-08-29 12:13:01 +02:00

Added subscriber: @TimothyDrewFilms

Timothy commented

2020-08-29 12:13:01 +02:00

I've been looking all over for a way to fix this and I think I finally found one!

The bug has been reported on Intel Gitlab page and a suggested workaround was to run blender prefixed as such:

MESA_LOADER_DRIVER_OVERRIDE=i965 blender

I've been using blender 2.83-5 for hours now with no GPU hang!

OS: Manjaro Mikah 20.1
CPU: i7-6700 (Skylake GT2) (no gpu)

(I've also switched from video-linux to video-modesetting, although I haven't tested to see whether or not this makes a difference)

Hope this is helpful!

I've been looking all over for a way to fix this and I think I finally found one! The bug has been reported on [Intel Gitlab page ](https://gitlab.freedesktop.org/drm/intel/-/issues/2380) and a suggested workaround was to run blender prefixed as such: ``` MESA_LOADER_DRIVER_OVERRIDE=i965 blender ``` I've been using blender 2.83-5 for hours now with no GPU hang! OS: Manjaro Mikah 20.1 CPU: i7-6700 (Skylake GT2) (no gpu) (I've also switched from video-linux to video-modesetting, although I haven't tested to see whether or not this makes a difference) Hope this is helpful!

Andrew Dionne commented

2020-08-29 19:01:41 +02:00

It's worth noting, that somewhere between my post on August 17th, and August 21st, when the most recent Arch Linux build of Blender was packaged(At the time of this post,) a pull request was done that fixed the GPU hang on Blender's side, rather than the driver side. Version 2.83.5 still hangs the GPU, but 2.83.5-2 in Arch Linux does not.

So, it looks like our three possible choices are to use 2.82, the latest git version, or switch from the Iris driver, as Timothy suggested.

I'd still like to find the cause, though... So I'll get off my lazy posterior and get to work on poking the Linux kernel.

It's worth noting, that somewhere between my post on August 17th, and August 21st, when the most recent Arch Linux build of Blender was packaged(At the time of this post,) a pull request was done that fixed the GPU hang on Blender's side, rather than the driver side. Version 2.83.5 still hangs the GPU, but 2.83.5-2 in Arch Linux does not. So, it looks like our three possible choices are to use 2.82, the latest git version, or switch from the Iris driver, as Timothy suggested. I'd still like to find the cause, though... So I'll get off my lazy posterior and get to work on poking the Linux kernel.

Erin M commented

2020-08-30 05:16:55 +02:00

In #77466#1004981, @CepheidVariable wrote:
It's worth noting, that somewhere between my post on August 17th, and August 21st, when the most recent Arch Linux build of Blender was packaged(At the time of this post,) a pull request was done that fixed the GPU hang on Blender's side, rather than the driver side. Version 2.83.5 still hangs the GPU, but 2.83.5-2 in Arch Linux does not.

So, it looks like our three possible choices are to use 2.82, the latest git version, or switch from the Iris driver, as Timothy suggested.

I'd still like to find the cause, though... So I'll get off my lazy posterior and get to work on poking the Linux kernel.

Andrew, is 2.83.5-2 working only with the env variable that Timothy mentioned? Or is it working for you regardless of whether you set that variable?

> In #77466#1004981, @CepheidVariable wrote: > It's worth noting, that somewhere between my post on August 17th, and August 21st, when the most recent Arch Linux build of Blender was packaged(At the time of this post,) a pull request was done that fixed the GPU hang on Blender's side, rather than the driver side. Version 2.83.5 still hangs the GPU, but 2.83.5-2 in Arch Linux does not. > > So, it looks like our three possible choices are to use 2.82, the latest git version, or switch from the Iris driver, as Timothy suggested. > > I'd still like to find the cause, though... So I'll get off my lazy posterior and get to work on poking the Linux kernel. Andrew, is 2.83.5-2 working only with the env variable that Timothy mentioned? Or is it working for you regardless of whether you set that variable?

Andrew Dionne commented

2020-08-30 14:30:24 +02:00

I'm going to have to rescind my previous post actually. Today, it's gone back to crashing again. Happens within 30-60 seconds of starting blender. GPU just hangs. Also seeing a report(https://gitlab.freedesktop.org/drm/intel/-/issues/2380#note_611990) on the Intel gitlab that 2.83.5-2 is still crashing for another user, only that it's taking up to an hour to occur instead of minutes for them.

When it WAS working for longer though, it was without the environment variable.

I'm gonna go poke the i915 kernel module today. Got nothing better to do with my time today.

I'm going to have to rescind my previous post actually. Today, it's gone back to crashing again. Happens within 30-60 seconds of starting blender. GPU just hangs. Also seeing a report(https://gitlab.freedesktop.org/drm/intel/-/issues/2380#note_611990) on the Intel gitlab that 2.83.5-2 is still crashing for another user, only that it's taking up to an hour to occur instead of minutes for them. When it *WAS* working for longer though, it was without the environment variable. I'm gonna go poke the i915 kernel module today. Got nothing better to do with my time today.

Richard Antalik commented

2020-08-31 11:19:34 +02:00

Added subscribers: @fho, @iss

Richard Antalik commented

2020-10-27 06:03:12 +01:00

Added subscribers: @Cowhead, @kursadk

Benoit commented

2020-11-07 13:08:02 +01:00

Added subscriber: @Yquux

Benoit commented

2020-11-07 13:08:02 +01:00

Blender user for 3 years now here, on Linux - I have that same issues since 2.83 came. 2.82a is still working flawlessly.

My computer :
Kernel: 5.4.0-52-generic x86_64
Desktop: Cinnamon 4.6.7 Distro: Linux Mint 20 Ulyana
RAM : 32Go
CPU : Intel Core i7-6700T bits: 64

GPU:
Device-1: Intel HD Graphics 530 driver: i915 v: kernel
Device-2: NVIDIA TU106 [GeForce RTX 2070] driver: nvidia v: 450.80.02
Device-3: NVIDIA TU106 [GeForce RTX 2070] driver: nvidia v: 450.80.02
Device-4: NVIDIA GP107 [GeForce GTX 1050 Ti] driver: nvidia v: 450.80.02
Display: x11 server: X.Org 1.20.8 driver: modesetting,nvidia
OpenGL: renderer: Mesa Intel HD Graphics 530 (SKL GT2) v: 4.6 Mesa 20.0.8

I use Blender not installed, downloading tarball and launching it with links on my deskptop.
I used to install headless NVIDIA drivers (before *buntu 20.04) and keep on using the intel embedded graphics as display, and NVIDIA graphic cards for computing (CUDA).
Now since Linux Mint 20 is introducing the Prime applet quite easily, I just stay in 'onDemand' mode, then NVIDIA graphic cards are recognized and used by Blender, while all display works on Intel.
(to be noted that with the mode : 'intel for power saving', Blender just does not see the NVIDIA graphic cards, I never understood why).

When being on "onDemand" use:

Version 2.82a : works perfectly
Version 2.83.8 (an all from 83.0) : lags then freeze
Version 2.90.1 : lags then freeze
Version 2.91 beta : lags then freeze

"Lags and freeze" means when being in object mode, trying to switch to rendered view, it will work a few times, and then, randomly but quickly, especially when you're moving a rendered scene, it will stop at something like "Path Tracing Sample 32/144". For a few times it will just hang 10s and then finish the render - and then it will freeze for good, and Blender will just stop - like if it was killed.

I tried using the now available kernel 5.8.0-25 : when a freeze occurs, it is even more annoying since it freezes everything (except mouse) - I can't get the panel board of Cinnamon to come back - CTRL ALT F2 is the only option. I tried kernel 5.8 on 2.90.1 only though.

BUT :
on POWER mode (i.e. display on NVIDIA) : no issue at all -

Therefore I also tend to consider it is an issue with i915 intel driver.

I'd like to use that "MESA_LOADER_DRIVER_OVERRIDE=i965 blender" option... but I confess I don't understand how to use it !!

Blender user for 3 years now here, on Linux - I have that same issues since 2.83 came. 2.82a is still working flawlessly. My computer : Kernel: 5.4.0-52-generic x86_64 Desktop: Cinnamon 4.6.7 Distro: Linux Mint 20 Ulyana RAM : 32Go CPU : Intel Core i7-6700T bits: 64 GPU: Device-1: Intel HD Graphics 530 driver: i915 v: kernel Device-2: NVIDIA TU106 [GeForce RTX 2070] driver: nvidia v: 450.80.02 Device-3: NVIDIA TU106 [GeForce RTX 2070] driver: nvidia v: 450.80.02 Device-4: NVIDIA GP107 [GeForce GTX 1050 Ti] driver: nvidia v: 450.80.02 Display: x11 server: X.Org 1.20.8 driver: modesetting,nvidia OpenGL: renderer: Mesa Intel HD Graphics 530 (SKL GT2) v: 4.6 Mesa 20.0.8 I use Blender not installed, downloading tarball and launching it with links on my deskptop. I used to install headless NVIDIA drivers (before *buntu 20.04) and keep on using the intel embedded graphics as display, and NVIDIA graphic cards for computing (CUDA). Now since Linux Mint 20 is introducing the Prime applet quite easily, I just stay in 'onDemand' mode, then NVIDIA graphic cards are recognized and used by Blender, while all display works on Intel. (to be noted that with the mode : 'intel for power saving', Blender just does not see the NVIDIA graphic cards, I never understood why). When being on "onDemand" use: Version 2.82a : works perfectly Version 2.83.8 (an all from 83.0) : lags then freeze Version 2.90.1 : lags then freeze Version 2.91 beta : lags then freeze "Lags and freeze" means when being in object mode, trying to switch to rendered view, it will work a few times, and then, randomly but quickly, especially when you're moving a rendered scene, it will stop at something like "Path Tracing Sample 32/144". For a few times it will just hang 10s and then finish the render - and then it will freeze for good, and Blender will just stop - like if it was killed. I tried using the now available kernel 5.8.0-25 : when a freeze occurs, it is even more annoying since it freezes everything (except mouse) - I can't get the panel board of Cinnamon to come back - CTRL ALT F2 is the only option. I tried kernel 5.8 on 2.90.1 only though. BUT : on POWER mode (i.e. display on NVIDIA) : no issue at all - Therefore I also tend to consider it is an issue with i915 intel driver. I'd like to use that "MESA_LOADER_DRIVER_OVERRIDE=i965 blender" option... but I confess I don't understand how to use it !!

Kotra commented

2020-11-07 14:04:38 +01:00

Removed subscriber: @kotrfa

Richard Antalik commented

2020-11-26 20:12:27 +01:00

Closed as duplicate of #80458

Richard Antalik closed this issue

2020-11-26 20:12:27 +01:00

Sign in to join this conversation.

No Label

Download

What's New

Blender Studio

Manual

Developers Blog

Documentation

Benchmark

Blender Conference

Development Fund

One-time Donations

Blender 2.83 crashing and lagging when rendering the viewport (cycles). #77466