I see in your config you set the scheduling profile to ‘multiple GPUs’. Try ‘very fast GPU’ or ‘default’.
For what it’s worth, with a single NVidia 1060, these are my settings (note that I use a very large value for mandatory_timeout
):
opencl=TRUE
opencl_device_priority=+0/*/+0/*/*
opencl_disable_drivers_blacklist=false
opencl_library=
opencl_mandatory_timeout=20000
opencl_scheduling_profile=very fast GPU
opencl_synch_cache=active module
opencl_tune_headroom=FALSE
opencl_tuning_mode=nothing
opencl_use_cpu_devices=false
My resource
settings:
resource_default=512 32 128 700
resource_large=700 64 128 900
resource_small=128 16 64 400
resource_unrestricted=16384 1024 128 1024
resourcelevel=large
The clplatform
settings (your card name will be different):
cldevice_v5_nvidiacudanvidiageforcegtx10606gb=0 250 0 64 64 1024 1 0 0.000 0.000 0.250
cldevice_v5_nvidiacudanvidiageforcegtx10606gb_building=-cl-fast-relaxed-math
cldevice_v5_nvidiacudanvidiageforcegtx10606gb_id0=600
clplatform_amdacceleratedparallelprocessing=FALSE
clplatform_apple=FALSE
clplatform_intelropenclhdgraphics=FALSE
clplatform_nvidiacuda=TRUE
clplatform_openclon12=FALSE
clplatform_other=FALSE
clplatform_rusticl=FALSE
You can force compiled ‘kernels’ (processing code) to be regenerated by deleting them:
rm -rf ~/.cache/darktable/*kernel*
Could you try running with -d opencl -d verbose -d perf
instead of -d common
? (edit: I now see you have tried that already, still, I think it’s worth having it on at least while you’re troubleshooting; you can combine it with -d common
and with -d verbose
, too). Certain warnings are only shown if opencl
logging is on (for example reaching the mandatory_timeout and falling back to CPU, or ‘memory allocation failed’). Maybe it’s a bit less verbose, and you may get more OpenCL logs. I get output like this (notice lines like [preview] processed 'colorout' on CPU, blended on CPU
and processed 'channelmixerrgb' on GPU, blended on GPU
; the words in square brackets, like [full]
and [preview]
are the names of the pipelines):
[opencl_init] opencl_scheduling_profile: 'very fast GPU'
[opencl_init] opencl_device_priority: '+0/*/+0/*/*'
[opencl_init] opencl_mandatory_timeout: 20000
[dt_opencl_update_priorities] these are your device priorities:
[dt_opencl_update_priorities] image preview export thumbs preview2
[dt_opencl_update_priorities] 0 0 0 0 0
[dt_opencl_update_priorities] show if opencl use is mandatory for a given pixelpipe:
[dt_opencl_update_priorities] image preview export thumbs preview2
[dt_opencl_update_priorities] 1 1 1 1 1
[opencl_synchronization_timeout] synchronization timeout set to 0
[dt_opencl_update_priorities] these are your device priorities:
[dt_opencl_update_priorities] image preview export thumbs preview2
[dt_opencl_update_priorities] 0 0 0 0 0
[dt_opencl_update_priorities] show if opencl use is mandatory for a given pixelpipe:
[dt_opencl_update_priorities] image preview export thumbs preview2
[dt_opencl_update_priorities] 1 1 1 1 1
[opencl_synchronization_timeout] synchronization timeout set to 0
1.1096 [opencl_update_settings] scheduling profile set to very fast GPU
23.2498 [dt_dev_load_raw] loading the image. took 0.171 secs (0.139 CPU)
23.7243 [histogram] took 0.001 secs (0.001 CPU) scope draw
23.7671 [dt_dev_process_image_job] loading image. took 0.000 secs (0.000 CPU)
23.7846 [dt_opencl_check_tuning] use 4808MB (headroom=OFF, pinning=OFF) on device `NVIDIA CUDA NVIDIA GeForce GTX 1060 6GB' id=0
23.7849 [dev_pixelpipe] took 0.000 secs (0.000 CPU) initing base buffer [full]
23.7922 [dev_pixelpipe] took 0.007 secs (0.007 CPU) [full] processed `rawprepare' on GPU, blended on GPU
23.7936 [dev_pixelpipe] took 0.001 secs (0.002 CPU) [full] processed `temperature' on GPU, blended on GPU
23.8003 [dev_pixelpipe] took 0.007 secs (0.005 CPU) [full] processed `highlights' on GPU, blended on GPU
23.8294 [resample_cl] plan 0.000 secs (0.000 CPU) resample 0.001 secs (0.001 CPU)
23.8302 [histogram] took 0.000 secs (0.000 CPU) scope draw
23.8461 [dev_pixelpipe] took 0.046 secs (0.044 CPU) [full] processed `demosaic' on GPU, blended on GPU
23.8989 [dev_pixelpipe] took 0.053 secs (0.101 CPU) [full] processed `lens' on GPU, blended on GPU
23.8992 [histogram] took 0.004 secs (0.001 CPU) scope draw
23.9019 [dev_pixelpipe] took 0.003 secs (0.005 CPU) [full] processed `exposure' on GPU, blended on GPU
23.9059 [dev_pixelpipe] took 0.004 secs (0.003 CPU) [full] processed `colorin' on GPU, blended on GPU
23.9092 [dt_ioppr_transform_image_colorspace_cl] IOP_CS_LAB-->IOP_CS_RGB took 0.003 secs (0.000 GPU) [channelmixerrgb]
23.9139 [dev_pixelpipe] took 0.008 secs (0.002 CPU) [full] processed `channelmixerrgb' on GPU, blended on GPU
23.9197 [dev_pixelpipe] took 0.006 secs (0.007 CPU) [full] processed `filmicrgb' on GPU, blended on GPU
23.9602 [dt_ioppr_transform_image_colorspace] IOP_CS_RGB-->IOP_CS_LAB took 0.005 secs (0.053 CPU) [colorout]
24.0420 [dev_pixelpipe] took 0.122 secs (0.852 CPU) [full] processed `colorout' on CPU, blended on CPU
24.0495 [dev_pixelpipe] took 0.007 secs (0.072 CPU) [full] processed `gamma' on CPU, blended on CPU
24.0495 [opencl_profiling] profiling device 0 ('NVIDIA CUDA NVIDIA GeForce GTX 1060 6GB'):
24.0495 [opencl_profiling] spent 0.0026 seconds in [Write Image (from host to device)]
....
24.0496 [opencl_profiling] spent 0.0015 seconds in filmicrgb_chroma
24.0496 [opencl_profiling] spent 0.0350 seconds in [Read Image (from device to host)]
24.0496 [opencl_profiling] spent 0.1068 seconds totally in command queue (with 0 events missing)
24.0562 [dev_process_image] pixel pipeline took 0.272 secs (1.102 CPU) processing `2024-02-04_17-22-10_P1030065.RW2'
24.0587 [histogram] took 0.000 secs (0.000 CPU) scope draw
24.0670 [dt_dev_process_image_job] loading image. took 0.000 secs (0.000 CPU)
24.0670 [dt_dev_process_image_job] loading image. took 0.000 secs (0.001 CPU)
24.0782 [dev_process_image] pixel pipeline took 0.011 secs (0.045 CPU) processing `2024-02-04_17-22-10_P1030065.RW2'
24.0900 [histogram] took 0.000 secs (0.000 CPU) scope draw
24.0932 [dt_dev_process_image_job] loading image. took 0.000 secs (0.000 CPU)
24.1138 [dev_pixelpipe] took 0.000 secs (0.000 CPU) initing base buffer [preview]
24.1153 [dev_pixelpipe] took 0.001 secs (0.001 CPU) [preview] processed `rawprepare' on GPU, blended on GPU
24.1160 [dev_pixelpipe] took 0.001 secs (0.000 CPU) [preview] processed `temperature' on GPU, blended on GPU
24.1168 [dev_pixelpipe] took 0.001 secs (0.000 CPU) [preview] processed `highlights' on GPU, blended on GPU
24.1212 [dev_pixelpipe] took 0.004 secs (0.000 CPU) [preview] processed `demosaic' on GPU, blended on GPU
24.1330 [dev_pixelpipe] took 0.012 secs (0.000 CPU) [preview] processed `lens' on GPU, blended on GPU
24.1338 [dev_pixelpipe] took 0.001 secs (0.000 CPU) [preview] processed `exposure' on GPU, blended on GPU
24.1355 [dev_pixelpipe] took 0.002 secs (0.000 CPU) [preview] processed `colorin' on GPU, blended on GPU
24.1373 [dt_ioppr_transform_image_colorspace_cl] IOP_CS_LAB-->IOP_CS_RGB took 0.001 secs (0.000 GPU) [channelmixerrgb]
24.1390 [dev_pixelpipe] took 0.003 secs (0.000 CPU) [preview] processed `channelmixerrgb' on GPU, blended on GPU
24.1432 [dev_pixelpipe] took 0.004 secs (0.007 CPU) [preview] processed `filmicrgb' on GPU, blended on GPU
24.1559 [dt_ioppr_transform_image_colorspace] IOP_CS_RGB-->IOP_CS_LAB took 0.001 secs (0.010 CPU) [colorout]
24.1781 [dev_pixelpipe] took 0.035 secs (0.237 CPU) [preview] processed `colorout' on CPU, blended on CPU
24.1808 [dev_pixelpipe] took 0.003 secs (0.019 CPU) [preview] processed `gamma' on CPU, blended on CPU
24.2185 [dt_ioppr_transform_image_colorspace_rgb] RGB-->RGB took 0.038 secs (0.386 lcms2) [final histogram]
24.2722 [histogram] took 0.091 secs (0.452 CPU) final RGB parade
24.2722 [opencl_profiling] profiling device 0 ('NVIDIA CUDA NVIDIA GeForce GTX 1060 6GB'):
24.2722 [opencl_profiling] spent 0.0003 seconds in [Write Image (from host to device)]
24.2722 [opencl_profiling] spent 0.0001 seconds in rawprepare_1f
....
24.2723 [opencl_profiling] spent 0.0206 seconds totally in command queue (with 0 events missing)
24.2741 [dev_process_image] pixel pipeline took 0.161 secs (0.721 CPU) processing `2024-02-04_17-22-10_P1030065.RW2'