GPU disabled, no opencl library found (Linux)

[threeqc]

FAH seems to be detecting CUDA properly, but not OpenCL. The system info says it couldn't find libOpenCL.so. Any ideas on how to fix it?

nvidia-smi:

Code: Select all

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 525.85.05    Driver Version: 525.85.05    CUDA Version: 12.0     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  NVIDIA GeForce ...  Off  | 00000000:01:00.0  On |                  N/A |
|  0%   36C    P5     8W / 130W |    342MiB /  8192MiB |      5%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+

(last bit of) clinfo:

Code: Select all

ICD loader properties
  ICD loader Name                                 OpenCL ICD Loader
  ICD loader Vendor                               OCL Icd free software
  ICD loader Version                              2.2.14
  ICD loader Profile                              OpenCL 3.0

[bikeaddict]

Install ocl-icd-devel (Fedora) or ocl-icd-opencl-dev (Ubuntu) or similar package.

[threeqc]

Install ocl-icd-devel (Fedora) or ocl-icd-opencl-dev (Ubuntu) or similar package.

This worked, thanks.

[threeqc]

Now it's doing this:

Code: Select all

01:19:19:WU01:FS01:0x22:Error initializing CUDA: CUDA_ERROR_UNKNOWN (999) at /home/conda/feedstock_root/build_artifacts/openmm_1640541150081/work/platforms/cuda/src/CudaContext.cpp:139

I had to pause it briefly and then enter suspend mode (linux mint) and that broke the unit it spent several hours running. Then it got stuck downloading units and failing to run them. Now it just doesn't work at all.

clinfo says this now

Code: Select all

Number of platforms                               0

So I guess I managed to nuke my system somehow. Great.

edit: After a reboot, clinfo works. F@H does not.
edit 2: It finds a "CUDA Device" and an "OpenCL Device", so I have no idea what's going wrong.

[threeqc]

Bump

I'm still having this issue.

[JimboPalmer]

Here is instructions on how to post a log so we know what the software thinks is happening.

viewtopic.php?p=327413&f=24#p327413

Usually the first 100 lines is enough to see the configuration.

[threeqc]

Here is instructions on how to post a log so we know what the software thinks is happening.

viewtopic.php?p=327413&f=24#p327413

Usually the first 100 lines is enough to see the configuration.

Code: Select all

*********************** Log Started 2023-03-30T19:38:22Z ***********************
19:38:22:******************************* libFAH ********************************
19:38:22:           Date: Oct 20 2020
19:38:22:           Time: 20:36:39
19:38:22:       Revision: 5ca109d295a6245e2a2f590b3d0085ad5e567aeb
19:38:22:         Branch: master
19:38:22:       Compiler: GNU 8.3.0
19:38:22:        Options: -faligned-new -std=c++11 -fsigned-char -ffunction-sections
19:38:22:                 -fdata-sections -O3 -funroll-loops -fno-pie
19:38:22:       Platform: linux2 5.8.0-1-amd64
19:38:22:           Bits: 64
19:38:22:           Mode: Release
19:38:22:****************************** FAHClient ******************************
19:38:22:        Version: 7.6.21
19:38:22:         Author: Joseph Coffland <[email protected]>
19:38:22:      Copyright: 2020 foldingathome.org
19:38:22:       Homepage: https://foldingathome.org/
19:38:22:           Date: Oct 20 2020
19:38:22:           Time: 20:39:00
19:38:22:       Revision: 6efbf0e138e22d3963e6a291f78dcb9c6422a278
19:38:22:         Branch: master
19:38:22:       Compiler: GNU 8.3.0
19:38:22:        Options: -faligned-new -std=c++11 -fsigned-char -ffunction-sections
19:38:22:                 -fdata-sections -O3 -funroll-loops -fno-pie
19:38:22:       Platform: linux2 5.8.0-1-amd64
19:38:22:           Bits: 64
19:38:22:           Mode: Release
19:38:22:           Args: --child /etc/fahclient/config.xml --run-as fahclient
19:38:22:                 --pid-file=/var/run/fahclient.pid --daemon
19:38:22:         Config: /etc/fahclient/config.xml
19:38:22:******************************** CBang ********************************
19:38:22:           Date: Oct 20 2020
19:38:22:           Time: 18:37:59
19:38:22:       Revision: 7e4ce85225d7eaeb775e87c31740181ca603de60
19:38:22:         Branch: master
19:38:22:       Compiler: GNU 8.3.0
19:38:22:        Options: -faligned-new -std=c++11 -fsigned-char -ffunction-sections
19:38:22:                 -fdata-sections -O3 -funroll-loops -fno-pie -fPIC
19:38:22:       Platform: linux2 5.8.0-1-amd64
19:38:22:           Bits: 64
19:38:22:           Mode: Release
19:38:22:******************************* System ********************************
19:38:22:            CPU: 13th Gen Intel(R) Core(TM) i5-13600K
19:38:22:         CPU ID: GenuineIntel Family 6 Model 183 Stepping 1
19:38:22:           CPUs: 20
19:38:22:         Memory: 15.42GiB
19:38:22:    Free Memory: 14.38GiB
19:38:22:        Threads: POSIX_THREADS
19:38:22:     OS Version: 5.15
19:38:22:    Has Battery: false
19:38:22:     On Battery: false
19:38:22:     UTC Offset: -4
19:38:22:            PID: 1515
19:38:22:            CWD: /var/lib/fahclient
19:38:22:             OS: Linux 5.15.0-69-generic x86_64
19:38:22:        OS Arch: AMD64
19:38:22:           GPUs: 1
19:38:22:          GPU 0: Bus:1 Slot:0 Func:0 NVIDIA:8 GA106 [Geforce RTX 3050]
19:38:22:  CUDA Device 0: Platform:0 Device:0 Bus:1 Slot:0 Compute:8.6 Driver:12.0
19:38:22:OpenCL Device 0: Platform:0 Device:0 Bus:1 Slot:0 Compute:3.0 Driver:525.89
19:38:22:***********************************************************************
19:38:22:<config>
19:38:22:  <!-- Client Control -->
19:38:22:  <fold-anon v='true'/>
19:38:22:
19:38:22:  <!-- Folding Slot Configuration -->
19:38:22:  <gpu v='false'/>
19:38:22:
19:38:22:  <!-- Network -->
19:38:22:  <proxy v=':8080'/>
19:38:22:
19:38:22:  <!-- Slot Control -->
19:38:22:  <pause-on-start v='true'/>
19:38:22:  <power v='MEDIUM'/>
19:38:22:
19:38:22:  <!-- User Information -->
19:38:22:  <passkey v='*****'/>
19:38:22:  <team v='223518'/>
19:38:22:  <user v='threeqc'/>
19:38:22:
19:38:22:  <!-- Folding Slots -->
19:38:22:  <slot id='0' type='CPU'/>
19:38:22:</config>
19:38:22:Trying to access database...
19:38:22:Successfully acquired database lock
19:38:22:FS00:Initialized folding slot 00: cpu:19