You are not logged in.
I'm trying to run foldingathome-noroot, but it fails to identify my NVidia GTX970 GPU. Any idea what I'm missing?
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19:INFO(1):Read GPUs.txt
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19:************************* Folding@home Client *************************
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Website: https://foldingathome.org/
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Copyright: (c) 2009-2018 foldingathome.org
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Author: Joseph Coffland <joseph@cauldrondevelopment.com>
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Args:
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Config: /var/lib/private/fah/config.xml
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19:******************************** Build ********************************
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Version: 7.5.1
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Date: May 11 2018
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Time: 19:59:04
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Repository: Git
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Revision: 4705bf53c635f88b8fe85af7675557e15d491ff0
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Branch: master
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Compiler: GNU 6.3.0 20170516
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Options: -std=gnu++98 -O3 -funroll-loops
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Platform: linux2 4.14.0-3-amd64
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Bits: 64
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Mode: Release
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19:******************************* System ********************************
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: CPU: Intel(R) Core(TM) i5-4590 CPU @ 3.30GHz
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: CPU ID: GenuineIntel Family 6 Model 60 Stepping 3
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: CPUs: 4
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Memory: 7.70GiB
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19:Free Memory: 3.17GiB
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: Threads: POSIX_THREADS
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: OS Version: 5.5
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19:Has Battery: false
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: On Battery: false
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: UTC Offset: -4
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: PID: 7983
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: CWD: /var/lib/private/fah
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: OS: Linux 5.5.13-arch1-1 x86_64
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: OS Arch: AMD64
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: GPUs: 1
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: GPU 0: Bus:1 Slot:0 Func:0 NVIDIA:5 GM204 [GeForce GTX 970]
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: CUDA: Not detected: cuInit() returned 999
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: OpenCL: Not detected: clGetPlatformIDs() returned -1001
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19:***********************************************************************
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19:<config>
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: <!-- User Information -->
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: <team v='241205'/>
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: <user v='rcorre'/>
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19:
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19: <!-- Folding Slots -->
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19:</config>
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19:Trying to access database...
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19:Successfully acquired database lock
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19:Enabled folding slot 00: READY cpu:2
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19:Enabled folding slot 01: READY gpu:0:GM204 [GeForce GTX 970]
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19:ERROR:No compute devices matched GPU #0 NVIDIA:5 GM204 [GeForce GTX 970]. You may need to update your graphics drivers.
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19:WARNING:WU00:No longer matches Slot 0's configuration, migrating to FS01
Mar 26 09:37:19 midova FAHClient[7983]: 13:37:19:ERROR:Exception: Unit not found
That GPU is listed in GPUs.txt:
[rcorre@midova ~]$ sudo grep "GTX 970" /var/lib/private/fah/GPUs.txt
0x10de:0x13c2:2:5:GM204 [GeForce GTX 970]
0x10de:0x13d8:2:5:GM204 [GeForce GTX 970M] 2365
0x10de:0x1618:2:5:GM204 [GeForce GTX 970M] 2365
I've installed the necessary opencl packages:
[rcorre@midova ~]$ pacman -Qs opencl
local/clinfo 2.2.18.04.06-1
A simple OpenCL application that enumerates all available platform and device properties
local/ocl-icd 2.2.12-3
OpenCL ICD Bindings
local/opencl-nvidia 440.64-2
OpenCL implemention for NVIDIA
The fah user is in the video group
[rcorre@midova ~]$ groups fah
video fah
Here's the full clinfo output:
Number of platforms 1
Platform Name NVIDIA CUDA
Platform Vendor NVIDIA Corporation
Platform Version OpenCL 1.2 CUDA 10.2.141
Platform Profile FULL_PROFILE
Platform Extensions cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_fp64 cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll cl_nv_copy_opts cl_nv_create_buffer cl_khr_int64_base_atomics cl_khr_int64_extended_atomics
Platform Extensions function suffix NV
Platform Name NVIDIA CUDA
Number of devices 1
Device Name GeForce GTX 970
Device Vendor NVIDIA Corporation
Device Vendor ID 0x10de
Device Version OpenCL 1.2 CUDA
Driver Version 440.64
Device OpenCL C Version OpenCL C 1.2
Device Type GPU
Device Topology (NV) PCI-E, 01:00.0
Device Profile FULL_PROFILE
Device Available Yes
Compiler Available Yes
Linker Available Yes
Max compute units 13
Max clock frequency 1253MHz
Compute Capability (NV) 5.2
Device Partition (core)
Max number of sub-devices 1
Supported partition types None
Supported affinity domains (n/a)
Max work item dimensions 3
Max work item sizes 1024x1024x64
Max work group size 1024
Preferred work group size multiple 32
Warp size (NV) 32
Preferred / native vector sizes
char 1 / 1
short 1 / 1
int 1 / 1
long 1 / 1
half 0 / 0 (n/a)
float 1 / 1
double 1 / 1 (cl_khr_fp64)
Half-precision Floating-point support (n/a)
Single-precision Floating-point support (core)
Denormals Yes
Infinity and NANs Yes
Round to nearest Yes
Round to zero Yes
Round to infinity Yes
IEEE754-2008 fused multiply-add Yes
Support is emulated in software No
Correctly-rounded divide and sqrt operations Yes
Double-precision Floating-point support (cl_khr_fp64)
Denormals Yes
Infinity and NANs Yes
Round to nearest Yes
Round to zero Yes
Round to infinity Yes
IEEE754-2008 fused multiply-add Yes
Support is emulated in software No
Address bits 64, Little-Endian
Global memory size 4236115968 (3.945GiB)
Error Correction support No
Max memory allocation 1059028992 (1010MiB)
Unified memory for Host and Device No
Integrated memory (NV) No
Minimum alignment for any data type 128 bytes
Alignment of base address 4096 bits (512 bytes)
Global Memory cache type Read/Write
Global Memory cache size 638976 (624KiB)
Global Memory cache line size 128 bytes
Image support Yes
Max number of samplers per kernel 32
Max size for 1D images from buffer 268435456 pixels
Max 1D or 2D image array size 2048 images
Max 2D image size 16384x16384 pixels
Max 3D image size 4096x4096x4096 pixels
Max number of read image args 256
Max number of write image args 16
Local memory type Local
Local memory size 49152 (48KiB)
Registers per block (NV) 65536
Max number of constant args 9
Max constant buffer size 65536 (64KiB)
Max size of kernel argument 4352 (4.25KiB)
Queue properties
Out-of-order execution Yes
Profiling Yes
Prefer user sync for interop No
Profiling timer resolution 1000ns
Execution capabilities
Run OpenCL kernels Yes
Run native kernels No
Kernel execution timeout (NV) Yes
Concurrent copy and kernel execution (NV) Yes
Number of async copy engines 2
printf() buffer size 1048576 (1024KiB)
Built-in kernels (n/a)
Device Extensions cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_fp64 cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll cl_nv_copy_opts cl_nv_create_buffer cl_khr_int64_base_atomics cl_khr_int64_extended_atomics
NULL platform behavior
clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...) NVIDIA CUDA
clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...) Success [NV]
clCreateContext(NULL, ...) [default] Success [NV]
clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT) No platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU) No devices found in platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU) No platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_ACCELERATOR) No devices found in platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_CUSTOM) Invalid device type for platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_ALL) No platform
ICD loader properties
ICD loader Name OpenCL ICD Loader
ICD loader Vendor OCL Icd free software
ICD loader Version 2.2.12
ICD loader Profile OpenCL 2.2
[rcorre@midova ~]$ lsmod | grep nvidia
nvidia_uvm 1101824 0
nvidia_drm 53248 3
nvidia_modeset 1118208 6 nvidia_drm
nvidia 20471808 286 nvidia_uvm,nvidia_modeset
drm_kms_helper 237568 1 nvidia_drm
ipmi_msghandler 73728 2 ipmi_devintf,nvidia
drm 557056 6 drm_kms_helper,nvidia_drm
Last edited by rcorre (2020-04-02 11:46:59)
Offline
Is the foldingathome.service active? I tried installing the other day and couldn't get the daemon to start.
Offline
Yes, the service status is running, but it shows those errors in the log.
Offline
I forgot to mention: I also made the ld.so.conf adjustment mentioned at https://wiki.archlinux.org/index.php/GP … OpenCL.so)
[rcorre@midova ~]$ sudo cat /etc/ld.so.conf.d/00-usrlib.conf
/usr/lib
Offline
And nvidia is listed in my opencl vendors:
[rcorre@midova ~]$ ls /etc/OpenCL/vendors/
nvidia.icd
Offline
Confirmed, I've got the same problem with my 1070 mobile. I've checked all the Outputs of the original post, exactly the same for me, only gpu is different. Everything else is absolutely identical (all the outputs), same packages installed, fah user is in video group.
ERROR:No compute devices matched GPU #0 NVIDIA:5 GP104BM [GeForce GTX 1070 Mobile] 6463. You may need to update your graphics drivers.
Interesting enough, it worked for me some days ago (<= 10). I tried to downgrade nvidia driver, but then mkinitcpio complains that there's no nvidia and no nvidia-drm module. By the way, back when it worked, fah user was not even in video group. I just tried this as one of the fixes lately, when it already had stopped working.
I'm not sure how to proceed from here, but it might be that one of the latest driver updates introduced this, nvidia-440.64-5 or nvidia-440.64-8.
Last edited by OOFscr (2020-03-31 08:36:35)
Offline
Yes, I was pretty sure I had it working a few days before I posted this too.
Offline
Ooh, it is working after an update today.
[2020-04-01T11:33:44-0400] [ALPM] upgraded linux (5.5.13.arch1-1 -> 5.5.13.arch2-1)
[2020-04-01T11:33:47-0400] [ALPM] upgraded nvidia (440.64-8 -> 440.64-9)
Last edited by rcorre (2020-04-01 23:48:31)
Offline
Sadly, not for me, still the same error although I did the same update. Was already happy after reading your post :(
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18:Free Memory: 27.84GiB
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: Threads: POSIX_THREADS
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: OS Version: 5.5
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18:Has Battery: true
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: On Battery: false
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: UTC Offset: 2
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: PID: 2088
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: CWD: /var/lib/private/fah
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: OS: Linux 5.5.13-arch2-1 x86_64
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: OS Arch: AMD64
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: GPUs: 1
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: GPU 0: Bus:1 Slot:0 Func:0 NVIDIA:5 GP104BM [GeForce GTX 1070 Mobile] 6463
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: CUDA: Not detected: cuInit() returned 999
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: OpenCL: Not detected: clGetPlatformIDs() returned -1001
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18:***********************************************************************
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18:<config>
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: <!-- Slot Control -->
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: <power v='FULL'/>
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18:
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: <!-- User Information -->
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: <team v='45032'/>
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18:
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: <!-- Folding Slots -->
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: <slot id='0' type='CPU'/>
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18: <slot id='1' type='GPU'/>
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18:</config>
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18:Trying to access database...
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18:Successfully acquired database lock
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18:Enabled folding slot 00: READY cpu:7
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18:Enabled folding slot 01: READY gpu:0:GP104BM [GeForce GTX 1070 Mobile] 6463
Apr 02 03:11:18 ArchBoxJ FAHClient[2088]: 01:11:18:ERROR:No compute devices matched GPU #0 NVIDIA:5 GP104BM [GeForce GTX 1070 Mobile] 6463. You may need to update your graphics drivers.
Offline
And it already happened, and I already rebootet just to be sure
[2020-04-02T02:57:25+0200] [ALPM] upgraded linux (5.5.13.arch1-1 -> 5.5.13.arch2-1)
[2020-04-02T02:57:29+0200] [ALPM] upgraded linux-headers (5.5.13.arch1-1 -> 5.5.13.arch2-1)
[2020-04-02T02:57:29+0200] [ALPM] upgraded nvidia (440.64-8 -> 440.64-9)
Offline
Argh, I fooled myself too. After booting this morning I'm back to the same issue. Something got it working, but it wasn't the update
Offline
It seems the magic might be running `sudo clinfo` (sudo is needed!) _before_ starting the service. https://bbs.archlinux.org/viewtopic.php?id=177991 suggests that you need to run an opencl/cuda program once as root, then it will work rootless after that. Weird, but seems to work for me.
Offline
Seriously, WTF? This seems to work for me, too. I am pretty sure I clinfo'ed from root account yesterday, so yes, it seems sudo is indeed neccessary. Big thanks! But again: WTF?
Edit: Because, on my system FAHClient runs as user fah, who is in no way connected to sudo and doesn't use sudo. I don't get it. Does sudo + clinfo somehow expose otherwise restricted OpenCL / nvidia components to the userspace? Any Captain here?!
Last edited by OOFscr (2020-04-02 17:53:08)
Offline
same for me. why does sudo clinfo make it work?
Offline
I can confirm.
Stopped the foldingathome service, then did sudo clinfo and afterwards, started the service again.
Now, CUDA and OpenCL appear to be detected correctly.
UPDATE
Two questions:
1. Why does that happen this way? -> this is explained here.
2. Will this survive a reboot? -> No.
And here's a solution:
First, run
/usr/bin/nvidia-modprobe -c0 -u
Then start the foldingathome service:
systemctl start foldingathome.service
I tried to include nvidia-modprobe in the foldingathome.service file to read like this (note the ExecStartPre line). But unfortunately, this fails (most likely due to the service started too early - the modprobe command only works when the graphical desktop is fully loaded).
[Service]
User=fah
Group=fah
WorkingDirectory=/var/lib/fah
ExecStartPre=/usr/bin/nvidia-modprobe -c0 -u
ExecStart=
ExecStart=/opt/fah/FAHClient --config /var/lib/fah/config.xml --exec-directory=/opt/fah --data-directory=/var/lib/fah
One strange thing is that you must keep the empty ExecStart= line as otherwise, systemd complains about a "bad file status", having multiple ExecStart= lines.
So you will HAVE to have multiple ExecStart= lines in order to prevent systemd from complaining about multiple ExecStart= lines.
I assume that only Lennart Poettering can understand this logic
Last edited by DAC324 (2020-04-15 11:18:34)
Offline
Seems to be of no interest anymore. Looks like the Corona virus is already beaten. That's at least good news
Offline
Thanks for the link to the explanation. i can live with the workaround of calling clinfo first.
Offline