Page 1 of 1

Missing GPU broke software 7.6.13

Posted: Mon Aug 31, 2020 12:50 am
by sam6861
I used to have 2 GPUs. I shut down computer, remove a GPU from inside the computer, and turn on with just 1 GPU. Now, folding at home went mostly broken. FAHControl have endless connecting and updating loop and is mostly blank. Can't delete missing GPU in FAHControl. The web control, after waiting for more then 5 minutes, eventually shows as disconnected.

Failing to connect from missing GPU also happens in windows device manager, disable any GPU in windows device manager and restart folding at home software.

Slow fix: End task Folding at Home software, go to %appdata%\FAHClient, edit config.xml, delete a GPU, save changes.

I would like to see some fix to FAHControl fail to connect over missing GPU, to speed up fixing the settings using FAHControl.

Code: Select all

*********************** Log Started 2020-08-30T22:46:47Z ***********************
22:46:47:Trying to access database...
22:46:47:Successfully acquired database lock
22:46:47:Downloading GPUs.txt from assign1.foldingathome.org:80
22:46:47:Connecting to 192.168.3.1:8080
22:46:47:Read GPUs.txt
22:46:47:Enabled folding slot 00: PAUSED cpu:8 (by user)
22:46:47:Enabled folding slot 01: PAUSED gpu:0:Baffin [Polaris11] (by user)
22:46:47:ERROR:Exception: No available GPUs
22:46:47:ERROR:Exception: Option 'gpu-index' has no default and is not set.
22:46:47:****************************** FAHClient ******************************
22:46:47:        Version: 7.6.13
22:46:47:         Author: Joseph Coffland <joseph@cauldrondevelopment.com>
22:46:47:      Copyright: 2020 foldingathome.org
22:46:47:       Homepage: https://foldingathome.org/
22:46:47:           Date: Apr 27 2020
22:46:47:           Time: 21:21:01
22:46:47:       Revision: 5a652817f46116b6e135503af97f18e094414e3b
22:46:47:         Branch: master
22:46:47:       Compiler: Visual C++ 2008
22:46:47:        Options: /TP /nologo /EHa /wd4297 /wd4103 /Ox /MT
22:46:47:       Platform: win32 10
22:46:47:           Bits: 32
22:46:47:           Mode: Release
22:46:47:           Args: --open-web-control
22:46:47:         Config: C:\Users\sam86\AppData\Roaming\FAHClient\config.xml
22:46:47:******************************** CBang ********************************
22:46:47:           Date: Apr 24 2020
22:46:47:           Time: 17:07:55
22:46:47:       Revision: ea081a3b3b0f4a37c4d0440b4f1bc184197c7797
22:46:47:         Branch: master
22:46:47:       Compiler: Visual C++ 2008
22:46:47:        Options: /TP /nologo /EHa /wd4297 /wd4103 /Ox /MT
22:46:47:       Platform: win32 10
22:46:47:           Bits: 32
22:46:47:           Mode: Release
22:46:47:******************************* System ********************************
22:46:47:            CPU: AMD Ryzen 7 2700X Eight-Core Processor
22:46:47:         CPU ID: AuthenticAMD Family 23 Model 8 Stepping 2
22:46:47:           CPUs: 16
22:46:47:         Memory: 31.95GiB
22:46:47:    Free Memory: 17.86GiB
22:46:47:        Threads: WINDOWS_THREADS
22:46:47:     OS Version: 6.2
22:46:47:    Has Battery: false
22:46:47:     On Battery: false
22:46:47:     UTC Offset: -5
22:46:47:            PID: 14148
22:46:47:            CWD: C:\Users\sam86\AppData\Roaming\FAHClient
22:46:47:  Win32 Service: false
22:46:47:             OS: Windows 10 Home
22:46:47:        OS Arch: AMD64
22:46:47:           GPUs: 1
22:46:47:          GPU 0: Bus:38 Slot:0 Func:0 AMD:5 Baffin [Polaris11]
22:46:47:           CUDA: Not detected: Failed to open dynamic library 'nvcuda.dll': The
22:46:47:                 specified module could not be found.
22:46:47:
22:46:47:OpenCL Device 0: Platform:0 Device:0 Bus:38 Slot:0 Compute:1.2 Driver:3004.8
22:46:47:******************************* libFAH ********************************
22:46:47:           Date: Apr 15 2020
22:46:47:           Time: 14:53:14
22:46:47:       Revision: 216968bc7025029c841ed6e36e81a03a316890d3
22:46:47:         Branch: master
22:46:47:       Compiler: Visual C++ 2008
22:46:47:        Options: /TP /nologo /EHa /wd4297 /wd4103 /Ox /MT
22:46:47:       Platform: win32 10
22:46:47:           Bits: 32
22:46:47:           Mode: Release
22:46:47:***********************************************************************
22:46:47:<config>
22:46:47:  <!-- HTTP Server -->
22:46:47:  <allow v='127.0.0.1 192.168.3.0/24'/>
22:46:47:
22:46:47:  <!-- Network -->
22:46:47:  <proxy v='192.168.3.1:8080'/>
22:46:47:  <proxy-enable v='true'/>
22:46:47:
22:46:47:  <!-- Remote Command Server -->
22:46:47:  <password v='*****'/>
22:46:47:
22:46:47:  <!-- Slot Control -->
22:46:47:  <pause-on-start v='true'/>
22:46:47:  <power v='full'/>
22:46:47:
22:46:47:  <!-- User Information -->
22:46:47:  <passkey v='*****'/>
22:46:47:  <user v='sam6861'/>
22:46:47:
22:46:47:  <!-- Work Unit Control -->
22:46:47:  <next-unit-percentage v='98'/>
22:46:47:
22:46:47:  <!-- Folding Slots -->
22:46:47:  <slot id='0' type='CPU'>
22:46:47:    <client-type v='advanced'/>
22:46:47:    <cpus v='8'/>
22:46:47:  </slot>
22:46:47:  <slot id='1' type='GPU'>
22:46:47:    <client-type v='advanced'/>
22:46:47:  </slot>
22:46:47:  <slot id='2' type='GPU'>
22:46:47:    <client-type v='advanced'/>
22:46:47:  </slot>
22:46:47:</config>
22:46:48:WU00:FS02:Connecting to 192.168.3.1:8080
22:46:48:WU00:FS02:Assigned to work server 3.21.157.11
22:46:48:WU00:FS02:Requesting new work unit for slot 
22:46:48:ERROR:WU00:FS02:Exception: Option 'gpu-index' has no default and is not set.
22:46:49:WU00:FS02:Connecting to 192.168.3.1:8080
22:46:49:WU00:FS02:Assigned to work server 3.21.157.11
22:46:49:WU00:FS02:Requesting new work unit for slot 
22:46:49:ERROR:WU00:FS02:Exception: Option 'gpu-index' has no default and is not set.
22:46:51:ERROR:Exception: Connection terminated
22:46:51:ERROR:Socket not open
22:46:54:ERROR:Exception: Connection terminated
22:46:54:ERROR:Socket not open
22:46:59:ERROR:Exception: Connection terminated
22:46:59:ERROR:Socket not open
22:47:05:ERROR:Exception: Connection terminated
22:47:05:ERROR:Socket not open
22:47:09:ERROR:Exception: Connection terminated
22:47:09:ERROR:Socket not open
22:47:14:ERROR:Exception: Connection terminated
22:47:14:ERROR:Socket not open
22:47:19:ERROR:Exception: Connection terminated
22:47:19:ERROR:Socket not open
22:47:24:ERROR:Exception: Connection terminated
22:47:24:ERROR:Socket not open
22:47:35:27:127.0.0.1:New Web session
22:47:36:ERROR:Exception: Connection terminated
22:47:36:ERROR:Socket not open
22:47:37:WARNING:36:127.0.0.1:500 HTTP INTERNAL SERVER ERROR /api/updates?_=1598827654332&sid=ef460ad47b9491aa6c5a3ad168198d12: Option 'gpu-index' has no default and is not set.
22:47:37:ERROR:HTTP_INTERNAL_SERVER_ERROR: Option 'gpu-index' has no default and is not set.
22:47:40:WARNING:36:127.0.0.1:500 HTTP INTERNAL SERVER ERROR /api/updates?_=1598827654333&sid=ef460ad47b9491aa6c5a3ad168198d12: Option 'gpu-index' has no default and is not set.
22:47:40:ERROR:HTTP_INTERNAL_SERVER_ERROR: Option 'gpu-index' has no default and is not set.
22:47:41:ERROR:Exception: Connection terminated
22:47:41:ERROR:Socket not open
22:47:42:WARNING:37:127.0.0.1:500 HTTP INTERNAL SERVER ERROR /api/updates?_=1598827654334&sid=ef460ad47b9491aa6c5a3ad168198d12: Option 'gpu-index' has no default and is not set.
22:47:42:ERROR:HTTP_INTERNAL_SERVER_ERROR: Option 'gpu-index' has no default and is not set.
22:47:45:WARNING:41:127.0.0.1:500 HTTP INTERNAL SERVER ERROR /api/updates?_=1598827654340&sid=ef460ad47b9491aa6c5a3ad168198d12: Option 'gpu-index' has no default and is not set.
22:47:45:ERROR:HTTP_INTERNAL_SERVER_ERROR: Option 'gpu-index' has no default and is not set.

Re: Missing GPU broke software 7.6.13

Posted: Mon Aug 31, 2020 8:32 am
by PantherX
May I suggest that you open up an issue on GitHub: https://github.com/FoldingAtHome/fah-issues/issues as a feature enhancement?

Generally speaking, any physical change of the GPU*, might trigger the client and fixing it can be a bit of a challenge. A fresh installation can be the quickest path assuming that it was planned which means that you set your slots to finish.

*For those who would like to know:
Switching GPU vendors -> Fresh installation is the safest option.
Upgrading GPU -> Ensure that you finish your GPU Slot and then upgrade.
Removing GPU -> Ensure that you finish your GPU Slot and then delete it.
Adding GPU -> Fresh installation is the safest option.

Of course, that's the behavior with V7.6.13 which isn't an indication of future client versions. Getting GPU detection to work correctly across different platforms and drivers isn't an easy task :(