Page 1 of 1
F@H using only 1 of 2 (identical) GPUs - is this expected?
Posted: Thu Apr 30, 2020 10:55 pm
by TyMarbut
I'm relatively new to this so please forgive me if this is a common issue - a search didn't turn up anything directly relevant.
My F@H client, which has been running for several weeks, is now finally getting enough work to run the CPU and one of two identical GPUs (MSI 1070ti) pretty much flat out. The client interface lists both GPUs, but I'm not sure that I've ever seen it use the second GPU, even now that I'm running full bore on the CPU and GPU1.
So, the question: Is this expected, if the amount of work available is only enough for me to use the one GPU? That is, I assume it doesn't do any sort of load balancing between the GPUs. Would you expect different behavior than what I'm seeing? I'd love to contribute both GPUs and the power to run them if I could just get them both to perform work.
System specs:
OS: Windows 10
CPU: Ryzen 1950x
GPU: 2x MSi 1070ti "Duke"
RAM: 64GB @2666mhz
Re: F@H using only 1 of 2 (identical) GPUs - is this expecte
Posted: Thu Apr 30, 2020 11:26 pm
by bruce
You can post your log per below, but I'm going to guess that you have two slots configured, one for the CPU and one for the GPU. If so, you can add a second GPU slot.
If you already have 3 slots, it's a fact that they work independently, so when the servers' bandwidth is limited you may not get assignments for all three concurrently.
Re: F@H using only 1 of 2 (identical) GPUs - is this expecte
Posted: Fri May 01, 2020 12:16 am
by TyMarbut
Thanks Bruce! I will both check the logs (post if I can locate them) and learn about configuring >2 slots.
Meanwhile, though, if only 2 slots were configured, would you expect to still see all 3 devices listed on the client web interface? I ask because all 3 are (correctly) identified on the interface so maybe things are running just fine and there isn't enough available bandwidth/work.
Re: F@H using only 1 of 2 (identical) GPUs - is this expecte
Posted: Fri May 01, 2020 12:20 am
by TyMarbut
bruce wrote:You can post your log per below
Here's a relevant selection of the most recent log:
Code: Select all
03:26:00:******************************** Build ********************************
03:26:00: Version: 7.5.1
03:26:00: Date: May 11 2018
03:26:00: Time: 13:06:32
03:26:00: Repository: Git
03:26:00: Revision: 4705bf53c635f88b8fe85af7675557e15d491ff0
03:26:00: Branch: master
03:26:00: Compiler: Visual C++ 2008
03:26:00: Options: /TP /nologo /EHa /wd4297 /wd4103 /Ox /MT
03:26:00: Platform: win32 10
03:26:00: Bits: 32
03:26:00: Mode: Release
03:26:00:******************************* System ********************************
03:26:00: CPU: AMD Ryzen Threadripper 1950X 16-Core Processor
03:26:00: CPU ID: AuthenticAMD Family 23 Model 1 Stepping 1
03:26:00: CPUs: 32
03:26:00: Memory: 63.90GiB
03:26:00: Free Memory: 58.69GiB
03:26:00: Threads: WINDOWS_THREADS
03:26:00: OS Version: 6.2
03:26:00: Has Battery: false
03:26:00: On Battery: false
03:26:00: UTC Offset: -6
03:26:00: PID: 20632
03:26:00: CWD: Q:\FAH
03:26:00: OS: Windows 10 Enterprise
03:26:00: OS Arch: AMD64
03:26:00: GPUs: 2
03:26:00: GPU 0: Bus:10 Slot:0 Func:0 NVIDIA:7 GP104 [GeForce GTX 1070 Ti] 8186
03:26:00: GPU 1: Bus:67 Slot:0 Func:0 NVIDIA:7 GP104 [GeForce GTX 1070 Ti] 8186
03:26:00: CUDA Device 0: Platform:0 Device:0 Bus:10 Slot:0 Compute:6.1 Driver:10.1
03:26:00: CUDA Device 1: Platform:0 Device:1 Bus:67 Slot:0 Compute:6.1 Driver:10.1
03:26:00:OpenCL Device 0: Platform:0 Device:0 Bus:10 Slot:0 Compute:1.2 Driver:432.0
03:26:00:OpenCL Device 1: Platform:0 Device:1 Bus:67 Slot:0 Compute:1.2 Driver:432.0
03:26:00: Win32 Service: false
03:26:00:***********************************************************************
03:26:00:<config>
03:26:00: <!-- Slot Control -->
03:26:00: <power v='FULL'/>
03:26:00:
03:26:00: <!-- User Information -->
03:26:00: <team v='250915'/>
03:26:00: <user v='TyMarbut'/>
03:26:00:
03:26:00: <!-- Folding Slots -->
03:26:00: <slot id='0' type='CPU'/>
03:26:00: <slot id='1' type='GPU'/>
03:26:00: <slot id='2' type='GPU'/>
03:26:00:</config>
03:26:00:Trying to access database...
03:26:00:Successfully acquired database lock
03:26:00:Enabled folding slot 00: READY cpu:30
03:26:00:Enabled folding slot 01: READY gpu:0:GP104 [GeForce GTX 1070 Ti] 8186
03:26:00:Enabled folding slot 02: READY gpu:1:GP104 [GeForce GTX 1070 Ti] 8186
03:26:01:WU00:FS00:Connecting to 65.254.110.245:8080
03:26:01:WU01:FS01:Connecting to 65.254.110.245:8080
03:26:01:WU02:FS02:Connecting to 65.254.110.245:8080
03:26:01:WARNING:WU00:FS00:Failed to get assignment from '65.254.110.245:8080': No WUs available for this configuration
03:26:01:WU00:FS00:Connecting to 18.218.241.186:80
03:26:01:WARNING:WU01:FS01:Failed to get assignment from '65.254.110.245:8080': No WUs available for this configuration
03:26:01:WU01:FS01:Connecting to 18.218.241.186:80
03:26:01:WARNING:WU02:FS02:Failed to get assignment from '65.254.110.245:8080': No WUs available for this configuration
03:26:01:WU02:FS02:Connecting to 18.218.241.186:80
03:26:01:WARNING:WU00:FS00:Failed to get assignment from '18.218.241.186:80': No WUs available for this configuration
03:26:01:ERROR:WU00:FS00:Exception: Could not get an assignment
Re: F@H using only 1 of 2 (identical) GPUs - is this expecte
Posted: Fri May 01, 2020 6:57 am
by PantherX
Welcome to the F@H Forum TyMarbut,
Please note that you're missing a passkey which is recommenced to use for security reasons and bonus points:
https://foldingathome.org/support/faq/points/passkey/
It seems that your system is correctly configured. There's a high demand for GPU WUs and the supply can't meet it yet. Work is underway to resolve that so just leave your client running and it will be assigned WUs when they are available. If you would like to double check, use GPU-Z to verify that the GPUs are being used correctly.