Added 2nd GTX 580, second GPU slot. Both slots use same GPU

[debs3759]

I was folding with 6 CPU threads and one GPU (GTX 580) with no problems. I added a second GTX 580 then added a second GPU slot.

Both GPU slots started running, then I opened MSI Afterburner to check how much of each GPU is being used. Strangely, GPU0 (the new GPU was operating at the expected 99%. GPU1 (the original GPU, definitely working) was at 0%. I paused each slot separately and checked the GPU usage. With just one slot (no matter which) the GPU usage remained at 99%, with GPU1 sitting idle. Both slots are using 12% of the CPU (ie one full thread each, as expected)

My hardware is:

Maximus VIII Hero
Intel i7 6700K
8 GB DDR4 (can't remember the speed)
2 x GTX580

OS is Windows 10 x64 with all updates
NVidia drivers: 368.39 (downloaded today from NVidia, clean install)

Finally, I access this system via Windows Remote Desktop. Hoping that is not the problem.

Looking at the log file, it looks like it is trying to run on both GPUs. Has anyone any idea how to fix this?

Code: Select all

*********************** Log Started 2016-06-12T12:47:03Z ***********************
12:47:03:************************* Folding@home Client *************************
12:47:03:      Website: http://folding.stanford.edu/
12:47:03:    Copyright: (c) 2009-2014 Stanford University
12:47:03:       Author: Joseph Coffland <joseph@cauldrondevelopment.com>
12:47:03:         Args: --open-web-control
12:47:03:       Config: C:/Users/Debs/AppData/Roaming/FAHClient/config.xml
12:47:03:******************************** Build ********************************
12:47:03:      Version: 7.4.4
12:47:03:         Date: Mar 4 2014
12:47:03:         Time: 20:26:54
12:47:03:      SVN Rev: 4130
12:47:03:       Branch: fah/trunk/client
12:47:03:     Compiler: Intel(R) C++ MSVC 1500 mode 1200
12:47:03:      Options: /TP /nologo /EHa /Qdiag-disable:4297,4103,1786,279 /Ox -arch:SSE
12:47:03:               /QaxSSE2,SSE3,SSSE3,SSE4.1,SSE4.2 /Qopenmp /Qrestrict /MT /Qmkl
12:47:03:     Platform: win32 XP
12:47:03:         Bits: 32
12:47:03:         Mode: Release
12:47:03:******************************* System ********************************
12:47:03:          CPU: Intel(R) Core(TM) i7-6700K CPU @ 4.00GHz
12:47:03:       CPU ID: GenuineIntel Family 6 Model 94 Stepping 3
12:47:03:         CPUs: 8
12:47:03:       Memory: 7.93GiB
12:47:03:  Free Memory: 4.96GiB
12:47:03:      Threads: WINDOWS_THREADS
12:47:03:   OS Version: 6.2
12:47:03:  Has Battery: false
12:47:03:   On Battery: false
12:47:03:   UTC Offset: 1
12:47:03:          PID: 5004
12:47:03:          CWD: C:/Users/Debs/AppData/Roaming/FAHClient
12:47:03:           OS: Windows 10 Pro
12:47:03:      OS Arch: AMD64
12:47:03:         GPUs: 2
12:47:03:        GPU 0: NVIDIA:2 GF110 [GeForce GTX 580]
12:47:03:        GPU 1: NVIDIA:2 GF110 [GeForce GTX 580]
12:47:03:         CUDA: 2.0
12:47:03:  CUDA Driver: 8000
12:47:03:Win32 Service: false
12:47:03:***********************************************************************
12:47:03:<config>
12:47:03:  <!-- Network -->
12:47:03:  <proxy v=':8080'/>
12:47:03:
12:47:03:  <!-- Slot Control -->
12:47:03:  <power v='full'/>
12:47:03:
12:47:03:  <!-- User Information -->
12:47:03:  <passkey v='********************************'/>
12:47:03:  <team v='35947'/>
12:47:03:  <user v='debs3759'/>
12:47:03:
12:47:03:  <!-- Folding Slots -->
12:47:03:  <slot id='0' type='CPU'>
12:47:03:    <cpus v='6'/>
12:47:03:    <next-unit-percentage v='100'/>
12:47:03:  </slot>
12:47:03:  <slot id='1' type='GPU'>
12:47:03:    <next-unit-percentage v='100'/>
12:47:03:  </slot>
12:47:03:  <slot id='2' type='GPU'>
12:47:03:    <next-unit-percentage v='100'/>
12:47:03:  </slot>
12:47:03:</config>
12:47:03:Trying to access database...
12:47:03:Successfully acquired database lock
12:47:03:Enabled folding slot 00: READY cpu:6
12:47:03:Enabled folding slot 01: READY gpu:0:GF110 [GeForce GTX 580]
12:47:03:Enabled folding slot 02: READY gpu:1:GF110 [GeForce GTX 580]
12:47:03:WU01:FS01:Starting
12:47:03:WU01:FS01:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:/Users/Debs/AppData/Roaming/FAHClient/cores/web.stanford.edu/~pande/Win32/AMD64/NVIDIA/Fermi/Core_18.fah/FahCore_18.exe -dir 01 -suffix 01 -version 704 -lifeline 5004 -checkpoint 15 -gpu 0 -gpu-vendor nvidia
12:47:03:WU01:FS01:Started FahCore on PID 7804
12:47:03:WU01:FS01:Core PID:5636
12:47:03:WU01:FS01:FahCore 0x18 started
12:47:03:WU02:FS02:Starting
12:47:03:WU02:FS02:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:/Users/Debs/AppData/Roaming/FAHClient/cores/web.stanford.edu/~pande/Win32/AMD64/NVIDIA/Fermi/Core_18.fah/FahCore_18.exe -dir 02 -suffix 01 -version 704 -lifeline 5004 -checkpoint 15 -gpu 1 -gpu-vendor nvidia
12:47:03:WU02:FS02:Started FahCore on PID 8020
12:47:03:WU02:FS02:Core PID:6324
12:47:03:WU02:FS02:FahCore 0x18 started
12:47:03:WU00:FS00:Starting
12:47:03:WU00:FS00:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:/Users/Debs/AppData/Roaming/FAHClient/cores/web.stanford.edu/~pande/Win32/AMD64/Core_a4.fah/FahCore_a4.exe -dir 00 -suffix 01 -version 704 -lifeline 5004 -checkpoint 15 -np 6
12:47:03:WU00:FS00:Started FahCore on PID 6236
12:47:03:WU00:FS00:Core PID:7608
12:47:03:WU00:FS00:FahCore 0xa4 started
12:47:04:WU01:FS01:0x18:*********************** Log Started 2016-06-12T12:47:03Z ***********************
12:47:04:WU01:FS01:0x18:Project: 10490 (Run 5, Clone 0, Gen 417)
12:47:04:WU01:FS01:0x18:Unit: 0x000001d78ca304f45537e8e0d1a06b53
12:47:04:WU01:FS01:0x18:CPU: 0x00000000000000000000000000000000
12:47:04:WU01:FS01:0x18:Machine: 1
12:47:04:WU01:FS01:0x18:Digital signatures verified
12:47:04:WU01:FS01:0x18:Folding@home GPU core18
12:47:04:WU01:FS01:0x18:Version 0.0.4
12:47:04:WU02:FS02:0x18:*********************** Log Started 2016-06-12T12:47:03Z ***********************
12:47:04:WU02:FS02:0x18:Project: 9158 (Run 86, Clone 0, Gen 360)
12:47:04:WU02:FS02:0x18:Unit: 0x00000194ab4041545673c9c95c391aec
12:47:04:WU02:FS02:0x18:CPU: 0x00000000000000000000000000000000
12:47:04:WU02:FS02:0x18:Machine: 2
12:47:04:WU02:FS02:0x18:Digital signatures verified
12:47:04:WU02:FS02:0x18:Folding@home GPU core18
12:47:04:WU02:FS02:0x18:Version 0.0.4
12:47:04:WU00:FS00:0xa4:
12:47:04:WU00:FS00:0xa4:*------------------------------*
12:47:04:WU00:FS00:0xa4:Folding@Home Gromacs GB Core
12:47:04:WU00:FS00:0xa4:Version 2.27 (Dec. 15, 2010)
12:47:04:WU00:FS00:0xa4:
12:47:04:WU00:FS00:0xa4:Preparing to commence simulation
12:47:04:WU00:FS00:0xa4:- Looking at optimizations...
12:47:04:WU00:FS00:0xa4:- Files status OK
12:47:04:WU00:FS00:0xa4:- Expanded 118692 -> 269896 (decompressed 227.3 percent)
12:47:04:WU00:FS00:0xa4:Called DecompressByteArray: compressed_data_size=118692 data_size=269896, decompressed_data_size=269896 diff=0
12:47:04:WU00:FS00:0xa4:- Digital signature verified
12:47:04:WU00:FS00:0xa4:
12:47:04:WU00:FS00:0xa4:Project: 6392 (Run 23, Clone 33, Gen 88)
12:47:04:WU00:FS00:0xa4:
12:47:04:WU00:FS00:0xa4:Assembly optimizations on if available.
12:47:04:WU00:FS00:0xa4:Entering M.D.
12:47:09:WU00:FS00:0xa4:Mapping NT from 6 to 6 
12:47:09:WU00:FS00:0xa4:Completed 0 out of 5000000 steps  (0%)
12:47:13:WU01:FS01:0x18:Completed 0 out of 5000000 steps (0%)
12:47:13:WU01:FS01:0x18:Temperature control disabled. Requirements: single Nvidia GPU, tmax must be < 110 and twait >= 900
12:47:15:WU02:FS02:0x18:Completed 0 out of 2500000 steps (0%)
12:47:15:WU02:FS02:0x18:Temperature control disabled. Requirements: single Nvidia GPU, tmax must be < 110 and twait >= 900
12:49:04:WU00:FS00:0xa4:Completed 50000 out of 5000000 steps  (1%)
12:50:58:WU00:FS00:0xa4:Completed 100000 out of 5000000 steps  (2%)
12:52:53:WU00:FS00:0xa4:Completed 150000 out of 5000000 steps  (3%)
12:54:53:WU00:FS00:0xa4:Completed 200000 out of 5000000 steps  (4%)
12:56:51:WU00:FS00:0xa4:Completed 250000 out of 5000000 steps  (5%)
12:57:18:WU02:FS02:0x18:Completed 25000 out of 2500000 steps (1%)
12:58:44:WU00:FS00:0xa4:Completed 300000 out of 5000000 steps  (6%)
13:00:16:Removing old file 'configs/config-20160612-120637.xml'
13:00:16:Saving configuration to config.xml
13:00:16:<config>
13:00:16:  <!-- Network -->
13:00:16:  <proxy v=':8080'/>
13:00:16:
13:00:16:  <!-- User Information -->
13:00:16:  <passkey v='********************************'/>
13:00:16:  <team v='35947'/>
13:00:16:  <user v='debs3759'/>
13:00:16:
13:00:16:  <!-- Folding Slots -->
13:00:16:  <slot id='0' type='CPU'>
13:00:16:    <cpus v='6'/>
13:00:16:    <next-unit-percentage v='100'/>
13:00:16:  </slot>
13:00:16:  <slot id='1' type='GPU'>
13:00:16:    <next-unit-percentage v='100'/>
13:00:16:  </slot>
13:00:16:  <slot id='2' type='GPU'>
13:00:16:    <next-unit-percentage v='100'/>
13:00:16:  </slot>
13:00:16:</config>
13:00:38:WU00:FS00:0xa4:Completed 350000 out of 5000000 steps  (7%)
13:02:36:WU00:FS00:0xa4:Completed 400000 out of 5000000 steps  (8%)
13:04:37:WU00:FS00:0xa4:Completed 450000 out of 5000000 steps  (9%)
13:05:36:WU01:FS01:0x18:Completed 50000 out of 5000000 steps (1%)
13:06:35:WU00:FS00:0xa4:Completed 500000 out of 5000000 steps  (10%)
13:07:15:WU02:FS02:0x18:Completed 50000 out of 2500000 steps (2%)
13:08:32:WU00:FS00:0xa4:Completed 550000 out of 5000000 steps  (11%)
13:10:28:WU00:FS00:0xa4:Completed 600000 out of 5000000 steps  (12%)
13:12:25:WU00:FS00:0xa4:Completed 650000 out of 5000000 steps  (13%)
13:14:22:WU00:FS00:0xa4:Completed 700000 out of 5000000 steps  (14%)
13:16:21:WU00:FS00:0xa4:Completed 750000 out of 5000000 steps  (15%)
13:17:15:WU02:FS02:0x18:Completed 75000 out of 2500000 steps (3%)
13:18:22:WU00:FS00:0xa4:Completed 800000 out of 5000000 steps  (16%)
13:20:24:WU00:FS00:0xa4:Completed 850000 out of 5000000 steps  (17%)
13:22:30:WU00:FS00:0xa4:Completed 900000 out of 5000000 steps  (18%)
13:24:01:WU01:FS01:0x18:Completed 100000 out of 5000000 steps (2%)
13:24:34:WU00:FS00:0xa4:Completed 950000 out of 5000000 steps  (19%)
13:26:38:WU00:FS00:0xa4:Completed 1000000 out of 5000000 steps  (20%)
13:27:32:WU02:FS02:0x18:Completed 100000 out of 2500000 steps (4%)
13:28:42:WU00:FS00:0xa4:Completed 1050000 out of 5000000 steps  (21%)
13:30:47:WU00:FS00:0xa4:Completed 1100000 out of 5000000 steps  (22%)
13:32:51:WU00:FS00:0xa4:Completed 1150000 out of 5000000 steps  (23%)
13:34:59:WU00:FS00:0xa4:Completed 1200000 out of 5000000 steps  (24%)
13:37:11:WU00:FS00:0xa4:Completed 1250000 out of 5000000 steps  (25%)

[toTOW]

Finally, I access this system via Windows Remote Desktop. Hoping that is not the problem.

This is a problem : in a RDP session, there's is no driver for the GPU. Don't start the client on RDP session.

Since you have two GPUs, I guess one gets disabled in RDP (the "main" one, used to drive the display) but the other remains active, probably because it is detected as a coprocessor (Physics or whatever). If you had only one GPU in your system, the client launched in RDP session wouldn't be able to start any GPU slot.

Try to run the client locally to confirm that everything is set up fine and that you can fold on both GPUs. Also, try VNC instead of RDP if you really need remote access ...

[debs3759]

Problem fixed, although I am unsure how.

I hooked a monitor up to GPU1, and rebooted. Logged in on the server as normal, and both GPUs were folding. I logged back in with Remote Desktop, and hey presto everything is working as expected/desired. I can only guess that after adding the second card, the drivers needed me to log in on the system to initialise the dual card setup.

[7im]

RDP is known to cause issues with folding. Any further problems should start with a change to a different remote client.

[Rel25917]

Has anyone confirmed that windows 10 acts the same as previous versions as far as remote desktop replacing the video driver and killing FaH?

[debs3759]

I have changed my setup. Instead of rdp, I now use an HDMI splitter, and as well as evcerything working better, it has increased my ppd per client.

[Joe_H]

Has anyone confirmed that windows 10 acts the same as previous versions as far as remote desktop replacing the video driver and killing FaH?

Yes it has been confirmed, the protocol used by MS is the same in all versions. It is recommended that a different remote screen access product be used if GPU folding is running and you do not want to interrupt it. Clients based on VNC are not reported to have this problem.

[bruce]

Has anyone confirmed that windows 10 acts the same as previous versions as far as remote desktop replacing the video driver and killing FaH?

It should be noted that if you let Windows Update install your video drivers FAH will not work but the Desktop will work just fine. FAH needs capabilities that nV provides but which MS chooses not to support. Why should we expect it be any different for RDT?