Page 1 of 2
No new SMP WU for several WEEKS(!?)
Posted: Sat Apr 23, 2016 3:19 pm
by ewal
Hi
I´ve been running SMP-clients since I got my first multicore computer many years ago. they have almost always run perfectly but lately there has been long periods with no work and now the computer has been running dry for several weeks! Currently I have only one computer online because I can´t keep wasting electricity like this just waiting. What´s the problem?
The performance has also been very bad lately, but that might have to do with Windows. It should have been reinstalled years ago, but I can´t do that right now.
Thankful for all help and a qick reply. (I don´t like falling in rank so fast every day)
I use client version 6.34. (I have to, because I don´t think newer versions work with HFM.NET)
Hardware: Core 2 Quad Q9450 @ 3200 MHz, 8 GB DDR3 @ 1600 MHz, intel X38 chipset
OS: Windows 7 Home premium SP1. 64 bit, swedish
My last log:
------------------------------------------------------
Code: Select all
[14:27:23] Could not open work queue, generating new queue...
[14:27:24] - Preparing to get new work unit...
[14:27:24] - Autosending finished units... [April 23 14:27:24 UTC]
[14:27:24] Cleaning up work directory
[14:27:24] Trying to send all finished work units
[14:27:24] + Attempting to get work packet
[14:27:24] + No unsent completed units remaining.
[14:27:24] Passkey found
[14:27:24] - Autosend completed
[14:27:24] - Will indicate memory of 8191 MB
[14:27:24] - Detect CPU. Vendor: GenuineIntel, Family: 6, Model: 7, Stepping: 7
[14:27:24] - Connecting to assignment server
[14:27:24] Connecting to http://assign.stanford.edu:8080/
[14:27:24] Posted data.
[14:27:24] Initial: 0000; + No appropriate work server was available; will try a
gain in a bit.
[14:27:24] + Couldn't get work instructions.
[14:27:24] - Attempt #1 to get work failed, and no other work to do.
Waiting before retry.
[14:27:35] + Attempting to get work packet
[14:27:35] Passkey found
[14:27:35] - Will indicate memory of 8191 MB
[14:27:35] - Connecting to assignment server
[14:27:35] Connecting to http://assign.stanford.edu:8080/
[14:27:35] Posted data.
[14:27:35] Initial: 0000; + No appropriate work server was available; will try a
gain in a bit.
[14:27:35] + Couldn't get work instructions.
[14:27:35] - Attempt #2 to get work failed, and no other work to do.
Waiting before retry.
[14:27:53] + Attempting to get work packet
[14:27:53] Passkey found
[14:27:53] - Will indicate memory of 8191 MB
[14:27:53] - Connecting to assignment server
[14:27:53] Connecting to http://assign.stanford.edu:8080/
[14:27:54] Posted data.
[14:27:54] Initial: 0000; + No appropriate work server was available; will try a
gain in a bit.
[14:27:54] + Couldn't get work instructions.
[14:27:54] - Attempt #3 to get work failed, and no other work to do.
Waiting before retry.
[14:28:21] + Attempting to get work packet
[14:28:21] Passkey found
[14:28:21] - Will indicate memory of 8191 MB
[14:28:21] - Connecting to assignment server
[14:28:21] Connecting to http://assign.stanford.edu:8080/
[14:28:22] Posted data.
[14:28:22] Initial: 0000; + No appropriate work server was available; will try a
gain in a bit.
[14:28:22] + Couldn't get work instructions.
[14:28:22] - Attempt #4 to get work failed, and no other work to do.
Waiting before retry.
[14:29:16] + Attempting to get work packet
[14:29:16] Passkey found
[14:29:16] - Will indicate memory of 8191 MB
[14:29:16] - Connecting to assignment server
[14:29:16] Connecting to http://assign.stanford.edu:8080/
[14:29:16] Posted data.
[14:29:16] Initial: 0000; + No appropriate work server was available; will try a
gain in a bit.
[14:29:16] + Couldn't get work instructions.
[14:29:16] - Attempt #5 to get work failed, and no other work to do.
Waiting before retry.
[14:30:37] + Attempting to get work packet
[14:30:43] Passkey found
[14:30:43] - Will indicate memory of 8191 MB
[14:30:43] - Connecting to assignment server
[14:30:43] Connecting to http://assign.stanford.edu:8080/
[14:30:44] Posted data.
[14:30:44] Initial: 0000; + No appropriate work server was available; will try a
gain in a bit.
[14:30:44] + Couldn't get work instructions.
[14:30:44] - Attempt #6 to get work failed, and no other work to do.
Waiting before retry.
...and so on.
----------------------------------
Mod edit: added Code tags to log file posting
Re: No new SMP WU for several WEEKS(!?)
Posted: Sat Apr 23, 2016 4:12 pm
by Joe_H
I would strongly encourage you to upgrade your client to version 7. Fewer servers are supporting connections to version 6 clients, and require V7 as a minimum.
As for HFM.net, it was upgraded to work with V7 a long time ago. Recent updates were done to work with changes in the supported format for information from the project summary.
Re: No new SMP WU for several WEEKS(!?)
Posted: Sat Apr 23, 2016 4:21 pm
by bruce
I use both HFM and the current V7 client without difficulties.
There have been some shortages of CPU projects for specific numbers of cores but there have not been shortages for commonplace hardware. I've been pushing FAH to be sure that some projects are available for a wide range of CPU-counts, but a lot of projects simply do not run well without limitations. It's impossible for any single project to work with 1-CPU and with 64-CPUs. (The same is true for much narrower ranges.)
Re: No new SMP WU for several WEEKS(!?)
Posted: Wed Apr 27, 2016 4:50 pm
by ewal
Thank you so much for your help
Now 7.4.4 runs. I haven't testet with HFM yet,
BUT it doesn´t utilize all the available processing power. With version 6 the CPU usage was always 100,0%
Now it veries between ~85-95% ..and yes, the Power is set to "full" in the web control panel.
Any thoughts?
Re: No new SMP WU for several WEEKS(!?)
Posted: Wed Apr 27, 2016 4:55 pm
by Joe_H
Please post the first 100 or so lines of your log file to show the system information and folding configuration. Bruce's signature has a link to guides on where to locate the log file and how to post it if you need that information.
It is possible that one of the default settings would need to be modified for how you want to run your system for folding. The defaults make a few assumptions that hold for most folders, but not all.
Re: No new SMP WU for several WEEKS(!?)
Posted: Wed Apr 27, 2016 5:22 pm
by ewal
Here you go
Code: Select all
*********************** Log Started 2016-04-26T22:27:24Z ***********************
22:27:24:************************* Folding@home Client *************************
22:27:24: Website: http://folding.stanford.edu/
22:27:24: Copyright: (c) 2009-2014 Stanford University
22:27:24: Author: Joseph Coffland <[email protected]>
22:27:24: Args:
22:27:24: Config: <none>
22:27:24:******************************** Build ********************************
22:27:24: Version: 7.4.4
22:27:24: Date: Mar 4 2014
22:27:24: Time: 20:26:54
22:27:24: SVN Rev: 4130
22:27:24: Branch: fah/trunk/client
22:27:24: Compiler: Intel(R) C++ MSVC 1500 mode 1200
22:27:24: Options: /TP /nologo /EHa /Qdiag-disable:4297,4103,1786,279 /Ox -arch:SSE
22:27:24: /QaxSSE2,SSE3,SSSE3,SSE4.1,SSE4.2 /Qopenmp /Qrestrict /MT /Qmkl
22:27:24: Platform: win32 XP
22:27:24: Bits: 32
22:27:24: Mode: Release
22:27:24:******************************* System ********************************
22:27:24: CPU: Intel(R) Core(TM)2 Quad CPU Q9450 @ 2.66GHz
22:27:24: CPU ID: GenuineIntel Family 6 Model 23 Stepping 7
22:27:24: CPUs: 4
22:27:24: Memory: 8.00GiB
22:27:24: Free Memory: 1.46GiB
22:27:24: Threads: WINDOWS_THREADS
22:27:24: OS Version: 6.1
22:27:24: Has Battery: false
22:27:24: On Battery: false
22:27:24: UTC Offset: 2
22:27:24: PID: 24132
22:27:24: CWD: C:/Windows/system32
22:27:24: OS: Windows 7 Home Premium
22:27:24: OS Arch: AMD64
22:27:24: GPUs: 1
22:27:24: GPU 0: NVIDIA:1 GT200b [GeForce GTX 285]
22:27:24: CUDA: Not detected
22:27:24:Win32 Service: true
22:27:24:***********************************************************************
22:27:24:<config>
22:27:24: <!-- Folding Slots -->
22:27:24:</config>
22:27:24:Connecting to assign-GPU.stanford.edu:80
22:27:25:Updated GPUs.txt
22:27:26:Read GPUs.txt
22:27:26:Trying to access database...
22:27:27:Successfully acquired database lock
22:27:27:Enabled folding slot 00: PAUSED cpu:2 (not configured)
22:27:27:Enabled folding slot 01: PAUSED gpu:0:GT200b [GeForce GTX 285] (no gpu access)
22:27:33:8:127.0.0.1:New Web connection
22:28:28:Saving configuration to config.xml
22:28:28:<config>
22:28:28: <!-- Folding Slots -->
22:28:28: <slot id='0' type='CPU'/>
22:28:28: <slot id='1' type='GPU'/>
22:28:28:</config>
22:28:28:Set client configured
22:28:29:WU00:FS00:Connecting to 171.67.108.45:8080
22:28:30:WU00:FS00:Connecting to 171.67.108.45:8080
22:28:30:WU00:FS00:Assigned to work server 171.67.108.158
22:28:30:WU00:FS00:Requesting new work unit for slot 00: READY cpu:2 from 171.67.108.158
22:28:30:WU00:FS00:Connecting to 171.67.108.158:8080
22:28:32:WU00:FS00:Downloading 806.35KiB
22:28:33:WU00:FS00:Download complete
22:28:33:WU00:FS00:Received Unit: id:00 state:DOWNLOAD error:NO_ERROR project:9039 run:585 clone:0 gen:127 core:0xa4 unit:0x00000094ab436c9e569827e1f0f5b34f
22:28:34:WU00:FS00:Downloading core from http://web.stanford.edu/~pande/Win32/AMD64/Core_a4.fah
22:28:34:WU00:FS00:Connecting to web.stanford.edu:80
22:28:34:WU00:FS00:FahCore a4: Downloading 2.89MiB
22:28:40:WU00:FS00:FahCore a4: 54.10%
22:28:43:WU00:FS00:FahCore a4: Download complete
22:28:44:WU00:FS00:Valid core signature
22:28:44:WU00:FS00:Unpacked 9.59MiB to cores/web.stanford.edu/~pande/Win32/AMD64/Core_a4.fah/FahCore_a4.exe
22:28:44:WU00:FS00:Starting
22:28:44:WU00:FS00:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:/ProgramData/FAHClient/cores/web.stanford.edu/~pande/Win32/AMD64/Core_a4.fah/FahCore_a4.exe -dir 00 -suffix 01 -version 704 -lifeline 24132 -checkpoint 15 -np 2 -service
22:28:44:WU00:FS00:Started FahCore on PID 25900
22:28:45:WU00:FS00:Core PID:33240
22:28:45:WU00:FS00:FahCore 0xa4 started
22:28:45:WU00:FS00:0xa4:
22:28:45:WU00:FS00:0xa4:*------------------------------*
22:28:45:WU00:FS00:0xa4:Folding@Home Gromacs GB Core
22:28:45:WU00:FS00:0xa4:Version 2.27 (Dec. 15, 2010)
22:28:45:WU00:FS00:0xa4:
22:28:45:WU00:FS00:0xa4:Preparing to commence simulation
22:28:45:WU00:FS00:0xa4:- Looking at optimizations...
22:28:45:WU00:FS00:0xa4:- Created dyn
22:28:45:WU00:FS00:0xa4:- Files status OK
22:28:45:WU00:FS00:0xa4:- Expanded 825187 -> 1398040 (decompressed 169.4 percent)
22:28:45:WU00:FS00:0xa4:Called DecompressByteArray: compressed_data_size=825187 data_size=1398040, decompressed_data_size=1398040 diff=0
22:28:45:WU00:FS00:0xa4:- Digital signature verified
22:28:45:WU00:FS00:0xa4:
22:28:45:WU00:FS00:0xa4:Project: 9039 (Run 585, Clone 0, Gen 127)
22:28:45:WU00:FS00:0xa4:
22:28:45:WU00:FS00:0xa4:Assembly optimizations on if available.
22:28:45:WU00:FS00:0xa4:Entering M.D.
22:28:51:WU00:FS00:0xa4:Mapping NT from 2 to 2
22:28:52:WU00:FS00:0xa4:Completed 0 out of 250000 steps (0%)
22:30:20:FS00:Shutting core down
22:30:20:WARNING:FS00:FahCore not accepting gentle shutdown, killing
22:30:20:WARNING:FS00:Killing WU00
22:30:21:WU00:FS00:FahCore terminated
22:30:21:WU00:FS00:Starting
22:30:21:WARNING:WU00:FS00:Changed SMP threads from 2 to 3 this can cause some work units to fail
22:30:21:WU00:FS00:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:/ProgramData/FAHClient/cores/web.stanford.edu/~pande/Win32/AMD64/Core_a4.fah/FahCore_a4.exe -dir 00 -suffix 01 -version 704 -lifeline 24132 -checkpoint 15 -np 3 -service
22:30:21:WU00:FS00:Started FahCore on PID 21256
22:30:21:WU00:FS00:Core PID:33416
22:30:21:WU00:FS00:FahCore 0xa4 started
22:30:21:WU00:FS00:0xa4:
22:30:21:WU00:FS00:0xa4:*------------------------------*
22:30:21:WU00:FS00:0xa4:Folding@Home Gromacs GB Core
22:30:21:WU00:FS00:0xa4:Version 2.27 (Dec. 15, 2010)
22:30:21:WU00:FS00:0xa4:
22:30:21:WU00:FS00:0xa4:Preparing to commence simulation
22:30:21:WU00:FS00:0xa4:- Ensuring status. Please wait.
22:30:30:Saving configuration to config.xml
22:30:30:<config>
22:30:30: <!-- Slot Control -->
22:30:30: <power v='FULL'/>
22:30:30:
22:30:30: <!-- User Information -->
22:30:30: <passkey v='********************************'/>
22:30:30: <team v='42956'/>
22:30:30: <user v='ew_kalmar_sweden'/>
22:30:30:
22:30:30: <!-- Folding Slots -->
22:30:30: <slot id='0' type='CPU'/>
22:30:30: <slot id='1' type='GPU'/>
22:30:30:</config>
22:30:31:WU00:FS00:0xa4:- Looking at optimizations...
22:30:31:WU00:FS00:0xa4:- Working with standard loops on this execution.
22:30:31:WU00:FS00:0xa4:- Previous termination of core was improper.
22:30:31:WU00:FS00:0xa4:- Files status OK
22:30:31:WU00:FS00:0xa4:- Expanded 825187 -> 1398040 (decompressed 169.4 percent)
22:30:31:WU00:FS00:0xa4:Called DecompressByteArray: compressed_data_size=825187 data_size=1398040, decompressed_data_size=1398040 diff=0
22:30:31:WU00:FS00:0xa4:- Digital signature verified
22:30:31:WU00:FS00:0xa4:
22:30:31:WU00:FS00:0xa4:Project: 9039 (Run 585, Clone 0, Gen 127)
22:30:31:WU00:FS00:0xa4:
22:30:31:WU00:FS00:0xa4:Entering M.D.
22:30:37:WU00:FS00:0xa4:Mapping NT from 3 to 3
22:30:37:WU00:FS00:0xa4:Completed 0 out of 250000 steps (0%)
22:33:37:WU00:FS00:0xa4:Completed 2500 out of 250000 steps (1%)
22:36:41:WU00:FS00:0xa4:Completed 5000 out of 250000 steps (2%)
22:39:34:WU00:FS00:0xa4:Completed 7500 out of 250000 steps (3%)
22:42:25:WU00:FS00:0xa4:Completed 10000 out of 250000 steps (4%)
22:45:15:WU00:FS00:0xa4:Completed 12500 out of 250000 steps (5%)
22:48:05:WU00:FS00:0xa4:Completed 15000 out of 250000 steps (6%)
22:50:53:WU00:FS00:0xa4:Completed 17500 out of 250000 steps (7%)
22:53:41:WU00:FS00:0xa4:Completed 20000 out of 250000 steps (8%)
22:56:30:WU00:FS00:0xa4:Completed 22500 out of 250000 steps (9%)
22:59:20:WU00:FS00:0xa4:Completed 25000 out of 250000 steps (10%)
Thanks in advance.
Mod edit: Added Code tags to log file
Re: No new SMP WU for several WEEKS(!?)
Posted: Wed Apr 27, 2016 6:57 pm
by Joe_H
Okay, looking at the log file, the client has detected an usable GPU and reserved one CPU core for processing connected with that. As an older pre-Fermi GPU, it is semi supported. The projects that used to be assigned to GPU's of that era have run out, some pre-Fermi cards can still fold on some later projects using another end-of-life folding core.
You have two options. First is just delete the GPU folding slot, that should free up the CPU core and let the client fold on all 4 cores at the Full setting.
The second option is more involved. You would need to install video drivers that provide CUDA support and then configure the GPU slot so it would request WU's from the projects that use Core_15. The extra configuration is needed because pre-Fermi cards are not automatically assigned to these projects and later GPU's are on an opt-in basis as well.
Re: No new SMP WU for several WEEKS(!?)
Posted: Wed Apr 27, 2016 8:19 pm
by ewal
But all the cores are running at ~90% Not just one.
A random screenshot from Open Hardware Monitor shows:
CPU Total 87,7%
CPU Core #1 90,8%
CPU Core #2 89.2%
CPU Core #3 87.7%
CPU Core #4 83.1%
I have fairly recent drivers for the GPU (340.52), both OpenCL, CUDA, PhysX and DirectCompute 4.0 are marked in the "computing"-area in GPU-Z AND I have no problem running memtestG80, so I don´t understand what I´m supposed to be lacking.
How do I tell F@H to use only Core_15?
Best regards / Erik W
Edit: I´m running F@H as a service
Re: No new SMP WU for several WEEKS(!?)
Posted: Wed Apr 27, 2016 9:32 pm
by 7im
Correct. Windows will balance the 100% load from 3 cores over all 4 of the cores. That is the nature of load balancing in Windows.
Windows does not allow GPU access from any program running as a service, so no GPU folding while running the client as a service.
Another option, if you do not want to fold on the GPU slot is to leave that slot paused, and configure the CPU slot to use all 4 CPU cores.
Open FAHControl (advanced control). Click the configure button. Select the slots tab. Select the CPU slot and click the Edit button. Change the number of CPU threads from -1 (automatic) to 4. Click OK, click Save. CPU usage should go up.
Re: No new SMP WU for several WEEKS(!?)
Posted: Wed Apr 27, 2016 10:40 pm
by Joe_H
7im has answered your questions, I missed the "Win32 Service: true" entry in your log file.
If you do decide that you want to try GPU folding on your system, first you would need to uninstall folding as a service and reinstall. The options you would need to add to the GPU slot are the following:
max-packet-size with a setting of small
client-type set to beta
The projects aren't actually beta anymore, but the flag is being used to indicate that you want to do Core_15 folding on your GPU. There were no flags already defined and available to use when this situation came up.
Re: No new SMP WU for several WEEKS(!?)
Posted: Wed Apr 27, 2016 10:46 pm
by bruce
Reinstalling will remove the service, though a reboot may be required.
Re: No new SMP WU for several WEEKS(!?)
Posted: Thu Apr 28, 2016 4:37 am
by Phildesign
I am getting the same issue.
I got a server to do my folding as my gaming laptop was getting got and loud, it does not have any problem picking up folding jobs.
But the server goes days without picking up jobs, at the moment its just running cpu jobs as I waiting for bios updates.
I deleted the laptop folding in case its holding jobs back.
Help
Btw I got boinc going while folding decides to give jobs to my server......
Re: No new SMP WU for several WEEKS(!?)
Posted: Thu Apr 28, 2016 4:45 am
by Joe_H
Welcome to the folding forum.
If your server is running version 6 of the client, the answer is the same as given to ewal. We recommend upgrading to the latest version 7 client. Otherwise we would need more information about your system to suggest methods to have it get work. There is work available for most CPU folding setups, yours may fall outside the the norm in one way or another.
P.S. I have three systems at home, none have had any issues getting CPU work.
Re: No new SMP WU for several WEEKS(!?)
Posted: Fri Apr 29, 2016 2:47 am
by Phildesign
Seems to working again, very low points reward but at least its running.
I have 3 systems, one is running now. one is being built, Pretty much the same config.
Dual Xeon e5-2670v1, 32-64 gb ram, gpus to come later.
Thanks
Re: No new SMP WU for several WEEKS(!?)
Posted: Mon May 02, 2016 4:54 am
by WhitehawkEQ
Is A5 WU's a thing of the past now? Or are we going to see more? I have 2 quad socket servers with 4 Opteron 6276 16 core, for a total core count of 64 for each server. I don't need to have these sitting doing nothing like the others have said.
So "if" A5's are gone, is there anything coming down the pipeline that I can use my servers for?
I'm using V7.4.4