Page 1 of 1
Virtual folding
Posted: Mon Jun 03, 2013 1:34 pm
by ChristianVirtual
Hello,
I'm new to folding and its quite addictive. Right now I have two boxes running: one Windows with i7-2600 and GTX660ti and one machine with a Xeon E3 1240v2, VMWare ESXi 5.1 and Ubuntu 12.4 as a folding guest. This VM has 3 cores assigned and make around 12kPPD with FahCore 0xa3. My question: is this smart or should I better use 2 cores and build a separate 1 core VM for additional folding. I need to keep a bit of CPU for other tasks on the box.
Thanks for any suggestion ...
Re: Virtual folding
Posted: Mon Jun 03, 2013 3:37 pm
by 7im
Keep the three.
Re: Virtual folding
Posted: Thu Jun 06, 2013 10:07 pm
by Tim_H
I would use as many as possible, then reduce the priority for that vm
Re: Virtual folding
Posted: Fri Aug 09, 2013 4:09 pm
by ChristianVirtual
One more question:
I recently upgraded the software form 7.2.9 to 7.3.6 on my virtual folding guest and seems I lost the capability running SMP.
I tried several times to force SMP via config file (smp v=true); but failed. The slot always went back to CPU.
Was fine the software version earlier as I could see in log files.
What would be the steps to ensure SMP comes back ?
System info
Code: Select all
15:49:02:************************* Folding@home Client *************************
15:49:02: Website: http://folding.stanford.edu/
15:49:02: Copyright: (c) 2009-2013 Stanford University
15:49:02: Author: Joseph Coffland <[email protected]>
15:49:02: Args: --child --lifeline 2190 /etc/fahclient/config.xml --run-as
15:49:02: fahclient --pid-file=/var/run/fahclient.pid --daemon
15:49:02: Config: /etc/fahclient/config.xml
15:49:02:******************************** Build ********************************
15:49:02: Version: 7.3.6
15:49:02: Date: Feb 18 2013
15:49:02: Time: 07:24:08
15:49:02: SVN Rev: 3923
15:49:02: Branch: fah/trunk/client
15:49:02: Compiler: GNU 4.4.7
15:49:02: Options: -std=gnu++98 -O3 -funroll-loops -mfpmath=sse -ffast-math
15:49:02: -fno-unsafe-math-optimizations -msse2
15:49:02: Platform: linux2 3.2.0-1-amd64
15:49:02: Bits: 64
15:49:02: Mode: Release
15:49:02:******************************* System ********************************
15:49:02: CPU: Intel(R) Xeon(R) CPU E3-1240 V2 @ 3.40GHz
15:49:02: CPU ID: GenuineIntel Family 6 Model 58 Stepping 9
15:49:02: CPUs: 3
15:49:02: Memory: 995.34MiB
15:49:02:Free Memory: 812.18MiB
15:49:02: Threads: POSIX_THREADS
15:49:02:Has Battery: false
15:49:02: On Battery: false
15:49:02: UTC offset: 9
15:49:02: PID: 2197
15:49:02: CWD: /var/lib/fahclient
15:49:02: OS: Linux 3.2.0-48-generic x86_64
15:49:02: OS Arch: AMD64
15:49:02: GPUs: 0
15:49:02: CUDA: Not detected
15:49:02:***********************************************************************
15:49:02:<config>
15:49:02: <!-- Folding Slot Configuration -->
15:49:02: <power v='full'/>
15:49:02:
15:49:02: <!-- Logging -->
15:49:02: <log-rotate-max v='1024'/>
15:49:02:
15:49:02: <!-- Network -->
15:49:02: <proxy v=':8080'/>
15:49:02:
15:49:02: <!-- Remote Command Server -->
15:49:02: <password v='************'/>
15:49:02:
15:49:02: <!-- User Information -->
15:49:02: <passkey v='********************************'/>
15:49:02: <team v='3446'/>
15:49:02: <user v='ChristianFAH'/>
15:49:02:
15:49:02: <!-- Folding Slots -->
15:49:02:</config>
15:49:02:Switching to user fahclient
15:49:02:Trying to access database...
15:49:02:Successfully acquired database lock
15:49:02:Enabled folding slot 00: READY cpu:3
15:49:03:WU00:FS00:Connecting to assign3.stanford.edu:8080
15:49:04:WU00:FS00:News: Welcome to Folding@Home
15:49:04:WU00:FS00:Assigned to work server 171.64.65.81
15:49:04:WU00:FS00:Requesting new work unit for slot 00: READY cpu:3 from 171.64.65.81
15:49:04:WU00:FS00:Connecting to 171.64.65.81:8080
15:49:04:WU00:FS00:Downloading 368.21KiB
15:49:06:WU00:FS00:Download complete
15:49:06:WU00:FS00:Received Unit: id:00 state:DOWNLOAD error:NO_ERROR project:10450 run:0 clone:1 gen:81 core:0xa4 unit:0x000000790a3b1e75515c687c15780001
15:49:06:WU00:FS00:Downloading core from http://www.stanford.edu/~pande/Linux/AMD64/Core_a4.fah
15:49:06:WU00:FS00:Connecting to www.stanford.edu:80
15:49:06:WU00:FS00:FahCore a4: Downloading 2.56MiB
15:49:12:WU00:FS00:FahCore a4: 51.29%
15:49:16:WU00:FS00:FahCore a4: Download complete
15:49:16:WU00:FS00:Valid core signature
15:49:16:WU00:FS00:Unpacked 5.98MiB to cores/www.stanford.edu/~pande/Linux/AMD64/Core_a4.fah/FahCore_a4
15:49:16:WU00:FS00:Starting
15:49:16:WU00:FS00:Running FahCore: /usr/bin/FAHCoreWrapper /var/lib/fahclient/cores/www.stanford.edu/~pande/Linux/AMD64/Core_a4.fah/FahCore_a4 -dir 00 -suffix 01 -version 703 -lifeline 2197 -checkpoint 15 -np 3
15:49:16:WU00:FS00:Started FahCore on PID 2209
15:49:16:WU00:FS00:Core PID:2213
15:49:16:WU00:FS00:FahCore 0xa4 started
15:49:17:WU00:FS00:0xa4:
15:49:17:WU00:FS00:0xa4:*------------------------------*
15:49:17:WU00:FS00:0xa4:Folding@Home Gromacs GB Core
15:49:17:WU00:FS00:0xa4:Version 2.27 (Dec. 15, 2010)
15:49:17:WU00:FS00:0xa4:
15:49:17:WU00:FS00:0xa4:Preparing to commence simulation
15:49:17:WU00:FS00:0xa4:- Looking at optimizations...
15:49:17:WU00:FS00:0xa4:- Created dyn
15:49:17:WU00:FS00:0xa4:- Files status OK
15:49:17:WU00:FS00:0xa4:- Expanded 376540 -> 690140 (decompressed 183.2 percent)
15:49:17:WU00:FS00:0xa4:Called DecompressByteArray: compressed_data_size=376540 data_size=690140, decompressed_data_size=690140 diff=0
15:49:17:WU00:FS00:0xa4:- Digital signature verified
15:49:17:WU00:FS00:0xa4:
15:49:17:WU00:FS00:0xa4:Project: 10450 (Run 0, Clone 1, Gen 81)
15:49:17:WU00:FS00:0xa4:
15:49:17:WU00:FS00:0xa4:Assembly optimizations on if available.
15:49:17:WU00:FS00:0xa4:Entering M.D.
15:49:23:WU00:FS00:0xa4:Completed 0 out of 2000000 steps (0%)
15:50:04:FS00:Finishing
15:51:29:Saving configuration to /etc/fahclient/config.xml
15:51:29:<config>
15:51:29: <!-- Folding Slot Configuration -->
15:51:29: <power v='full'/>
15:51:29:
15:51:29: <!-- Logging -->
15:51:29: <log-rotate-max v='1024'/>
15:51:29:
15:51:29: <!-- Network -->
15:51:29: <proxy v=':8080'/>
15:51:29:
15:51:29: <!-- Remote Command Server -->
15:51:29: <password v='************'/>
15:51:29:
15:51:29: <!-- User Information -->
15:51:29: <passkey v='********************************'/>
15:51:29: <team v='3446'/>
15:51:29: <user v='ChristianFAH'/>
15:51:29:
15:51:29: <!-- Folding Slots -->
15:51:29: <slot id='0' type='CPU'/>
15:51:29:</config>
15:52:35:WU00:FS00:0xa4:Completed 20000 out of 2000000 steps (1%)
15:55:47:WU00:FS00:0xa4:Completed 40000 out of 2000000 steps (2%)
Re: Virtual folding
Posted: Fri Aug 09, 2013 4:29 pm
by bruce
In V7, the former uniprocessor slot and the former SMP slot have been merged into a slot called CPU. We sometimes casually refer to it as SMP, but the number of CPUs can also be set to 1 so it got a new name. Note that you have a slot that's cpu:3 that's using FahCore_a4 which has progressed to 2%.
It looks fine to me.
Re: Virtual folding
Posted: Fri Aug 09, 2013 9:08 pm
by ChristianVirtual
Thanks Bruce,
Seems I got confused by this. I read about the merger but also in 7.2.9 I still saw SMP slot in config and log files.
The above CPU:3 slot got auto detected and (unintended) auto started after removing my manual slot (which had pause-on-start). Once that slot is finished I will reapply my other options again I had in my old config.
Re: Virtual folding
Posted: Sat Aug 10, 2013 12:49 am
by bruce
Perhaps you were confused by the way V7 handles invalid arguments. In V7.3.6 can go to Configure + Slots and under "Add" I have two choices: I can add a CPU slot or a GPU slot. I cannot add an SMP slot. That doesn't prevent me from thinking I need a SMP slot and manually editing config.xml to define one (or from using an old version of config.xml that had one). Nevertheless, when I start FAHControl the formerly valid statements will be parsed and will be removed. Similarly, I can go to Configure + Expert and I can add an Extra_client_option saying smp=true which will do absolutely nothing, since it's now only correct if it says cpu=true. (Notice that it disappears from the listing in the log.)
If you have one concept in your mind (perhaps one that was valid in an earlier version) and the current version supports something else, there are no clear error messages explaining what you've done wrong. I might not have written the client this way, but it is the way it is.