Project 14349 ... AS lowers CPUs to 12

[Neil-B]

Not seen this behaviour before so thought I'd post on it for info: (and posted here as more about the general behaviour of AS stepping down a slot not about the Project WU which I reckon is doing what is intended)

Two slots on my server (client-type = advanced) both received WUs for this project … I noticed TPFs seemed quite long with corresponding drop in PPD of over 2/3rds range I usually see so I checked the logs which showed that the slots were both stepped down to 12 cores for this WU … I have no issue with it this - it seems intended from a look at the beta forum and the WUs folded fine - I just have not seen this behaviour before (had assumed my slots just would not be assigned low core restricted WUs) and the only mention I could find searching the forums was an installation error scenario (which this obviously isn't).

Project: 14349 (Run 78, Clone 2, Gen 1) for 24/56 slot …

Code: Select all

21:30:04:WU00:FS00:Starting
21:30:04:WU01:FS00:Connecting to 13.82.98.119:8080
21:30:04:WARNING:WU00:FS00:AS lowered CPUs from 24 to 12
21:30:04:WU00:FS00:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:\Users\OpDoubleHelix\AppData\Roaming\FAHClient\cores/cores.foldingathome.org/v7/win/64bit/avx/Core_a7.fah/FahCore_a7.exe -dir 00 -suffix 01 -version 705 -lifeline 10928 -checkpoint 5 -np 12
21:30:04:WU00:FS00:Started FahCore on PID 13708
21:30:04:WU00:FS00:Core PID:11612
21:30:04:WU00:FS00:FahCore 0xa7 started
21:30:05:WU00:FS00:0xa7:*********************** Log Started 2020-03-28T21:30:04Z ***********************
21:30:05:WU00:FS00:0xa7:************************** Gromacs Folding@home Core ***************************
21:30:05:WU00:FS00:0xa7:       Type: 0xa7
21:30:05:WU00:FS00:0xa7:       Core: Gromacs
21:30:05:WU00:FS00:0xa7:       Args: -dir 00 -suffix 01 -version 705 -lifeline 13708 -checkpoint 5 -np
21:30:05:WU00:FS00:0xa7:             12
21:30:05:WU00:FS00:0xa7:************************************ CBang *************************************
21:30:05:WU00:FS00:0xa7:       Date: Oct 26 2019
21:30:05:WU00:FS00:0xa7:       Time: 01:38:25
21:30:05:WU00:FS00:0xa7:   Revision: c46a1a011a24143739ac7218c5a435f66777f62f
21:30:05:WU00:FS00:0xa7:     Branch: master
21:30:05:WU00:FS00:0xa7:   Compiler: Visual C++ 2008
21:30:05:WU00:FS00:0xa7:    Options: /TP /nologo /EHa /wd4297 /wd4103 /Ox /MT
21:30:05:WU00:FS00:0xa7:   Platform: win32 10
21:30:05:WU00:FS00:0xa7:       Bits: 64
21:30:05:WU00:FS00:0xa7:       Mode: Release
21:30:05:WU00:FS00:0xa7:************************************ System ************************************
21:30:05:WU00:FS00:0xa7:        CPU: Intel(R) Xeon(R) CPU E5-2697 v3 @ 2.60GHz
21:30:05:WU00:FS00:0xa7:     CPU ID: GenuineIntel Family 6 Model 63 Stepping 2
21:30:05:WU00:FS00:0xa7:       CPUs: 56
21:30:05:WU00:FS00:0xa7:     Memory: 511.75GiB
21:30:05:WU00:FS00:0xa7:Free Memory: 500.03GiB
21:30:05:WU00:FS00:0xa7:    Threads: WINDOWS_THREADS
21:30:05:WU00:FS00:0xa7: OS Version: 6.2
21:30:05:WU00:FS00:0xa7:Has Battery: false
21:30:05:WU00:FS00:0xa7: On Battery: false
21:30:05:WU00:FS00:0xa7: UTC Offset: 0
21:30:05:WU00:FS00:0xa7:        PID: 11612
21:30:05:WU00:FS00:0xa7:        CWD: C:\Users\OpDoubleHelix\AppData\Roaming\FAHClient\work
21:30:05:WU00:FS00:0xa7:******************************** Build - libFAH ********************************
21:30:05:WU00:FS00:0xa7:    Version: 0.0.18
21:30:05:WU00:FS00:0xa7:     Author: Joseph Coffland <joseph@cauldrondevelopment.com>
21:30:05:WU00:FS00:0xa7:  Copyright: 2019 foldingathome.org
21:30:05:WU00:FS00:0xa7:   Homepage: https://foldingathome.org/
21:30:05:WU00:FS00:0xa7:       Date: Oct 26 2019
21:30:05:WU00:FS00:0xa7:       Time: 01:52:30
21:30:05:WU00:FS00:0xa7:   Revision: c1e3513b1bc0c16013668f2173ee969e5995b38e
21:30:05:WU00:FS00:0xa7:     Branch: master
21:30:05:WU00:FS00:0xa7:   Compiler: Visual C++ 2008
21:30:05:WU00:FS00:0xa7:    Options: /TP /nologo /EHa /wd4297 /wd4103 /Ox /MT
21:30:05:WU00:FS00:0xa7:   Platform: win32 10
21:30:05:WU00:FS00:0xa7:       Bits: 64
21:30:05:WU00:FS00:0xa7:       Mode: Release
21:30:05:WU00:FS00:0xa7:************************************ Build *************************************
21:30:05:WU00:FS00:0xa7:       SIMD: avx_256
21:30:05:WU00:FS00:0xa7:********************************************************************************
21:30:05:WU00:FS00:0xa7:Project: 14349 (Run 78, Clone 2, Gen 1)
21:30:05:WU00:FS00:0xa7:Unit: 0x0000000125bb0c305e77dcb906d661f9
21:30:05:WU00:FS00:0xa7:Reading tar file core.xml
21:30:05:WU00:FS00:0xa7:Reading tar file frame1.tpr
21:30:05:WU00:FS00:0xa7:Digital signatures verified
21:30:05:WU00:FS00:0xa7:Calling: mdrun -s frame1.tpr -o frame1.trr -cpt 5 -nt 12
21:30:05:WU00:FS00:0xa7:Steps: first=0 total=2500000
21:30:05:WU00:FS00:0xa7:Completed 1 out of 2500000 steps (0%)
21:30:10:WU01:FS00:Upload 20.15%
21:30:16:WU01:FS00:Upload 53.02%
21:30:22:WU01:FS00:Upload 84.84%
21:30:25:WU01:FS00:Upload complete
21:30:25:WU01:FS00:Server responded WORK_ACK (400)
21:30:25:WU01:FS00:Final credit estimate, 25456.00 points
21:30:25:WU01:FS00:Cleaning up
21:31:15:WU00:FS00:0xa7:Completed 25000 out of 2500000 steps (1%)
21:32:24:WU00:FS00:0xa7:Completed 50000 out of 2500000 steps (2%)

and Project: 14349 (Run 35, Clone 3, Gen 0) for 30/56 slot

Code: Select all

21:07:18:WU03:FS02:Starting
21:07:18:WARNING:WU03:FS02:AS lowered CPUs from 30 to 12
21:07:18:WU03:FS02:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:\Users\OpDoubleHelix\AppData\Roaming\FAHClient\cores/cores.foldingathome.org/v7/win/64bit/avx/Core_a7.fah/FahCore_a7.exe -dir 03 -suffix 01 -version 705 -lifeline 10928 -checkpoint 5 -np 12
21:07:18:WU03:FS02:Started FahCore on PID 14552
21:07:18:WU03:FS02:Core PID:1108
21:07:18:WU03:FS02:FahCore 0xa7 started
21:07:18:WU03:FS02:0xa7:*********************** Log Started 2020-03-28T21:07:18Z ***********************
21:07:18:WU03:FS02:0xa7:************************** Gromacs Folding@home Core ***************************
21:07:18:WU03:FS02:0xa7:       Type: 0xa7
21:07:18:WU03:FS02:0xa7:       Core: Gromacs
21:07:18:WU03:FS02:0xa7:       Args: -dir 03 -suffix 01 -version 705 -lifeline 14552 -checkpoint 5 -np
21:07:18:WU03:FS02:0xa7:             12
21:07:18:WU03:FS02:0xa7:************************************ CBang *************************************
21:07:18:WU03:FS02:0xa7:       Date: Oct 26 2019
21:07:18:WU03:FS02:0xa7:       Time: 01:38:25
21:07:18:WU03:FS02:0xa7:   Revision: c46a1a011a24143739ac7218c5a435f66777f62f
21:07:18:WU03:FS02:0xa7:     Branch: master
21:07:18:WU03:FS02:0xa7:   Compiler: Visual C++ 2008
21:07:18:WU03:FS02:0xa7:    Options: /TP /nologo /EHa /wd4297 /wd4103 /Ox /MT
21:07:18:WU03:FS02:0xa7:   Platform: win32 10
21:07:18:WU03:FS02:0xa7:       Bits: 64
21:07:18:WU03:FS02:0xa7:       Mode: Release
21:07:18:WU03:FS02:0xa7:************************************ System ************************************
21:07:18:WU03:FS02:0xa7:        CPU: Intel(R) Xeon(R) CPU E5-2697 v3 @ 2.60GHz
21:07:18:WU03:FS02:0xa7:     CPU ID: GenuineIntel Family 6 Model 63 Stepping 2
21:07:18:WU03:FS02:0xa7:       CPUs: 56
21:07:18:WU03:FS02:0xa7:     Memory: 511.75GiB
21:07:18:WU03:FS02:0xa7:Free Memory: 497.94GiB
21:07:18:WU03:FS02:0xa7:    Threads: WINDOWS_THREADS
21:07:18:WU03:FS02:0xa7: OS Version: 6.2
21:07:18:WU03:FS02:0xa7:Has Battery: false
21:07:18:WU03:FS02:0xa7: On Battery: false
21:07:18:WU03:FS02:0xa7: UTC Offset: 0
21:07:18:WU03:FS02:0xa7:        PID: 1108
21:07:18:WU03:FS02:0xa7:        CWD: C:\Users\OpDoubleHelix\AppData\Roaming\FAHClient\work
21:07:18:WU03:FS02:0xa7:******************************** Build - libFAH ********************************
21:07:18:WU03:FS02:0xa7:    Version: 0.0.18
21:07:18:WU03:FS02:0xa7:     Author: Joseph Coffland <joseph@cauldrondevelopment.com>
21:07:18:WU03:FS02:0xa7:  Copyright: 2019 foldingathome.org
21:07:18:WU03:FS02:0xa7:   Homepage: https://foldingathome.org/
21:07:18:WU03:FS02:0xa7:       Date: Oct 26 2019
21:07:18:WU03:FS02:0xa7:       Time: 01:52:30
21:07:18:WU03:FS02:0xa7:   Revision: c1e3513b1bc0c16013668f2173ee969e5995b38e
21:07:18:WU03:FS02:0xa7:     Branch: master
21:07:18:WU03:FS02:0xa7:   Compiler: Visual C++ 2008
21:07:18:WU03:FS02:0xa7:    Options: /TP /nologo /EHa /wd4297 /wd4103 /Ox /MT
21:07:18:WU03:FS02:0xa7:   Platform: win32 10
21:07:18:WU03:FS02:0xa7:       Bits: 64
21:07:18:WU03:FS02:0xa7:       Mode: Release
21:07:18:WU03:FS02:0xa7:************************************ Build *************************************
21:07:18:WU03:FS02:0xa7:       SIMD: avx_256
21:07:18:WU03:FS02:0xa7:********************************************************************************
21:07:18:WU03:FS02:0xa7:Project: 14349 (Run 35, Clone 3, Gen 0)
21:07:18:WU03:FS02:0xa7:Unit: 0x0000000025bb0c305e77dcbb74ecec6e
21:07:18:WU03:FS02:0xa7:Reading tar file core.xml
21:07:18:WU03:FS02:0xa7:Reading tar file frame0.tpr
21:07:18:WU03:FS02:0xa7:Digital signatures verified
21:07:18:WU03:FS02:0xa7:Calling: mdrun -s frame0.tpr -o frame0.trr -cpt 5 -nt 12
21:07:18:WU03:FS02:0xa7:Steps: first=0 total=2500000
21:07:19:WU03:FS02:0xa7:Completed 1 out of 2500000 steps (0%)
21:07:23:WU00:FS02:Upload 84.25%
21:07:24:WU00:FS02:Upload complete
21:07:24:WU00:FS02:Server responded WORK_ACK (400)
21:07:24:WU00:FS02:Final credit estimate, 7464.00 points
21:07:24:WU00:FS02:Cleaning up
21:08:36:WU03:FS02:0xa7:Completed 25000 out of 2500000 steps (1%)
21:09:53:WU03:FS02:0xa7:Completed 50000 out of 2500000 steps (2%)

[Joe_H]

There is code within the server and client code to assign a WU that will use the maximum number of threads possible when a request comes in. So if a request comes in for 30 threads, usually it might drop down to 24. But there are a lot of max 12 thread WUs just coming in off Beta testing, so maybe nothing larger was available to assign. I have not seen the code do such a large drop, but that is what appears to have happened here

[Neil-B]

I am actually really happy it happened - I get to have something unusual happen to my kit - makes me feel special !!

[v00d00]

Nothing stops you from running two fahclient instances of 12 cores.

There was mention on beta about a lot of the covid workunits being set to 12 cores max due to limited numbers of atoms.

[Neil-B]

I know I could set a whole load of 12core slots … but have been running larger slots as the work has been there and it gets them done faster … what I was posting on was that I had never seen an AS step a slot down from the cores it was set at to something lower as I was under the mistaken belief that a 12cpu limit assignment just wouldn't be given to a larger core folding slot - but I know it can

Since those two WUs all WUs have made full use of the larger core counts (thereby completing faster) and so I am happy leaving them as is and if the occasional 12core limited WU comes along my server can run a bit cooler than normal for a while ... If of course 12core limited became the standard fodder I'd resize those slots and add a few more for a while.

[v00d00]

It happens from time to time. If there is only 10,000 atoms then it will probably not need a 64 core system to process it, so they will set the core count to something less thereby freeing up the 64 core system for work that needs that kind of power. But if no other workunits are available it will step down that 64 core system and only use 12.