18806 and 18820 Dumped WUs

[Manndolin]

For the last 11 days, my machine will download and attempt a WU from one of these projects before immediately dumping it. These are the only projects that my machine is dumping and it has dumped every single WU from either of these projects. My percent finished remains at 84.3% so my bonus is still applied, but I would still like to resolve this so that these can be run. Both are from the server, fahserver1.flatironinstitute.org

Processor 12th Gen Intel(R) Core(TM) i5-12400F 2.50 GHz

Graphics NVIDIA GeForce RTX 4060 Graphics Card

OS Windows 11 Home

Client v8-4

I have already tried updating my graphics driver from the august version to the latest version, and yielded no change.
Of course I have tried cycling power, etc. No change. I don't run any overclocking software.

Not sure how to proceed in addressing.

[muziqaz]

These are CPU projects and they have nothing to do with your GPU
Something went wrong with your system stability. BIOS update? Some chipset update? Stupid windows update? Maybe you added something to your system? Maybe some antivirus system is flagging fahcores as false positives?
Have you checked your system RAM for stability?
Download y-cruncher and run its stability test. See if that flags your system with anything.

[Manndolin]

Thanks for the reply!

I've only been folding for two weeks, so this isn't a "new" problem as much as an "always" problem. My machine has never successfully run either of these studies. But I will look for BIOS updates or windows updates that I may have missed.

The machine is a PowerSpec G235 Gaming PC used as-is, out of the box.
ASUS PRIME B660M-A AC D4 Motherboard
16GB DDR4-2666 RAM

RAM usage is unremarkable, ~65%.

When folding, only Firefox and Task Manager are open.

[Joe_H]

Only difference I can think of is that Projects 18806 and 18820 both use the A9 folding core. The last I checked that required a processor with support for at least AVX. I know that is a problem for older systems, but I think your board is new enough that any processor installed on it should meet that requirement.

Could you post the beginning section of your log from after a startup, and a section showing one of these WUs being downloaded and starting until it fails and dumps.

[Manndolin]

Okay I figured out the machine log viewer. All of the dumps I looked at had the same error at the end, so here is the one from 18806 WU126 as an examp

09:18:44:I1:WU126:Requesting WU assignment for user Alexander_Mann team 0
09:18:44:I1:WU126:Received WU assignment 68K0k6GyUo3RwNbm23iL3uapggoDZctRKJxWOmCRYX4
09:18:44:I1:WU126:Downloading WU
09:18:46:I1:WU126:Received WU P18806 R9 C23 G365
09:18:46:I3:WU126:Started FahCore on PID 10660
09:18:46:I1:WU126:*********************** Log Started 2025-03-18T09:18:46Z ***********************
09:18:46:I1:WU126:************************** Gromacs Folding@home Core ***************************
09:18:46:I1:WU126: Core: Gromacs
09:18:46:I1:WU126: Type: 0xa9
09:18:46:I1:WU126: Version: 0.0.12
09:18:46:I1:WU126: Author: Joseph Coffland <joseph@cauldrondevelopment.com>
09:18:46:I1:WU126: Copyright: 2022 foldingathome.org
09:18:46:I1:WU126: Homepage: https://foldingathome.org/
09:18:46:I1:WU126: Date: Nov 15 2022
09:18:46:I1:WU126: Time: 13:31:08
09:18:46:I1:WU126: Compiler: Visual C++
09:18:46:I1:WU126: Options: /TP /std:c++17 /nologo /EHa /wd4297 /wd4103 /O2 /Zc:throwingNew /MT
09:18:46:I1:WU126: Platform: win32 10
09:18:46:I1:WU126: Bits: 64
09:18:46:I1:WU126: Mode: Release
09:18:46:I1:WU126: SIMD: avx2_256
09:18:46:I1:WU126: OpenMP: ON
09:18:46:I1:WU126: CUDA: OFF
09:18:46:I1:WU126: OpenCL: OFF
09:18:46:I1:WU126: Args: -dir 68K0k6GyUo3RwNbm23iL3uapggoDZctRKJxWOmCRYX4 -suffix 01
09:18:46:I1:WU126: -version 8.4.9 -lifeline 18008 -np 10
09:18:46:I1:WU126:************************************ libFAH ************************************
09:18:46:I1:WU126: Date: Nov 15 2022
09:18:46:I1:WU126: Time: 13:30:33
09:18:46:I1:WU126: Compiler: Visual C++
09:18:46:I1:WU126: Options: /TP /std:c++14 /nologo /EHa /wd4297 /wd4103 /O2 /Zc:throwingNew /MT
09:18:46:I1:WU126: Platform: win32 10
09:18:46:I1:WU126: Bits: 64
09:18:46:I1:WU126: Mode: Release
09:18:46:I1:WU126:************************************ CBang *************************************
09:18:46:I1:WU126: Date: Nov 15 2022
09:18:46:I1:WU126: Time: 13:29:57
09:18:46:I1:WU126: Compiler: Visual C++
09:18:46:I1:WU126: Options: /TP /std:c++14 /nologo /EHa /wd4297 /wd4103 /O2 /Zc:throwingNew /MT
09:18:46:I1:WU126: Platform: win32 10
09:18:46:I1:WU126: Bits: 64
09:18:46:I1:WU126: Mode: Release
09:18:46:I1:WU126:************************************ System ************************************
09:18:46:I1:WU126: CPU: 12th Gen Intel(R) Core(TM) i5-12400F
09:18:46:I1:WU126: CPU ID: GenuineIntel Family 6 Model 151 Stepping 5
09:18:46:I1:WU126: CPUs: 12
09:18:46:I1:WU126: Memory: 15.82GiB
09:18:46:I1:WU126:Free Memory: 3.96GiB
09:18:46:I1:WU126: Threads: WINDOWS_THREADS
09:18:46:I1:WU126: OS Version: 6.2
09:18:46:I1:WU126:Has Battery: false
09:18:46:I1:WU126: On Battery: false
09:18:46:I1:WU126: UTC Offset: -4
09:18:46:I1:WU126: PID: 10660
09:18:46:I1:WU126: CWD: C:\ProgramData\FAHClient\work
09:18:46:I1:WU126: Exec: C:\ProgramData\FAHClient\cores\gromacs-core-a9\windows-10-64bit\cpu-avx2_256-release\fahcore-a9-windows-10-64bit-cpu-avx2_256-release-0.0.12\FahCore_a9.exe
09:18:46:I1:WU126:********************************************************************************
09:18:46:I1:WU126:Project: 18806 (Run 9, Clone 23, Gen 365)
09:18:46:I1:WU126:Reading tar file core.xml
09:18:46:I1:WU126:Reading tar file frame365.tpr
09:18:46:I1:WU126:Digital signatures verified
09:18:46:I1:WU126:Calling: mdrun -c frame365.gro -s frame365.tpr -x frame365.xtc -cpt 5 -nt 10 -ntmpi 1 -update cpu -nb cpu -bonded cpu -pme cpu -pmefft cpu
09:18:46:I1:WU126:ERROR:The total number of threads requested (10) does not match the thread-MPI ranks (1) times the OpenMP threads (12) requested
09:18:52:E :WU126:Core returned UNKNOWN_ERROR (121)
09:18:52:E :WU126:Run did not produce any results. Dumping WU
09:18:52:I1:WU126:Sending dump report
09:18:52:I1:WU126:Dumped

However, I don't understand CPUs well enough to parse this error.

[muziqaz]

OK, this is not your computer's issue. something weird is happening with the a9 core :/

[Joe_H]

Only thing I can think of is checking if HT is disabled in the BIOS. Your processor has 6 cores, 12 threads supported with HT.

Speaking of HT, I have often found the best efficiency doing CPU folding using the same number of threads as physical cores. You do gain some performance from using the HT threads, but at the cost of increased power usage and heat that is disproportionate to the performance gained folding.

[muziqaz]

These projects would have crapped out on my CPUs, I don't think HT is the deal breaker here. Maybe different ISA support is the thing on that particular CPU :/ isn't this one big.shittle design, where small cores do not support certain SIMD instructions.
Oh it has no efficiency cores
Though I don't know why fahclient is printing this:
09:18:46:I1:WU126: CPUs: 12
And then goes nt 10. I know why it goes nt 10, since that's what fahclient is configured to use.

[Joe_H]

No efficiency cores, and in the "F" variant no iGPU either. Hard to tell if made that way or if they are binned chips from the production line with faulty iGPUs and possibly efficiency cores.

[Joe_H]

Though I don't know why fahclient is printing this:
09:18:46:I1:WU126: CPUs: 12

That is the header portion of the A9 core startup. It just repeats the hardware detected by the client, in this case the type of processor, memory, and total number of CPU threads. It is not part of the client configuration and should not be what is sent as part of a WU request.

For example, core startup on my Mac Studio reports "CPUs: 10" for the 8 performance and 2 efficiency cores. I usually have it set to use 6 threads.

[Manndolin]

Okay I looked at Machine Settings and I see that my Resource Usage is set to 11/12 CPUs. Is that the source of the issue? Should I set that to 12/12?

[muziqaz]

Okay I looked at Machine Settings and I see that my Resource Usage is set to 11/12 CPUs. Is that the source of the issue? Should I set that to 12/12?

If you are folding on GPU, no, leave at 11 out of 12

[Manndolin]

Okay that's what I figured. I'll leave it there then.

[Manndolin]

Okay bellow I've pasted the summary of the "Machines" Page


[Device Name] v8.4.9 11 CPUs, NVIDIA GeForce RTX 4060
Resources Project Progress ETA Core Timeout Actions
NVIDIA GeForce RTX 4060 12702 5.9% 1h 43m 0x23 23h 53m
10 CPUs 12464 0.1% 9h 35m 0xa8 1d 23h

So to me it looks like the program reserves a CPU for the GPU folding on its own. I'm going to try setting it to 12/12 and seeing if this continues to happen.

[muziqaz]

If you fold on all 12 threads, your GPU will choke and will be folding very slowly