You seem to have something wrong with your BOINC client. it's impossible to say what, but your stderr output is just blank, which is not normal or an artifact of these tasks. since this is the same system that you saw weirdness with Asteroids also, i do think you have some kind of problem with BOINC itself. it's impossible for us to guess without access to your system though.

this is what a Windows output should look like: http://www.gpugrid.net/result.php?resultid=33743283

09:54:40 (15568): wrapper (7.9.26016): starting
09:54:40 (15568): wrapper: running python.exe (bin/conda-unpack)
09:54:42 (15568): python.exe exited; CPU time 0.000000
09:54:42 (15568): wrapper: running Library/usr/bin/tar.exe (xjvf input.tar.bz2)
atom.tar
JNK1_m35_m25_0.xml
JNK1_m35_m25_asyncre.cntl
JNK1_m35_m25.inpcrd
JNK1_m35_m25.prmtop
run.bat
run.sh
09:54:43 (15568): Library/usr/bin/tar.exe exited; CPU time 0.000000
09:54:43 (15568): wrapper: running C:/Windows/system32/cmd.exe (/c call run.bat)
ERROR: Invalid requirement: './Acellera-AToM-OpenMM-*'
09:54:46 (15568): C:/Windows/system32/cmd.exe exited; CPU time 0.000000
09:54:46 (15568): app exit status: 0xd
09:54:46 (15568): called boinc_finish(195)

so yes, there is still a problem on Windows (probably something wrong in the run.bat file, or a file missing from the environment package or input files. but you have a larger problem as well.

while troubleshooting your asteroids problem, I had recommended to upgrade your BOINC client, and I think you did that, but you may have performed an in-place upgrade rather than a fresh install. I would recommend removing all aspects of BOINC on this system. completely delete everything. and re-install from a fresh install package. do not keep anything from the previous install.

---

Not sure if he has the same problem, but for me, the past few jobs on Windows are sent to the wrong platform AFAIC.

Host: https://www.gpugrid.net/results.php?hostid=615737
Task example: https://www.gpugrid.net/result.php?resultid=33744637
Error: The operating system cannot run %1

I checked tasks where other hosts subsequently succeeded and they are all Linux.

Windows task failer again, this is a copy of the run file up to the point it hit the error

Setup environment

C:\ProgramData\BOINC\slots\9>set HOMEPATH=C:\ProgramData\BOINC\slots\9

C:\ProgramData\BOINC\slots\9>set PATH=C:\ProgramData\BOINC\slots\9;C:\ProgramData\BOINC\slots\9\Library\usr\bin;C:\ProgramData\BOINC\slots\9\Library\bin;C:\Windows\system32;C:\Windows

C:\ProgramData\BOINC\slots\9>set PYTHONPATH=C:\ProgramData\BOINC\slots\9\Lib\python3.9\site-packages

C:\ProgramData\BOINC\slots\9>set SYSTEMROOT=C:\Windows
Create a temporary directory

C:\ProgramData\BOINC\slots\9>set TEMP=C:\ProgramData\BOINC\slots\9\tmp

C:\ProgramData\BOINC\slots\9>mkdir C:\ProgramData\BOINC\slots\9\tmp
Install AToM

C:\ProgramData\BOINC\slots\9>tar.exe xvf atom.tar
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/.github/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/.github/workflows/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/.github/workflows/publish.yml
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/LICENSE
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/README.md
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/abfe_explicit.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/abfe_explicit_zrestr.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/abfe_structprep.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/async_re.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/atom_nnp_wrapper.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/environment.yml
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/README.md
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/ligands/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/ligands/but.mol2
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/ligands/dap.mol2
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/ligands/dapp.mol2
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/ligands/dmso.mol2
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/ligands/dss.mol2
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/ligands/prop.mol2
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/ligands/thi.mol2
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/receptor/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/receptor/fkbp.pdb
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/scripts/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/scripts/analyze.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/scripts/asyncre_template.cntl
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/scripts/equil_template.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/scripts/free_energies_template.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/scripts/mdlambda_template.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/scripts/mintherm_template.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/scripts/nodefile
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/scripts/prep_template.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/scripts/run_template.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/scripts/runopenmm
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/scripts/setup-atm.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/scripts/setup-settings.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/fkbp/scripts/uwham_analysis.R
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/scripts/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/scripts/analyze.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/scripts/nodefile
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/scripts/runopenmm
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/scripts/uwham_analysis.R
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/temoa-g1/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/temoa-g1/README.md
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/temoa-g1/equil.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/temoa-g1/mdlambda.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/temoa-g1/mintherm.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/temoa-g1/npt.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/temoa-g1/temoa-g1.inpcrd
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/temoa-g1/temoa-g1.prmtop
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/ABFE/temoa-g1/temoa-g1_asyncre.cntl
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/README.md
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/ligands/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/ligands/1H1Q.mol2
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/ligands/1H1Q.sdf
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/ligands/1H1R.mol2
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/ligands/1H1R.sdf
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/ligands/1H1S.mol2
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/ligands/1H1S.sdf
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/ligands/1OI9.mol2
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/ligands/1OI9.sdf
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/ligands/1OIU.mol2
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/ligands/1OIU.sdf
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/ligands/1OIY.mol2
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/ligands/1OIY.sdf
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/receptor/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/receptor/cdk2.pdb
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/scripts/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/scripts/analyze.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/scripts/asyncre_template.cntl
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/scripts/free_energies_template.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/scripts/prep_template.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/scripts/run_template.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/scripts/setup-atm.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/scripts/setup-settings.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/cdk2/scripts/uwham_analysis.R
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/README.md
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/ligands/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/ligands/2d.mol2
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/ligands/2e.mol2
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/ligands/3a.mol2
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/ligands/3b.mol2
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/receptor/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/receptor/eralpha.pdb
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/scripts/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/scripts/analyze.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/scripts/asyncre_template.cntl
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/scripts/equil_template.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/scripts/free_energies_template.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/scripts/mintherm_template.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/scripts/nodefile
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/scripts/prep_template.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/scripts/run_template.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/scripts/runopenmm
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/scripts/setup-atm.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/scripts/setup-settings.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/eralpha/scripts/uwham_analysis.R
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/scripts/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/scripts/analyze.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/scripts/nodefile
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/scripts/runopenmm
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/scripts/uwham_analysis.R
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/temoa-g1-g4/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/temoa-g1-g4/README.md
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/temoa-g1-g4/equil.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/temoa-g1-g4/mintherm.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/temoa-g1-g4/npt.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/temoa-g1-g4/temoa-g1-g4.inpcrd
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/temoa-g1-g4/temoa-g1-g4.prmtop
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/RBFE/temoa-g1-g4/temoa-g1-g4_asyncre.cntl
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/README.md
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/scripts/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/scripts/analyze.sh
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/scripts/nodefile
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/scripts/runopenmm
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/examples/scripts/uwham_analysis.R
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/gibbs_sampling.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/local_openmm_transport.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/ommreplica.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/ommsystem.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/ommworker.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/openmm_async_re.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/rbfe_explicit.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/rbfe_explicit_sync.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/rbfe_explicit_zrestr.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/rbfe_structprep.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/setup.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/sync/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/sync/__init__.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/sync/atm.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/sync/worker.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/temperatureRE_explicit.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/transport.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/utils/
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/utils/__init__.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/utils/logging.conf
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/utils/singal_guard.py
Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6/utils/timer.py

C:\ProgramData\BOINC\slots\9>python.exe -m pip install ./Acellera-AToM-OpenMM-* || exit 13

Just to add, I've now had multiple failers over the last 10 minutes, all of them are failing at the same point.

Hope this helps with the Windows debug.

Ian&Steve C. wrote yesterday:

You seem to have something wrong with your BOINC client. it's impossible to say what, but your stderr output is just blank, which is not normal or an artifact of these tasks. since this is the same system that you saw weirdness with Asteroids also, i do think you have some kind of problem with BOINC itself. it's impossible for us to guess without access to your system though.

...

so yes, there is still a problem on Windows (probably something wrong in the run.bat file, or a file missing from the environment package or input files. but you have a larger problem as well.

while troubleshooting your asteroids problem, I had recommended to upgrade your BOINC client, and I think you did that, but you may have performed an in-place upgrade rather than a fresh install. I would recommend removing all aspects of BOINC on this system. completely delete everything. and re-install from a fresh install package. do not keep anything from the previous install.

yes, you are right, there is obviously something wrong with this BOINC installation. I will remove it and install it from scratch, once the currently running Climateprediction tasks (which use to last up 14 days or even longer) are through.

Nevertheless, it's sad to learn that the Windows version of the ATM app is still faulty.
What I don't understand is: do they not test it before hundreds or thousands faulty tasks are being sent out? In fact, a testrun in their own lab would have shown within 5 minutes that still something is wrong. I think these 5 minutes would be worth the time, right?

fully agree

Nevertheless, it's sad to learn that the Windows version of the ATM app is still faulty.
What I don't understand is: do they not test it before hundreds or thousands faulty tasks are being sent out? In fact, a testrun in their own lab would have shown within 5 minutes that still something is wrong. I think these 5 minutes would be worth the time, right?

Néanmoins, il est triste d'apprendre que la version Windows de l'application ATM est toujours défectueuse.
Ce que je ne comprends pas, c'est : ne le testent-ils pas avant que des centaines ou des milliers de tâches défectueuses ne soient envoyées ? En fait, un test dans leur propre laboratoire aurait montré en 5 minutes que quelque chose ne va toujours pas. Je pense que ces 5 minutes en vaudraient la peine, non ?

---

Nevertheless, it's sad to learn that the Windows version of the ATM app is still faulty.
What I don't understand is: do they not test it before hundreds or thousands faulty tasks are being sent out? In fact, a testrun in their own lab would have shown within 5 minutes that still something is wrong. I think these 5 minutes would be worth the time, right?

Steve, the researcher, in his first few posts about these tasks said that they don't have any Windows machines in the lab.

They only have Linux.

I'll post to Gianni that he needs to help get the Windows apps sorted out.

Been 10 months since this was posted. Where is the "hoped for" windows version? Why are you wasting the potential of all of our Windows machines and new fast GPU's?

The'Energy is NaN' error is still around:
http://gpugrid.net/result.php?resultid=33745995

Nevertheless, it's sad to learn that the Windows version of the ATM app is still faulty.
What I don't understand is: do they not test it before hundreds or thousands faulty tasks are being sent out? In fact, a testrun in their own lab would have shown within 5 minutes that still something is wrong. I think these 5 minutes would be worth the time, right?

Steve, the researcher, in his first few posts about these tasks said that they don't have any Windows machines in the lab.

They only have Linux.

I'll post to Gianni that he needs to help get the Windows apps sorted out.

Hey Keith,

If you contact Gianni, pass on the following info I found from my testing.

There are 2 issues on the same line in this piece of code in run.bat:

 
@echo Install AToM
tar.exe xvf atom.tar
python.exe -m pip install ./Acellera-AToM-OpenMM-* || exit 13
python.exe -m pip list

1. The path separator '/' is wrong for Windows, should be '\' instead. This makes pip install choke. This should be a trivial fix.
2. Windows CMD shell scripts do not support inline expansion of the '*' wildcard. So pip install doesn't find the module in the location it expects, being "Acellera-AToM-OpenMM-*"
There are a few ways to fix this:
- Use the full name of the package folder 'Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6' This also implies that if this '2dd310b8027c68262906a8946f807896b49947b6' is variable, run.bat should be changed every time
- Generate a new atom.tar with a fixed folder name, for example always using 'Acellera-AToM-OpenMM' as the folder name of the package inside atom.tar - and adapting the run.bat pathname to .\Acellera-AToM-OpenMM accordingly
- use some scripting magic to pre-expand the wildcard into a variable (e.g. ATOM) and passing that variable to pip install. Something like this could work, but may have mixed results on different Windows installs - so solution 1 or 2 preferred.

@echo Install AToM
tar.exe xvf atom.tar

set PARM1=.\Acellera-AToM-OpenMM-*
for %%A in (%PARM1%) do set ATOM=%%A

python.exe -m pip install %ATOM% || exit 13
python.exe -m pip list

I posted to Gianni and he replied that he copied my message to Steve.

I will try and get a response from Steve directly via PM and reference your post and analysis.

Heres the top half of my dump:

<core_client_version>7.24.1</core_client_version>
<![CDATA[
<message>
The operating system cannot run %1.
(0xc3) - exit code 195 (0xc3)</message>
<stderr_txt>
00:14:03 (19992): wrapper (7.9.26016): starting
00:14:03 (19992): wrapper: running python.exe (bin/conda-unpack)
00:14:13 (19992): python.exe exited; CPU time 0.000000
00:14:13 (19992): wrapper: running Library/usr/bin/tar.exe (xjvf input.tar.bz2)
atom.tar
JNK1_m38_m58_0.xml
JNK1_m38_m58_asyncre.cntl
JNK1_m38_m58.inpcrd
JNK1_m38_m58.prmtop
run.bat
run.sh
00:14:14 (19992): Library/usr/bin/tar.exe exited; CPU time 0.015625
00:14:14 (19992): wrapper: running C:/Windows/system32/cmd.exe (/c call run.bat)
ERROR: Invalid requirement: './Acellera-AToM-OpenMM-*'
00:14:18 (19992): C:/Windows/system32/cmd.exe exited; CPU time 0.000000
00:14:18 (19992): app exit status: 0xd
00:14:18 (19992): called boinc_finish(195)
0 bytes in 0 Free Blocks.
310 bytes in 4 Normal Blocks.
1144 bytes in 1 CRT Blocks.
0 bytes in 0 Ignore Blocks.
0 bytes in 0 Client Blocks.
Largest number used: 0 bytes.
Total allocations: 434076 bytes.
Dumping objects ->

The python package is the large 1.9GB package that downloads to every host at first running of the ATMBeta tasks. It is static and sets up the python environment in the project folder.

It only needs to be downloaded once, not for every task.

The name won't change until Steve updates or make changes to it. If he fixes the package for Windows, the name should change. But he could then make the filename static and reference it directly without paths.

Thank you all for the windows debugging info. I am looking into this!

Thank you all for the windows debugging info. I am looking into this!

thank you Steve, I'm looking forward to crunching ATMs with my altogether 6 GPUs on Windows

Thank you all for the windows debugging info. I am looking into this!

Thanks for working on this, Steve!

I just got a WU called "T0_1-STEVE_TEST_ATM-1-5-RND5320" where I noticed you went for a pre-untarred folder "Acellera-AToM-OpenMM-gitrepo" inside the input file.

I'm happy to report that this went past the pip install statement without a hitch and is now happily simulating!

Good job!

And done successfully!
https://www.gpugrid.net/result.php?resultid=33751815

I've got another one in queue that is not yet corrected, so I'm going to suspend GPUGRID for now to avoid a string of error-WU's until you let us know the fix has been incorporated in all new WU's.

Great thanks for the help! The new changes have been passed onto the researchers. Next round of jobs should have the fix.

(Please note I can't inject this fix into already sent WU's)

Just to explain a bit about how this app currently works.

The "app" is a python environment we package as a zipfile (~1GB). This is downloaded once. It will be re-downloaded if we update the app. Updating an app is a rather time consuming process and error prone so we try and avoid it unless absolutely necessary.

In each work unit we include three main things: 1. The input molecular structures. 1. A few scripts that run the simulation. 3. A git code folder that contains the python code (Atom-OpenMM, ~a few MB size folder).

The code folder could have been packaged into the "app" python environment. However, this code is something we update regularly with different features so it is easier to include it on a per work unit basis.

+1