What does this mean: <message>
The operating system cannot run %1.
(0xc3) - exit code 195 (0xc3)</message>

Is this something like Linux trying to run on Windows or what?

What does this mean: <message>
The operating system cannot run %1.
(0xc3) - exit code 195 (0xc3)</message>

Is this something like Linux trying to run on Windows or what?

Googling shows this:

"The operation system cannot run %1" is shown when some apps try to open links

Sounds like the app is trying to open links that are invalid or badly formed.

Or a variation of this:

ImportError: DLL load failed: The operating system cannot run %1.

What does this mean: <message>
The operating system cannot run %1.
(0xc3) - exit code 195 (0xc3)</message>

Is this something like Linux trying to run on Windows or what?

Googling shows this:

"The operation system cannot run %1" is shown when some apps try to open links

Sounds like the app is trying to open links that are invalid or badly formed.

Or a variation of this:

ImportError: DLL load failed: The operating system cannot run %1.

Lovely

Got four failures for ATMbeta resends today, on two slightly differing versions of Linux Mint. All had

+ python -m pip install ./Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6 ./Acellera-AToM-OpenMM-gitrepo
ERROR: Cannot install acellera-atom-openmm 3.3.0rc4 (from /hdd/boinc-client/slots/4/Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6) and acellera-atom-openmm 3.3.0rc4 (from /hdd/boinc-client/slots/4/Acellera-AToM-OpenMM-gitrepo) because these package versions have conflicting dependencies.
ERROR: ResolutionImpossible: for help visit https://pip.pypa.io/en/latest/topics/dependency-resolution/#dealing-with-dependency-conflicts

Got four failures for ATMbeta resends today, on two slightly differing versions of Linux Mint. All had

+ python -m pip install ./Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6 ./Acellera-AToM-OpenMM-gitrepo
ERROR: Cannot install acellera-atom-openmm 3.3.0rc4 (from /hdd/boinc-client/slots/4/Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6) and acellera-atom-openmm 3.3.0rc4 (from /hdd/boinc-client/slots/4/Acellera-AToM-OpenMM-gitrepo) because these package versions have conflicting dependencies.
ERROR: ResolutionImpossible: for help visit https://pip.pypa.io/en/latest/topics/dependency-resolution/#dealing-with-dependency-conflicts

same. i had about 150 of these errors.

---

same. i had about 150 of these errors.

All of mine have gone to full workunit failure, with too many errors. Some of them have been tried by Windows computers, and have failed there too.

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)

Hi Steve,

I notice a new batch of ATMs released, but the fix has only been partially incorporated, leading to failures both on Windows and Linux.

The WU's do contain the pre-untarred 'Acellera-AToM-OpenMM-gitrepo' directory which is good, but the 'atom.tar' file is also still included which is both inefficient use of network bandwith as well as leading to errors.

Worse is, the run.bat and run.sh files both still have the following statement:

python.exe -m pip install ./Acellera-AToM-OpenMM-*

On Windows, this still leads to the same old error because of the invalid path separator / instead of \ and because the '*' is interpreted literally instead of as a wildcard. See Task error example here: https://www.gpugrid.net/result.php?resultid=33760402

On Linux, the same task leads to an error because it does interpret '*' as a wildcard, finding 2 AToM folders (from atom.tar and from Acellera-AToM-OpenMM-gitrepo) instead of 1, with a conflicting dependency error as result. See Task error example here: https://www.gpugrid.net/result.php?resultid=33760795

Resolution:
Replace the statement

python.exe -m pip install ./Acellera-AToM-OpenMM-*

by the following in run.bat (Windows):

python.exe -m pip install .\Acellera-AToM-OpenMM-gitrepo

by the following in run.sh (Linux):

python.exe -m pip install ./Acellera-AToM-OpenMM-gitrepo

As an optimization:
EITHER remove atom.tar from the WU, as well as the corresponding 'tar' statements in run.bat and run.sh
OR package Acellera-AToM-OpenMM-gitrepo folder inside of atom.tar instead of Acellera-AToM-OpenMM-2dd310b8027c68262906a8946f807896b49947b6

same. i had about 150 of these errors.

All of mine have gone to full workunit failure, with too many errors. Some of them have been tried by Windows computers, and have failed there too.

See the reason for both Linux and Windows errors in my post above...

thank you for the catch again, I will check the correct scripts are being used

Task TYK2_m10_m15_2_TEST-QUICO_ATM_500K_dih14fit-0-5-RND5394_0 (ATMbeta) received and is running correctly under Linux Mint 21.3

Hi!

I've got one on Win10 that so far still running!

Task https://www.gpugrid.net/result.php?resultid=33764570

Hi!

I've got one on Win10 that so far still running!

Task https://www.gpugrid.net/result.php?resultid=33764570

here, too :-)

Also the progress bar in the BOINC manager now works fine !

Has the ATM experiment been stopped? Or can we expect more work in future?

A new batch of ATMbeta tasks is on the field since yesterday.
This batch seems to be heavier in processing times than most of precedent ones.

A pinhead:
Answering my own request for multi GPU hosts at Message #61293.
Watching carefully Stderr output report for a certain ATMbeta task, can be found a line like this:

.
+ echo localhost,0:N,1,CUDA,,/var/lib/boinc-client/slots/X/tmp
.

Where "N" corresponds to the Device Number (GPU) where the task was run on.

With ATMbeta tasks I'm not currently experiencing reliability problems, but identifying every single GPU can be useful, for example, to characterize its performance.

A new batch of ATMbeta tasks is on the field since yesterday.
This batch seems to be heavier in processing times than most of precedent ones.

Does anyone else have problems loading ATMbeta even though they were explicitly selected in the GPUGrid settings?

No, none at all.

As long as you also toggle to use "test tasks" in your project preferences.

No, none at all.

As long as you also toggle to use "test tasks" in your project preferences.

this.

you need beta/test tasks selected.

---

you need beta/test tasks selected.

Silly me. I actually missed that. Thank you.

Got an error when pause and restart in my non-English system.
https://www.gpugrid.net/result.php?resultid=34111833

&#215;&#211;&#196;&#191;&#194;&#188;&#187;&#242;&#206;&#196;&#188;&#254; F:\apps\BOINC\data\slots\1\tmp &#210;&#209;&#190;&#173;&#180;&#230;&#212;&#218;&#161;&#163;

This may mean "A subdirectory or file F:\apps\BOINC\data\slots\1\tmp already exists."or so.

Warning: importing 'simtk.openmm' is deprecated. Import 'openmm' instead.

ImportError: DLL load failed while importing _openmm: &#213;&#210;&#178;&#187;&#181;&#189;&#214;&#184;&#182;&#168;&#181;&#196;&#196;&#163;&#191;&#233;&#161;&#163;

I can't read this. Is it because of my system lenguage? Or just something wrong with "importing _openmm" or so?

Just Windows thing. Doesn't have the character library or something.

You generally can't stop or pause GPUGrid tasks.

Depending on the app, it either errors out or starts calculation over from the beginning.

At least for the tasks that have been available lately.