Message boards :
Number crunching :
Quantum chemistry calculations on GPU
Message board moderation
Previous · 1 · 2
| Author | Message |
|---|---|
|
Send message Joined: 11 Jul 09 Posts: 1639 Credit: 10,159,968,649 RAC: 428 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
Heads up: problem with new Linux BOINC installation script. This will affect new users only - existing users need not make any adjustments. BOINC makes substantial use of a BOINC data directory, which for many Linux flavours has become established at '/var/lib/boinc-client/'. Instead, the new BOINC installer creates it at '/var/lib/boinc/'. This breaks the current Quantum Chemistry application, which fails with FileNotFoundError: [Errno 2] No such file or directory: '/var/lib/boinc-client/.cupy' (see errors for host 625407) If you have sudo access to your machine (and I assume you do, if you are installing your own software), you should find /var/lib/boinc, and ctreate a symlink folder which points to it, and call that /var/lib/boinc-client |
|
Send message Joined: 21 Feb 20 Posts: 1116 Credit: 40,839,470,595 RAC: 6,423 Level ![]() Scientific publications
|
Heads up: problem with new Linux BOINC installation script. i'm willing to bet that the directory listed is just a relative file path definition, not hard coded. something like assuming you're in the running slot. then going to "../../.cupy" or maybe utilizing an environment variable with $PATH or even BOINC's internal path variables. I don't even install boinc to that directory, nor do i have anything related to boinc in my /var/lib directory. i have it in my home folder. if it were hard coded, no one with a standalone install would be doing work, and no one has reported any issues, so... if you migrated an existing system with apps and stuff already downloaded, you might have some lingering configurations from the old setup? try resetting the project.
|
|
Send message Joined: 11 Jul 09 Posts: 1639 Credit: 10,159,968,649 RAC: 428 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
This is a brand-new machine (less than 3 weeks old) - supplied with no OS installed, so I've installed Linux Mint and BOINC from cold - no history. BOINC is running as a service, but the data file is actually called 'boinc-data' and is located on a different SSD - I've learned how to manage folder redirects! A GPUGrid contributor has joined the parallel conversation at BOINC, and we'll resolve it there. But the work-round works in the meantime, if you need it. |
|
Send message Joined: 21 Feb 20 Posts: 1116 Credit: 40,839,470,595 RAC: 6,423 Level ![]() Scientific publications
|
This is a brand-new machine (less than 3 weeks old) - supplied with no OS installed, so I've installed Linux Mint and BOINC from cold - no history. i think you running it on a separate SSD is likely the issue, or a problem with Linux Mint. and it wont impact most people. it's clearly not hard coded, since I do not have any /var/lib/boinc /var/lib/boinc-client /var/lib/boinc-data or otherwise in my directories at all. in fact, the .cupy directory in use on my system is just put in the users home directory (/home/ian/.cupy) and this works perfectly fine for QChem. you seem to have something else going on with the system to get the environment variables confused between $HOME and /var/lib/boinc-client, cause it's not hard coded to that on GPUGRID's end.
|
|
Send message Joined: 21 Dec 23 Posts: 51 Credit: 0 RAC: 0 Level ![]() Scientific publications ![]() |
Yes that temp dir used by cupy should be located at $HOME/.cupy by default as I mentioned here: https://github.com/BOINC/boinc/discussions/5811#discussioncomment-10670615 And as shown by Ian’s path. I don’t quite know why it would be different for your setup. Maybe a side effect of the new client version you have installed combined with our older wrapper script. Although as mentioned by the BOINC developers it is better if this runtime folder is instead located in the boinc slot or project directory. I will add this change to the next app update. |
|
Send message Joined: 12 Jul 17 Posts: 404 Credit: 17,408,899,587 RAC: 0 Level ![]() Scientific publications ![]() ![]()
|
This is a brand-new machine (less than 3 weeks old) - supplied with no OS installed, so I've installed Linux Mint and BOINC from cold - no history. I've got what may be the same problem with a fresh install of Linux Mint 22 Ubuntu 24.04 Noble. Can't run BOINC on it following these instructions: https://isaac.ssl.berkeley.edu/linux_install.php?os_num=6&build=alpha Please post a link to your "parallel conversation at BOINC" or better yet the solution. TIA |
|
Send message Joined: 21 Feb 20 Posts: 1116 Credit: 40,839,470,595 RAC: 6,423 Level ![]() Scientific publications
|
This is a brand-new machine (less than 3 weeks old) - supplied with no OS installed, so I've installed Linux Mint and BOINC from cold - no history. Sounds like richards problem was in relation to the QChem tasks specifically, not BOINC as a whole. if you're having a problem running BOINC, you have a separate issue.
|
|
Send message Joined: 11 Jul 09 Posts: 1639 Credit: 10,159,968,649 RAC: 428 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
Sounds like richards problem was in relation to the QChem tasks specifically, not BOINC as a whole. if you're having a problem running BOINC, you have a separate issue. Yes, it was specific to Quantum Chemistry tasks. As soon as I saw the results of the first night's run (all very quick errors), I switched to ATMML and they ran flawlessly. Then I investigated the error message about the missing file and the location it was looking in - I knew that didn't match the installation I'd only just completed. So I devised the workround I posted before, and it worked with no other changes. It's an easy fix, so I suggested BOINC cover it - but after discussion, it was decided to fix it at the project end. Unfortunately, the date on the apps page still shows an installation data of 9 Jul 2024, so the problem probably still exists. |
|
Send message Joined: 11 Jul 09 Posts: 1639 Credit: 10,159,968,649 RAC: 428 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
I've got what may be the same problem with a fresh install of Linux Mint 22 Ubuntu 24.04 Noble. Can't run BOINC on it following these instructions: I've had a quick look at your host 624473 - that's the only Mint 22 I can see in the list. You only let two tasks run to the point where they reported a failure. Both were ACEND 3 tasks. One was 'Particle coordinate is nan': the other was 'Cannot use a restart file on a different device!'. Both are well known processing errors here, and not related to the version of BOINC used. Please post a link to your "parallel conversation at BOINC" or better yet the solution. It's at https://github.com/BOINC/boinc/discussions/5811 |
|
Send message Joined: 15 Jul 20 Posts: 95 Credit: 2,550,803,412 RAC: 248 Level ![]() Scientific publications
|
https://www.gpugrid.net/hosts_user.php?userid=563937 pc sous linux mint 22 |
©2025 Universitat Pompeu Fabra