Message boards :
ATLAS application :
all ATLAS tasks fail after about 10 minutes
Message board moderation
Previous · 1 · 2 · 3
Author | Message |
---|---|
Send message Joined: 15 Jun 08 Posts: 2551 Credit: 255,492,121 RAC: 68,719 |
I can confirm that increasing the RAM setting to (nearly) 5000 MB for a 1-core WU the WUs complete successfully. Thanks to Erich56 and HerveUAE. Nonetheless this setting is far above the 3400 MB that can be calculated from David Cameron's RAM formula. This issue may affect all users running 1-core or 2-core WUs and although Erich56 mentioned it more than a week ago it is sad to say that nobody from the project team answered that part of his post. Sure, it's holiday time, but nobody believes that the project runs completely unattended. |
Send message Joined: 18 Dec 15 Posts: 1835 Credit: 119,681,086 RAC: 49,634 |
I can confirm that increasing the RAM setting to (nearly) 5000 MB for a 1-core WU the WUs complete successfully. This surely affects a lot of crunchers - either by making their tasks fail (and if the crunchers have not read here in the forum, they would not find out what the reason is), or crunchers like us here who now know that the RAM requirement has markedly increased, with the result that now not as many ATLAS tasks can be run simultaneously compared to before. For example, for me this means that on one of my machines I cannot run ATLAS any more at all, and on the other (main) machine, I had to change the relation of concurrently running ATLAS and CMS tasks from originally 4:4 to 2:6, which is also too bad. |
©2025 CERN