Windows work unit NWChem long XXL

Message boards : Number crunching : Windows work unit NWChem long XXL
Message board moderation

To post messages, you must log in.

AuthorMessage
PHILIPPE

Send message
Joined: 4 Jan 20
Posts: 60
Credit: 516,736
RAC: 0
Message 639 - Posted: 2 Mar 2020, 18:22:13 UTC

Some of us thought that crunching a long work unit with windows and virtualbox was impossible because of the weakness in the stability observed through our previous experiences.
But it 's not true if you follow some rules.
I succeeded to run a very long work unit with a no-dedicated slow computer:
2193598 1327142 26 Feb 2020, 14:52:03 UTC 2 Mar 2020, 16:52:11 UTC Terminé et validé 197,068.84 187,284.10 5,000.00 NWChem long v0.11 (vbox64_t1)
windows_x86_64

This wu began February 27 and ended March 2.It had been suspended and resumed 4 times.(computer no dedicated)

How-to :

To stop successfully:

1°) Go in Boinc manager and Set No new Task for all the projects running on your computer.
2°) Open Windows task manager to see the cpu activity and the read/write access disk.
3°) Suspend all the tasks slowly, beginning by the more resilient wus and finishing by the weakest one(s) (that is to say NWChem long).
4°) Give enough time to do the save state.This is important.
5°) Shutdown your computer.It's not necessary to close Boinc manager before...

To restart successfully :

1°) Start your computer.
2°) Open Windows task manager to see cpu activity and read/write access disk.
3°) Open Boinc manager.
4°) Wait the necessary time to let the windows initialization services make the cpu and the disk at a low activity level.( Telemetry , directX diagnostic tools , anti-virus, windows update , system , and so on)(For me 3 mins)
5°) When you think , good conditions are ready to go , resume the wus , beginning by the weakest ( that is to say NWChem Long).
6°) Wait the time necessary that the cpu activity is maximum for this wu. So you have all the chances that the resume is successfully.
7°) Resume the next wus in the same way , cautiously.
8°) Set Enable new tasks for the projects.
9°) Go and take a coffee.It 's all.You deserve it.

The main idea of this method is to manage the NWChem long wu in the best stable environment (lowest cpu activity and read/write access disk). So you reduce the risk the wu becomes unstable during this particular phase of "stop and start".
In some particular cases , you maybe have to let an idle core to your system to avoid unfortunate perturbation for the wus running under heavy load.

--A small step for QuChemPedIA , a big step for the Windows Cruncher Community.--

Good Luck
ID: 639 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
damotbe
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Help desk expert

Send message
Joined: 23 Jul 19
Posts: 289
Credit: 464,119,561
RAC: 0
Message 644 - Posted: 3 Mar 2020, 10:29:37 UTC - in response to Message 639.  

Smart tips !

Thank you.
ID: 644 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote

Message boards : Number crunching : Windows work unit NWChem long XXL

©2024 Benoit DA MOTA - LERIA, University of Angers, France