[QE-users] [SPAM] Re: Problem with restart_mode='restart'

[Lorenzo Paulatto]

Hello,

what you are getting seems pretty normal to me, but I think it can be 
improved. Your calculation converges very badly, partially because 
spin-orbit is difficult but also because the starting magnetization is 
off by a factor three. Now that you have an idea of how strong the 
magnetization is, it will be more efficient to use its value as a 
starting point.

The code is able to complete a single relax step before it is killed by 
the scheduler, and because it does not stop cleanly (setting variable 
max_seconds) the wavefunctions cannot be used to restart it. Hence when 
you restart, it has to redo all the convergence from scratch.

Also, you used a very lax threshold for the scf calculation, which may 
have led to noisy forces, but energy is indeed going down, so this may 
not be a problem. You also use an extremely large cutoff. If you cannot 
afford more CPU time, it is better to start with a lower cutoff and a 
tighter convergence threshold. Actually, it is better to start without 
spin-orbit coupling and do an initial relax. Then do a calculation with 
spin-orbit and check how strong are the forces, and eventually relax a 
bit more.

Finally, you are using 40 CPUs over three nodes, which seems very odd to 
me as 40 is not divisible by three.


On 18/07/2024 09:28, Suraj P wrote:
> Hello,
> Please find the attached script file (Ni111.sh).
>
> I have already attached the input and output files. The links of the same are as follows:
>
> input (Ni(111)):
> https://drive.google.com/file/d/1PPb6PvDpJECKrtNUSL8RzI73V7XL1zSW/view?usp=drive_link
>
> output1:
> https://drive.google.com/file/d/1Hb5mqXD5Wr3jpE81Aww7v0h_Uqa96bX4/view?usp=drive_link
>
> input2:-----> used for restarting
> https://drive.google.com/file/d/1VeHWLRoAoP7iVAvEAtCCQZrEit2N4eZZ/view?usp=drive_link
>
> output2afterrestarting:
> https://drive.google.com/file/d/12suf8CbOun1JpA_vewfQJgApD05QotpC/view?usp=drive_link
>
> The file size is more than 200kB so I have to upload it through link.
>
> Thanking you
> Suraj
>
> _______________________________________________
> The Quantum ESPRESSO community stands by the Ukrainian
> people and expresses its concerns about the devastating
> effects that the Russian military offensive has on their
> country and on the free and peaceful scientific, cultural,
> and economic cooperation amongst peoples
> _______________________________________________
> Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
> users mailing listusers at lists.quantum-espresso.org
> https://lists.quantum-espresso.org/mailman/listinfo/users
-- 
Dr. Lorenzo Paulatto
IdR @ IMPMC - CNRS UMR 7590 & Sorbonne Université
phone: +33 (0)1 442 79822 / skype: paulatz
http://www.impmc.upmc.fr/~paulatto/ - https://anharmonic.github.io/
23-24/423 B115, 4 place Jussieu 75252 Paris CX 05
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20240718/8730ce4b/attachment.html>