[QE-users] Help Figuring Out Why HSE06 Calculation is Freezing

[Alpin Novianus Tatan]

Dear Derek,

I think you may still have a memory problem.

"I increased the memory to 650 Gigabytes over 32 processors on 2 nodes.
That seemed to fix the problem, at least the calculation wasn't crashing
anymore. But now the calculation simply stalls after the initial SCF
convergence (which takes about 10-30 minutes) when it reaches the "ACE
projected onto   272 (nbndproj) and applied to   272 (nbnd) bands" line."

In your time-limit case,

     Estimated max dynamical RAM per process >      11.69 GB

     Estimated total dynamical RAM >     748.23 GB
     EXX: ACE will be initialized later

The total is more than 650 GB you have, and also you see that at memory per
process this isn't so different than your first crash (11 GB):

     Estimated max dynamical RAM per process >      11.81 GB

     Estimated total dynamical RAM >     378.02 GB
     EXX: ACE will be initialized later

You increased memory but also increased the number of processors. So the
memory per processor might not have increased sufficiently.
You should keep or reduce the number of processors but increase the memory
(and/or use multiple nodes) so that the memory room per process increases.
This may slow your calculation (so give it enough time to finish).

As a simple check if this is a memory issue, try reducing ecutwfc (and
ecutrho) drastically from 60 to 40 Ry, for example.
If your calculation finishes without problems, then you know it is a memory
issue and you can address it as suggested above.

Best regards,
Alpin N. Tatan

Department of Physics,
The University of Tokyo

On Thu, Jul 2, 2026 at 3:49 PM Derek Heidenfelder via users <
users at lists.quantum-espresso.org> wrote:

> Hello,
> I have been trying to perform an HSE06 Hybrid Functional SCF calculation
> on the material AlMgB14 to determine the band gap energy. I have been using
> Quantum Espresso version 7.3 on a High Performance Computing Cluster.
>
> Initially, the calculation was running into Out Of Memory problems, which
> wasn't surprising since the unit cell I am using is very large containing
> 64 atoms and 272 bands. I increased the memory to 650 Gigabytes over 32
> processors on 2 nodes. That seemed to fix the problem, at least the
> calculation wasn't crashing anymore. But now the calculation simply stalls
> after the initial SCF convergence (which takes about 10-30 minutes) when it
> reaches the "ACE projected onto   272 (nbndproj) and applied to   272
> (nbnd) bands" line. I have tried giving the calculation more time to see if
> it was just taking a while but the same thing happened. The calculation has
> never crashed or posted an error log, it has only been terminated by the
> time limit.
>
> I was hoping to get help if anyone has any ideas what the issue might be.
> I'm not the most experienced with Quantum Espresso as I only started
> learning it two months ago. I will attach my input file, output log
> (terminated by the time limit), and an older OOM crash log.
>
> Please let me know if any more information would be helpful and I
> appreciate any input.
> Thank you.
>
>
>
> _______________________________________________________________________________
> The Quantum ESPRESSO Foundation stands in solidarity with all civilians
> worldwide who are victims of terrorism, military aggression, and
> indiscriminate warfare.
>
> --------------------------------------------------------------------------------
> Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
> users mailing list users at lists.quantum-espresso.org
> https://lists.quantum-espresso.org/mailman/listinfo/users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20260702/264f1ed3/attachment.html>