[Pw_forum] need pwscf large-scale input test file HELP

Hi, You can find some much larger benchmarks on the QE Forge site (http://qe-forge.org/gf/) under Quantum ESPRESSO -> Files ->QE Benchmarks I?m not sure if the link changes but here is the current link: http://qe-forge.org/gf/project/q-e/frs/?action=FrsReleaseBrowse&frs_package_id=36 Mike micha

[Pw_forum] QE 5.0.2 compile problem and MKL problem

Hi, Yue-Wen, I think you made a simple mistake. In defining your libraries, it should be -lmkl_blacs_lp64.a and NOT -libmkl_blacs_lp64.a. Mike michael at ihpc.a-star.edu.sg http://www.sullivan.sg/ On 9 Dec, 2013, at 4:58 pm, Yue-Wen Fang mailto:yuewen.fang

[Pw_forum] CuO bulk Error during compilation

Hi, Please be polite and sign your message with your name and affiliation. I see one problem that you have ntyp=2 when it should be 3 since you have 3 things in ATOMIC_SPECIES. Mike Sullivan Institute of High Performance Computing, Singapore michael at ihpc.a-star.edu.sg http://www.sullivan.sg/

[Pw_forum] How to convince QE from local lapack/blas?

Thanks, Paolo, for the clarification. Mike michael at ihpc.a-star.edu.sg http://www.sullivan.sg/ On 27-Jul-2011, at 3:31 PM, Paolo Giannozzi wrote: > > On Jul 27, 2011, at 5:40 , Michael Sullivan wrote: > >> If I'm not wrong, that just means that you're not using ESS

[Pw_forum] How to convince QE from local lapack/blas?

Guntram: If I'm not wrong, that just means that you're not using ESSL. Mike michael at ihpc.a-star.edu.sg http://www.sullivan.sg/ On 27-Jul-2011, at 3:31 AM, Guntram Schmidt wrote: > Heureka! > > After several days of compiling and complaining I've found a possible > solution! > In contrast to

[Pw_forum] How to convince QE from local lapack/blas?

Guntram: I've never seen this error message, but if you can try to edit the make.sys file directly. Replace the LAPACK_LIBS flag to point to your own LAPACK and probably change LAPACK_LIBS_SWITCH to external. That said, though, I've found that usually if it can't find it, there's a reason. Goo

[Pw_forum] Single point energy calculation within PWSCF.

Hongsheng: It's just calculation='scf' in &CONTROL http://www.quantum-espresso.org/input-syntax/INPUT_PW.html#id2933659 Mike michael at ihpc.a-star.edu.sg http://www.sullivan.sg/ On 26-May-2011, at 12:49 AM, Hongsheng Zhao wrote: > Hi all, > > I've learned that the single point energy calculati

[Pw_forum] limit on atom count to compute total energy

Sakthi: I think you will need a VERY GOOD supercomputer to do this. 300 atoms is quite a challenge so I can't image trying 3000 atoms. If you really need to study a system with this size, it seems like a forcefield molecular dynamics approach would be the best. Mike Sullivan Institute of High

[Pw_forum] van der Waals density functional

Dear all, It's in the CVS thanks especially to Stefano de Gironcoli and Timo Thonhauser. From what I've gathered, you're suppose to use RPBE pseudopotentials and override the functional using this: input_dft='vdw-df' in &SYSTEM. You'll also need a generate a vdw-df kernel using generate_vdW_ker

[Pw_forum] changing vdW radii for dispersion correction

Martin: Have a look at the mm_dispersion.f90 file in the Modules folder. It's pretty obvious what to change. This is also the place you'd change the C6 if needed. Mike michael at ihpc.a-star.edu.sg http://www.sullivan.sg/ -Original Message- From: pw_forum-bounces at pwscf.org [mailto:pw

[Pw_forum] QM/MM and/or external charges in QE

PWSCFers: My colleague has this question/problem: We have a cluster of atoms (not necessarily periodical). One (small) region of this cluster must be treated by Quantum Mechanics, the rest can be tre

[Pw_forum] Van der Waals interaction in PWSCF

Okuno-san: Barone et al. has implemented Grimme's DFT-D by adding london=.true. into the &SYSTEM tag. In my opinion, this has some limitations that Grimme has recently addressed but hasn't been implemented in Espresso yet. He calls it DFT-D3. If you want to use Dion's method, have a look at JuN

[Pw_forum] constraint vc-relax

Hi, That would be cell_dofree in &CELL: http://www.quantum-espresso.org/input-syntax/INPUT_PW.html#id3902439 Mike michael at ihpc.a-star.edu.sg http://www.sullivan.sg/ From: pw_forum-bounces at pwscf.org [mailto:pw_forum-boun...@pwscf.org] On Behalf Of Moha

[Pw_forum] Predicting the cpu time usage for bigger system withour doing the simulation

Hi, I chose N=5 because the original post referred to 5 supercells. If your system is twice as big it will be 2**3, or 8 times bigger. For the second part of your message, I don't have a good way to know in advance how it's going to perform. Usually I run some tests and build up experience. If

[Pw_forum] Predicting the cpu time usage for bigger system withour doing the simulation

Dimpy: DFT usually scales as N^3 so I'd guess that it will take 5**3=125 times longer than the small system. You would probably gain somewhere due to fewer k-Points, etc. but that's my rough estimate. Others may have other comments that I welcome... Mike michael at ihpc.a-star.edu.sg

[Pw_forum] Problem in reading PP's taken from QE site?

Sonu Kumar: I'm not having any trouble with Intel 11. I'd double check my pseudo_dir and make sure the file is where you think it is (without any typos anywhere). I usually have a typo or point to the wrong directory. Good luck! Mike michael at ihpc.a-star.edu.sg

[Pw_forum] How can I search the mail list archive?

Haowei: I usually just use Google and start the search with pwscf and whatever I want to search on. If you want, you can also start your search with pw_forum site:democritos.it For example you could put this into Google: site:democritos.it diagonalization failed -OR- pwscf diagonalization faile

[Pw_forum] Relax calculation in quantum espresso with 56 atoms had not been converged

You may have set nstep to 40. If you are using BFGS, it could be that it actually took 50 steps (the default, I think) but it only used the last 39 steps. Have a look at the energies and see how it looks by using a grep command like this: grep "!total energy " If you add -c to grep then it

[Pw_forum] Relax calculation in quantum espresso with 56 atoms had not been converged

My two guesses are: (1) Ran out of time in the queue. (2) Only allowed 40 optimization steps. That said, we need more information. Usually something will show up in the error file that will give us a hint. Mike michael at ihpc.a-star.edu.sg http://www.sullivan.sg/ From: pw_forum-bounces at pws

[Pw_forum] TDDFT and GW

I was looking into this a bit this week? If I understand it correctly, both YAMBO and SaX can perform GW calculations. Is there any reason to use one over the other and in what situations? By the way, information about SaX can be found here: http://www.sax-project.org/ Mike Sullivan Institute o

[Pw_forum] Geometry Optimization

Prof Paulatto: Very good point. Mahasin: Listen to the master. :) For slabs, I usually optimize the bulk cell and use that lattice constant to generate a surface. With this surface, I don't change the cell (ie I use calculation='relax'). Hope that helps and sorry for the confusion. Mike mic

[Pw_forum] Geometry Optimization

Mahasin: You are very close. Trying using calculation='vc-relax'. Don't forget to add an &CELL section to your input. Here is the relevant section of the manual: http://www.quantum-espresso.org/input-syntax/INPUT_PW.html#id3667252 Mike michael at ihpc.a-star.edu.sg http://www.sullivan.sg/ From