<html><body>
<div style="font-family: Tahoma; color: rgb(0, 0, 0); font-size: 16px; font-weight: 400; font-style: normal; background: rgba(0, 0, 0, 0) none repeat scroll 0% 0%;">Dear Paolo,</div>
<div style="font-family: Tahoma; color: rgb(0, 0, 0); font-size: 16px; font-weight: 400; font-style: normal; background: rgba(0, 0, 0, 0) none repeat scroll 0% 0%;">Hi.</div>
<div style="font-family: Tahoma; color: rgb(0, 0, 0); font-size: 16px; font-weight: 400; font-style: normal; background: rgba(0, 0, 0, 0) none repeat scroll 0% 0%;">Thank you for your comments.</div>
<div style="font-family: Tahoma; color: rgb(0, 0, 0); font-size: 16px; font-weight: 400; font-style: normal; background: rgba(0, 0, 0, 0) none repeat scroll 0% 0%;">Your comments mean that the "small numerical differences" may cause to system converge to different metastable states in DFT+U, as is my case.</div>
<div style="font-family: Tahoma; color: rgb(0, 0, 0); font-size: 16px; font-weight: 400; font-style: normal; background: rgba(0, 0, 0, 0) none repeat scroll 0% 0%;">To solve this problem, I compiled the QE-7.1 with the following config.:</div>
<div style="font-family: Tahoma; color: rgb(0, 0, 0); font-size: 16px; font-weight: 400; font-style: normal; background: rgba(0, 0, 0, 0) none repeat scroll 0% 0%;">"./configure FC=ifort F77=ifort MPIF90=mpif90 CC=icc CXX=icpc F90FLAGS=-O0 FCFLAGS=-O0 FFLAGS=-O0 CFLAGS=-O0"</div>
<div style="font-family: Tahoma; color: rgb(0, 0, 0); font-size: 16px; font-weight: 400; font-style: normal; background: rgba(0, 0, 0, 0) none repeat scroll 0% 0%;"> </div>
<div style="font-family: Tahoma; color: rgb(0, 0, 0); font-size: 16px; font-weight: 400; font-style: normal; background: rgba(0, 0, 0, 0) none repeat scroll 0% 0%;">and the results for "-np 2" and "-np 8" becomes very much similar (slightly different), and so maybe acceptable.</div>
<div style="font-family: Tahoma; color: rgb(0, 0, 0); font-size: 16px; font-weight: 400; font-style: normal; background: rgba(0, 0, 0, 0) none repeat scroll 0% 0%;">However, I also tried compiling using the following command to increase precision:</div>
<div style="font-family: Tahoma; color: rgb(0, 0, 0); font-size: 16px; font-weight: 400; font-style: normal; background: rgba(0, 0, 0, 0) none repeat scroll 0% 0%;">./configure FC=ifort F77=ifort MPIF90=mpif90 CC=icc CXX=icpc F90FLAGS="-O0 -double-size 128" FCFLAGS="-O0 -double-size 128" FFLAGS="-O0 -double-size 128" CFLAGS="-O0"</div>
<div style="font-family: Tahoma; color: rgb(0, 0, 0); font-size: 16px; font-weight: 400; font-style: normal; background: rgba(0, 0, 0, 0) none repeat scroll 0% 0%;"> </div>
<div style="font-family: Tahoma; color: rgb(0, 0, 0); font-size: 16px; font-weight: 400; font-style: normal; background: rgba(0, 0, 0, 0) none repeat scroll 0% 0%;">and during "make pw" I get the following error message:</div>
<div style="font-family: Tahoma; color: rgb(0, 0, 0); font-size: 16px; font-weight: 400; font-style: normal; background: rgba(0, 0, 0, 0) none repeat scroll 0% 0%;"> </div>
<div style="font-family: Tahoma; color: rgb(0, 0, 0); font-size: 16px; font-weight: 400; font-style: normal; background: rgba(0, 0, 0, 0) none repeat scroll 0% 0%;">
<div>mpif90 -O0 -double-size 128 -fpp -allow nofpp_comments -D__DFTI -D__MPI -D__SCALAPACK -I/home/mahmoud/QE71O0-double_128/external/devxlib/src -I. -I/home/mahmoud/QE71O0-double_128/include -I/home/mahmoud/QE71O0-double_128/FoX/finclude -I/opt/intel/2017.8/compilers_and_libraries_2017.8.262/linux/mkl/include -c cdiaghg.f90<br />
cdiaghg.f90(540): error #6285: There is no matching specific subroutine for this generic subroutine call. [SQR_SETMAT]<br />
CALL sqr_setmat( 'U', n, ( 0.D0, 0.D0 ), ss, size(ss,1), idesc )<br />
----------^<br />
cdiaghg.f90(559): error #6285: There is no matching specific subroutine for this generic subroutine call. [SQR_MM_CANNON]<br />
CALL sqr_mm_cannon( 'N', 'N', n, ( 1.D0, 0.D0 ), ss, nx, hh, nx, ( 0.D0, 0.D0 ), v, nx, idesc )<br />
----------^<br />
cdiaghg.f90(567): error #6285: There is no matching specific subroutine for this generic subroutine call. [SQR_MM_CANNON]<br />
CALL sqr_mm_cannon( 'N', 'C', n, ( 1.D0, 0.D0 ), v, nx, ss, nx, ( 0.D0, 0.D0 ), hh, nx, idesc )<br />
----------^<br />
cdiaghg.f90(572): error #6285: There is no matching specific subroutine for this generic subroutine call. [SQR_SETMAT]<br />
CALL sqr_setmat( 'H', n, ( 0.D0, 0.D0 ), hh, size(hh,1), idesc )<br />
----------^<br />
cdiaghg.f90(607): error #6285: There is no matching specific subroutine for this generic subroutine call. [SQR_MM_CANNON]<br />
CALL sqr_mm_cannon( 'C', 'N', n, ( 1.D0, 0.D0 ), ss, nx, hh, nx, ( 0.D0, 0.D0 ), v, nx, idesc )<br />
----------^<br />
cdiaghg.f90(431): warning #6843: A dummy argument with an explicit INTENT(OUT) declaration is not given an explicit value. [V]<br />
SUBROUTINE laxlib_pcdiaghg( n, h, s, ldh, e, v, idesc )<br />
---------------------------------------------^<br />
compilation aborted for cdiaghg.f90 (code 1)<br />
make[1]: *** [cdiaghg.o] Error 1<br />
make[1]: Leaving directory `/home/mahmoud/QE71O0-double_128/LAXlib'</div>
<div> </div>
<div>So, I could not succeed here.</div>
<div> </div>
<div>Thirdly, from the QE development site at gillab I downloaded the code "q-e-mixed_precision" with the hope that maybe it is somehow a cure. It was QE-6.5 and developed by Carlo Cavazzoni. The compilation was as usual (without using "-O0" flags), and did not find any improvement over the default configuration.</div>
<div> </div>
<div>Could you please give me an advice on how can I compile PWscf with any desired higher precisions than the default?</div>
<div> </div>
<div>Thank you so much.</div>
<div>Best regards,</div>
<div>Mahmoud</div>
</div>
<div style="font-family: Tahoma; color: rgb(0, 0, 0); font-size: 16px; font-weight: 400; font-style: normal; background: rgba(0, 0, 0, 0) none repeat scroll 0% 0%;">
<blockquote class="as91pFk8AX8sDPOUp66RUy8IoCRjefxNG8oLk8fg" style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
<div style="font-family: monospace, courier new, courier">In systems with a difficult self-consistency, it is possible that the<br />
small numerical differences, coming from execution on different numbers<br />
of processors, are sufficient to drive the system towards<br />
non-convergence, or convergence towards a different final state (DFT+U<br />
seems to be especially unstable in this respect).<br />
<br />
If you are using some exotic, little tested feature, it is conceivable<br />
that some overlooked bug in parallelization exists. In this case, the<br />
problem is easily reproducible and clearly bound to a specific feature,<br />
though.<br />
<br />
I do not see any other possibility<br />
<br />
Paolo<br />
<br />
On 28/11/2022 17:05, Mahmoud Payami Shabestari via users wrote:<br />
> Hi All,<br />
> I am experiencing that the success in scf cycle depends on the number of<br />
> processes used in mpirun. That is, for example, for a given scf input<br />
> when I use "mpirun -np *2* pw.x < input.in",<br />
> it gives the converged result in a reasonable number of iteration; but,<br />
> on the other hand, if I instead use "mpirun -np *8* pw.x < input.in",<br />
> the system does not converge numerically.<br />
> First I thought it is somehow a problem of numerical error accumulation.<br />
> So, I used in compilation of QE the FFTW3 as prescribed in the manual.<br />
> And even used "-lfftw3l" (long-double) to cure this problem, but some<br />
> other numerical problems happened in vc-relax job.<br />
> I would appreciate if anybody helps to understand the roots of this<br />
> dependency and how to control it.<br />
> Bests,<br />
> Mahmoud Payami<br />
> NSTRI, AEOI, Tehran, Iran<br />
> Email: <a href="mailto:mpayami%40aeoi.org.ir">mpayami@aeoi.org.ir</a><br />
> Phone: +98 (0)21 82066504<br />
> --------------------------------------------------------<br />
><br />
> _______________________________________________<br />
> The Quantum ESPRESSO community stands by the Ukrainian<br />
> people and expresses its concerns about the devastating<br />
> effects that the Russian military offensive has on their<br />
> country and on the free and peaceful scientific, cultural,<br />
> and economic cooperation amongst peoples<br />
> _______________________________________________<br />
> Quantum ESPRESSO is supported by MaX (<a href="http://www.max-centre.eu" target="_blank">www.max-centre.eu</a>)<br />
> users mailing list <a href="mailto:users%40lists.quantum-espresso.org">users@lists.quantum-espresso.org</a><br />
> <a href="https://lists.quantum-espresso.org/mailman/listinfo/users" target="_blank">https://lists.quantum-espresso.org/mailman/listinfo/users</a><br />
<br />
--<br />
Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,<br />
Univ. Udine, via delle Scienze 206, 33100 Udine Italy, +39-0432-558216</div>
</blockquote>
</div>
</body></html>