<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=Windows-1252">
<style type="text/css" style="display:none;"> P {margin-top:0;margin-bottom:0;} </style>
</head>
<body dir="ltr">
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Dear Stefano, </div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Thanks a lot for the kind and fast reply! This is a good question, I also wondered that maybe the storage part was useless. I need this term because it enters the PAW equation for the "all-electron reconstruction" of the pseudopotential KS-states: </div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
\psi_{ae} = \psi_{ps} + \sum_{R,l} [\phi_{R,l}^{ae} - \phi_{R,l}^{ps}] x <\beta_{R,l}|\psi_{ps}></div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
where R and l are the atomic species and projectors indices, \phi the "atomic-like" states from the PP calculation and \psi the KS state (there might be slight mistake in the formula).</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
If I now want to apply this correction to each of the u_{k+G} pw coefficients that contributes to \psi_{ps}, I think I also need the <\beta_{R,l}|\psi_{ps}> in a "G-wise" fashion, as I was naming it in the question. </div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
I was therefore implementing this routine inside "gipaw", which already has precious stuff written for this PAW reconstruction. <br>
<br>
I hope this gives you more information, and please let me know if you think there is already a good place either in pw or gipaw where to perform this! </div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Best,</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Luca </div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
ETH Zurich, PhD student</div>
<div id="appendonsend"></div>
<hr style="display:inline-block;width:98%" tabindex="-1">
<div id="divRplyFwdMsg" dir="ltr"><font face="Calibri, sans-serif" style="font-size:11pt" color="#000000"><b>Da:</b> developers <developers-bounces@lists.quantum-espresso.org> per conto di Stefano de Gironcoli <degironc@sissa.it><br>
<b>Inviato:</b> sabato 8 febbraio 2025 15:55<br>
<b>A:</b> developers@lists.quantum-espresso.org <developers@lists.quantum-espresso.org><br>
<b>Oggetto:</b> Re: [QE-developers] Parallelization question for G-wise projector-wavefunctions product computation</font>
<div> </div>
</div>
<div>
<p>Dear Luca,</p>
<p> the array betapsi_g(1:npw,1:nkb,1:nbnd) looks very big to me, unless the system is very small</p>
<p> what do you need it for ? maybe one can compute the target quantity on the fly rather than storing this intermediate result</p>
<p><br>
</p>
<p>stefano</p>
<p><br>
</p>
<div class="x_moz-cite-prefix">On 08/02/25 15:31, Marin Luca wrote:<br>
</div>
<blockquote type="cite"><style type="text/css" style="display:none">
<!--
p
{margin-top:0;
margin-bottom:0}
-->
</style>
<div class="x_elementToProof" style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
Dear QE developers, <br>
</div>
<div class="x_elementToProof" style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
I had a conceptual question of how I should correctly modify a subroutine in order to avoid parallelization and memory’s absurdities. Specifically, my idea is to slightly modify one of the subroutines in the “calbec” (becmod.f90) interface that is computing
<\Beta|\psi> products, and I am using QE v.6.4 because I’m working on a specific code repository using this version. By calling such or similar line:<br>
</div>
<div style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
CALL ZGEMV( 'C', npw, nkb, (1.0_DP,0.0_DP), beta, npwx, psi, 1, &</div>
<div class="x_elementToProof" style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
(0.0_DP, 0.0_DP), betapsi, 1 )<br>
</div>
<div class="x_elementToProof" style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
and subsequently summing through bands:<br>
</div>
<div class="x_elementToProof" style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
CALL mp_sum( betapsi( :, 1:m ), intra_bgrp_comm )<br>
</div>
<div class="x_elementToProof" style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
“calbec_k” computes the betapsi product as a 2D matrix with dimension (number of projectors, number of bands). My goal is to compute the same product but accessing the elements at each G vectors, i.e. without summing over the G vectors themselves. To this end,
my (very basic, probably terrible) idea was to define another subroutine with an extra variable like this (without breaking the requirement of having unambiguous interface procedures in “calbec”):<br>
</div>
<div style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
calbec_k_modified ( npw, beta, psi, betapsi, betapsi_g, nbnd )</div>
<div class="x_elementToProof" style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
COMPLEX (DP), INTENT (out) :: betapsi_g(:,:,:) <br>
</div>
<div class="x_elementToProof" style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
and explicitly computing betapsi_g with nested do loops: <br>
</div>
<div style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
DO k = 1, npw</div>
<div style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
DO i = 1, nkb</div>
<div style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
DO j = 1, m</div>
<div style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
betapsi_g(k, i, j) = CONJG(beta(k, i)) * psi(k, j)</div>
<div style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
ENDDO</div>
<div style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
ENDDO</div>
<div class="x_elementToProof" style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
ENDDO<br>
</div>
<div class="x_elementToProof" style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
I am very unsure of whether this is a feasible approach and how I should handle the MPI parallelization in this case (I have little experience in QE development). I read the parallelization section of the developer’s manual, but I’m still quite confused. Any
suggestions, critics or explanations will be very appreciated. </div>
<div class="x_elementToProof" style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
Best regards,</div>
<div class="x_elementToProof" style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
Luca Marin</div>
<div class="x_elementToProof" style="line-height:100%; margin-top:1em; margin-bottom:0in; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
ETH Zurich, PhD student </div>
<div class="x_elementToProof" style="font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
<br>
</div>
<div style="font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
<br>
</div>
<br>
<fieldset class="x_moz-mime-attachment-header"></fieldset>
<pre class="x_moz-quote-pre">________________________________________________
The Quantum ESPRESSO Foundation stands in solidarity with all civilians worldwide who are victims of terrorism, military aggression, and indiscriminate warfare.
_______________________________________________
developers mailing list
<a class="x_moz-txt-link-abbreviated" href="mailto:developers@lists.quantum-espresso.org">developers@lists.quantum-espresso.org</a>
<a class="x_moz-txt-link-freetext" href="https://lists.quantum-espresso.org/mailman/listinfo/developers">https://lists.quantum-espresso.org/mailman/listinfo/developers</a>
</pre>
</blockquote>
</div>
</body>
</html>