1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
|
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
<!--Converted with LaTeX2HTML 2019.2 (Released June 5, 2019) -->
<HTML lang="EN">
<HEAD>
<TITLE>4 Parallelization</TITLE>
<META NAME="description" CONTENT="4 Parallelization">
<META NAME="keywords" CONTENT="developer_man">
<META NAME="resource-type" CONTENT="document">
<META NAME="distribution" CONTENT="global">
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=utf-8">
<META NAME="viewport" CONTENT="width=device-width, initial-scale=1.0">
<META NAME="Generator" CONTENT="LaTeX2HTML v2019.2">
<LINK REL="STYLESHEET" HREF="developer_man.css">
<LINK REL="next" HREF="node8.html">
<LINK REL="previous" HREF="node6.html">
<LINK REL="next" HREF="node8.html">
</HEAD>
<BODY >
<!--Navigation Panel-->
<A
HREF="node8.html">
<IMG WIDTH="37" HEIGHT="24" ALT="next" SRC="next.png"></A>
<A
HREF="developer_man.html">
<IMG WIDTH="26" HEIGHT="24" ALT="up" SRC="up.png"></A>
<A
HREF="node6.html">
<IMG WIDTH="63" HEIGHT="24" ALT="previous" SRC="prev.png"></A>
<A ID="tex2html40"
HREF="node1.html">
<IMG WIDTH="65" HEIGHT="24" ALT="contents" SRC="contents.png"></A>
<BR>
<B> Next:</B> <A
HREF="node8.html">5 Files produced by</A>
<B> Up:</B> <A
HREF="developer_man.html">User's Guide for the</A>
<B> Previous:</B> <A
HREF="node6.html">3 GRID parallelization and</A>
<B> <A ID="tex2html41"
HREF="node1.html">Contents</A></B>
<BR>
<BR>
<!--End of Navigation Panel-->
<H1><A ID="SECTION00050000000000000000">
4 Parallelization</A>
</H1>
<P>
The <TT>PHonon</TT> package uses the same parallelization mechanisms of the Q<SMALL>UANTUM </SMALL>ESPRESSO package. See the Developer manual in the <TT>Doc</TT> directory
two levels above this one for more information.
It is parallelized on plane-waves, pools, bands, and images.
The <TT>-ortho</TT> flag is not used. <TT>Scalapack</TT> routines are not
used in the <TT>ph.x</TT> code.
<P>
Each tensor should be collected as soon as it is calculated
and all processors must have the same tensors. Please avoid to collect
tensors in routines distant from where they are calculated. There might be
exception to this rule for efficiency, but please try not to abuse for
small arrays. Only collected quantities are saved on the <TT>.xml</TT> file,
so they should not depend on the parallelization level. Note that only ionode
writes the <TT>.xml</TT> files, so you have different <TT>xml</TT> files only
for different images. The variables are then broadcasted to all processors
in an image.
<P>
<BR><HR>
</BODY>
</HTML>
|
