scnc
/
quantum-espresso
mirror of https://gitlab.com/QEF/q-e.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Added some info on how to reduce memory usage 

git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@9809 c92efa57-630b-4861-b058-cf58834340f0
This commit is contained in:
giannozz 2013-01-19 14:33:51 +00:00
parent 12ec486aa7
commit 25a4992f24
2 changed files with 16 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
PW/Doc/INPUT_PW.def
View File

@ -1399,7 +1399,8 @@ input_description -distribution {Quantum Espresso} -package PWscf -program pw.x
var mixing_ndim -type INTEGER {
default { 8 }
info {
number of iterations used in mixing scheme
number of iterations used in mixing scheme.
If you are tight with memory, you may reduce it to 4 or so.
}
}

23
PW/Doc/user_guide.tex
View File

@ -1107,21 +1107,26 @@ have problems with the provided tests and examples.
or whenever more RAM is needed}
Possible solutions:
\begin{itemize}
\item increase the amount of RAM you are authorized to use (which may
\item Increase the amount of RAM you are authorized to use (which may
be much smaller than the available RAM). Ask your system
administrator if you don't know what to do. In some cases the
stack size can be a source of problems: if so, increase it with command
\texttt{limits} or \texttt{ulimit}).
\item reduce \texttt{nbnd} to the strict minimum, or reduce the cutoffs, or the
cell size , or a combination of them
\item use conjugate-gradient (\texttt{diagonalization='cg'}: slow but very
robust): it requires less memory than the default Davidson
algorithm. If you stick to the latter, use \texttt{diago\_david\_ndim=2}.
\item in parallel execution, use more processors, or use the same
\item Reduce \texttt{nbnd} to the strict minimum (for insulators, the
default is already the minimum, though).
\item Reduce the work space for Davidson diagonalization to the minimum
by setting \texttt{diago\_david\_ndim=2}; also consider using conjugate
gradient diagonalization (\texttt{diagonalization='cg'}), slow but very
robust, which requires almost no work space.
\item If the charge density takes a significant amount of RAM, reduce
\texttt{mixing\_ndim} from its default value (8) to 4 or so.
\item In parallel execution, use more processors, or use the same
number of processors with less pools. Remember that parallelization
with respect to k-points (pools) does not distribute memory:
parallelization with respect to R- (and G-) space does.
\item buggy or weird-behaving compiler.
only parallelization with respect to R- (and G-) space does.
\item If none of the above is sufficient or feasible, you have to either
reduce the cutoffs and/or the cell size, or to use a machine with
more RAM.
\end{itemize}
\paragraph{pw.x crashes with {\em error in davcio}}