From be0256375c56eb2d683ca19d1c3b26668f6e871e Mon Sep 17 00:00:00 2001
From: Giovanni Bussi <giovanni.bussi@gmail.com>
Date: Fri, 7 Dec 2018 12:59:57 +0100
Subject: [PATCH] Improved doc
---
user-doc/Miscelaneous.md | 8 ++++----
user-doc/Performances.md | 9 ++++++---
2 files changed, 10 insertions(+), 7 deletions(-)
diff --git a/user-doc/Miscelaneous.md b/user-doc/Miscelaneous.md
index 3742fca0a..e72943d21 100644
--- a/user-doc/Miscelaneous.md
+++ b/user-doc/Miscelaneous.md
@@ -144,14 +144,14 @@ autocompletion should be installed. Otherwise, configure will look for the prese
and, in case it is installed on the same prefix as PLUMED, also PLUMED autocompletion will be installed.
Finally, if none of these two conditions are satisfied, autocompletion will not be enabled. You will
have to change your bashrc file once adding the following lines:
-Now look at these lines:
\verbatim
_plumed() { eval "$(plumed --no-mpi completion 2>/dev/null)";}
complete -F _plumed -o default plumed
\endverbatim
-This is what you are expected to do if for instance you have multiple versions of PLUMED installed
-concurrently using separate env modules.
-The command `plumed completion` just writes on its standard output the body of a bash function.
+The command `plumed completion` just writes on its standard output the body of a bash function that is
+then used by bash to construct the autocompletion.
+The `--no-mpi` flag makes it more likely that the command can be executed correctly e.g. when you are on the login node of a cluster and
+PLUMED was compiled with MPI but the login node does not support MPI. In other cases, it is harmless.
The `-o default` options will make sure that if `plumed --no-mpi completion` returns an error the default bash completion
will be used. This is what will happen if you load an older PLUMED version for which the `completion` command is not available yet.
In future PLUMED versions the `plumed completion` command might return more sophisticated functions. You should
diff --git a/user-doc/Performances.md b/user-doc/Performances.md
index c45a7350b..87afac91f 100644
--- a/user-doc/Performances.md
+++ b/user-doc/Performances.md
@@ -230,8 +230,8 @@ At last, try to reduce the number of residues in the calculation.
In case you are using a lot of \ref CUSTOM functions or \ref switchingfunction "switching functions",
notice that these functionalities depend on the lepton library included in PLUMED.
-This library replace libmatheval since PLUMED 2.5, and by itself it is significantly faster than libmatheval.
-However, you can make it even faster using a [just-in-time compilater](https://github.com/asmjit/asmjit.git).
+This library replaces libmatheval since PLUMED 2.5, and by itself it is significantly faster than libmatheval.
+However, you can make it even faster using a [just-in-time compiler](https://github.com/asmjit/asmjit.git).
Currently, this is an experimental feature, so use it with care.
In order to enable it you should first install asmjit.
@@ -285,7 +285,10 @@ PLUMED: 4A 2 dfast 108 0.135210 0.
...
\endverbatim
-Notice the usage of `x2` as a variable for the switching function (see \ref switchingfunction).
+Notice the usage of `x2` as a variable for the switching function (see \ref switchingfunction), which
+avoids an unnecessary square root calculation (this is done automatically by the hard-coded switching functions
+when you use only even powers). The asmjit calculation (`dfast`) takes less than 10% more than the hard-coded
+one (`c`).
\page Time Time your Input
--
GitLab