There are two exceptions to this rule. First, variables of style equal and atom ARE redefined each time the command is encountered. This allows them to be reset, when their formulas contain a substitution for another variable, e.g. $x. This can be useful in a -loop. This also means an equal-style variable will re-define a -command-line switch -var setting, so an index-style variable should -be used for such settings instead, as in bench/in.lj. +loop.
Second, as described below, if a variable is iterated on to the end of its list of strings via the next command, it is removed @@ -298,13 +297,13 @@ compute. N is a global atom ID (positive integer).
Fix references access one or more quantities calculated by a -fix. The ID in the reference should be replaced by -the actual ID of the fix defined elsewhere in the input script. -See the doc pages for individual computes to see which ones calculate -global versus per-atom quantities. If the compute reference contains -empty brackets, then per-atom values calculated by the compute are -accessed. Otherwise a single value (global or per-atom) calculated by -the compute is accessed. +fix. The ID in the reference should be replaced by the +actual ID of the fix defined elsewhere in the input script. See the +doc pages for individual computes to see which ones calculate global +versus per-atom quantities. If the fix reference contains empty +brackets, then per-atom values calculated by the fix are accessed. +Otherwise a single value (global or per-atom) calculated by the +fix is accessed.
Note that some fixes only generate quantities on certain timesteps. If a variable attempts to access the fix on non-allowed timesteps, an @@ -338,27 +337,30 @@ print $a
then LAMMPS will run for a while when the print statement is invoked!
-Another way to reference a variable in a formula is using the $x form -instead of v_x. There is a subtle difference between the two -references that has to do with when the evaluation of the included -variable is done. +
It is useful to understand the distinction between referencing a +variable in a formula using the $x form instead of v_x. There is a +subtle difference between the two references that has to do with when +the evaluation of the included variable is done.
-Using a $x, the value of the include variable is substituted for -immediately when the line is read from the input script, just as it -would be in other input script command. This could be the desired -behavior if a static value is desired. Or it could be the desired -behavior for an equal-style variable if the variable command appears -in a loop (see the jump and next commands), -since the substitution will be performed anew each time thru the loop -as the command is re-read. Note that if the variable formula is -enclosed in double quotes, this prevents variable substitution and -thus an error will be generated when the variable formula is -evaluated. +
Referencing the variable as $x, the value of the include variable is +substituted for immediately when the line is read from the input +script, just as it would be in other input script command.
-Using a v_x, the value of the included variable will not be accessed -until the variable formula is evaluated. Thus the value may change -each time the evaluation is performed. This may also be desired -behavior. +
This could be the desired behavior if a static value is desired. Or +it could be the desired behavior for an equal-style variable if the +variable command appears in a loop (see the jump and +next commands), since the substitution will be performed +anew each time thru the loop as the command is re-read. Note that if +the variable formula is enclosed in double quotes, this prevents +variable substitution and thus an error will be generated when the +variable formula is evaluated. +
+Referencing the variable as v_x, the value of the included variable +will not be accessed until the variable formula is evaluated. Thus +the value may change each time the evaluation is performed. This may +also be desired behavior.
As an example, if the current simulation box volume is 1000.0, then these lines: diff --git a/doc/variable.txt b/doc/variable.txt index de38901eeb..2a7cc9fc05 100644 --- a/doc/variable.txt +++ b/doc/variable.txt @@ -35,7 +35,7 @@ style = {index} or {loop} or {world} or {universe} or {uloop} or {equal} or {ato vx\[\], vy\[\], vz\[\], fx\[\], fy\[\], fz\[\] compute references = c_ID, c_ID\[2\], c_ID\[N\], c_ID\[N\]\[2\], c_ID\[\], c_ID\[\]\[2\] fix references = f_ID, f_ID\[2\], f_ID\[N\], f_ID\[N\]\[2\], f_ID\[\], f_ID\[\]\[2\] - other variables = v_abc, v_abc\[N\], v_abc\[\] :pre + variable references = v_abc, v_abc\[N\], v_abc\[\] :pre :ule [Examples:] @@ -91,15 +91,14 @@ script (with 2 exceptions, read further). This is to allow an input script to be processed multiple times without resetting the variables; see the "jump"_jump.html or "include"_include.html commands. It also means that using the "command-line switch"_Section_start.html#2_6 -var -will override a corresponding variable setting in the input script. +will override a corresponding index variable setting in the input +script. There are two exceptions to this rule. First, variables of style {equal} and {atom} ARE redefined each time the command is encountered. This allows them to be reset, when their formulas contain a substitution for another variable, e.g. $x. This can be useful in a -loop. This also means an {equal}-style variable will re-define a -command-line switch -var setting, so an {index}-style variable should -be used for such settings instead, as in bench/in.lj. +loop. Second, as described below, if a variable is iterated on to the end of its list of strings via the "next"_next.html command, it is removed @@ -291,13 +290,13 @@ c_ID\[\]: per-atom scalar from a per-atom compute c_ID\[\]\[M\]: per-atom vector component from a per-atom compute :tb(s=:) Fix references access one or more quantities calculated by a -"fix"_fix.html. The ID in the reference should be replaced by -the actual ID of the fix defined elsewhere in the input script. -See the doc pages for individual computes to see which ones calculate -global versus per-atom quantities. If the compute reference contains -empty brackets, then per-atom values calculated by the compute are -accessed. Otherwise a single value (global or per-atom) calculated by -the compute is accessed. +"fix"_fix.html. The ID in the reference should be replaced by the +actual ID of the fix defined elsewhere in the input script. See the +doc pages for individual computes to see which ones calculate global +versus per-atom quantities. If the fix reference contains empty +brackets, then per-atom values calculated by the fix are accessed. +Otherwise a single value (global or per-atom) calculated by the +fix is accessed. Note that some fixes only generate quantities on certain timesteps. If a variable attempts to access the fix on non-allowed timesteps, an @@ -329,27 +328,30 @@ print $a :pre then LAMMPS will run for a while when the print statement is invoked! -Another way to reference a variable in a formula is using the $x form -instead of v_x. There is a subtle difference between the two -references that has to do with when the evaluation of the included -variable is done. +:line -Using a $x, the value of the include variable is substituted for -immediately when the line is read from the input script, just as it -would be in other input script command. This could be the desired -behavior if a static value is desired. Or it could be the desired -behavior for an equal-style variable if the variable command appears -in a loop (see the "jump"_jump.html and "next"_next.html commands), -since the substitution will be performed anew each time thru the loop -as the command is re-read. Note that if the variable formula is -enclosed in double quotes, this prevents variable substitution and -thus an error will be generated when the variable formula is -evaluated. +It is useful to understand the distinction between referencing a +variable in a formula using the $x form instead of v_x. There is a +subtle difference between the two references that has to do with when +the evaluation of the included variable is done. -Using a v_x, the value of the included variable will not be accessed -until the variable formula is evaluated. Thus the value may change -each time the evaluation is performed. This may also be desired -behavior. +Referencing the variable as $x, the value of the include variable is +substituted for immediately when the line is read from the input +script, just as it would be in other input script command. + +This could be the desired behavior if a static value is desired. Or +it could be the desired behavior for an equal-style variable if the +variable command appears in a loop (see the "jump"_jump.html and +"next"_next.html commands), since the substitution will be performed +anew each time thru the loop as the command is re-read. Note that if +the variable formula is enclosed in double quotes, this prevents +variable substitution and thus an error will be generated when the +variable formula is evaluated. + +Referencing the variable as v_x, the value of the included variable +will not be accessed until the variable formula is evaluated. Thus +the value may change each time the evaluation is performed. This may +also be desired behavior. As an example, if the current simulation box volume is 1000.0, then these lines: