mirrors/make
1
0
Fork 0
mirror of https://git.savannah.gnu.org/git/make.git synced 2024-11-24 04:09:41 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

* doc/make.texi: Clarify -j versus -jN behavior

Browse source
This commit is contained in:
Paul Smith 2024-09-02 16:23:36 -04:00
parent 9251546bac
commit 101bf5636f
1 changed files with 22 additions and 9 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

31
doc/make.texi
View file

@ -4380,11 +4380,17 @@ Disable, ,Disabling Parallel Execution}).
On MS-DOS, the @samp{-j} option has no effect, since that system doesn't On MS-DOS, the @samp{-j} option has no effect, since that system doesn't
support multi-processing. support multi-processing.
If the @samp{-j} option is followed by an integer, this is the number of Providing @samp{-j} with no argument tells @code{make} that it should run as
recipes to execute at once; this is called the number of @dfn{job slots}. many jobs in parallel as are allowed by the dependency graph. This is usually
If there is nothing looking like an integer after the @samp{-j} option, not what you want as it can overload your system very quickly. For example
there is no limit on the number of job slots. The default number of job suppose you have a large program that is built from 900 object files. Since
slots is one, which means serial execution (one thing at a time). object files normally do not depend on each other, using @samp{-j} by itself
will attempt to create all 900 object files in parallel, starting 900
instances of the compiler in simultaneously. However, @samp{-j} with no
argument is useful in conjunction with @samp{-l} (see below).
If the @samp{-j} option is followed by an integer, this is the maximum number
of recipes to execute at once; this is called the number of @dfn{job slots}.
Handling recursive @code{make} invocations raises issues for parallel Handling recursive @code{make} invocations raises issues for parallel
execution. For more information on this, see @ref{Options/Recursion, execution. For more information on this, see @ref{Options/Recursion,
@ -9746,10 +9752,10 @@ You can examine the current list of directories to be searched via the
@itemx --jobs[=@var{jobs}] @itemx --jobs[=@var{jobs}]
@cindex @code{--jobs} @cindex @code{--jobs}
Specifies the number of recipes (jobs) to run simultaneously. With no Specifies the number of recipes (jobs) to run simultaneously. With no
argument, @code{make} runs as many recipes simultaneously as possible. argument, @code{make} runs as many recipes simultaneously as are allowed by
If there is more than one @samp{-j} option, the last one is effective. the dependency graph. If there is more than one @samp{-j} option, the last
@xref{Parallel, ,Parallel Execution}, for more information on how one is effective. @xref{Parallel, ,Parallel Execution}, for more information
recipes are run. Note that this option is ignored on MS-DOS. on how recipes are run. Note that this option is ignored on MS-DOS.
@item --jobserver-style=[@var{style}] @item --jobserver-style=[@var{style}]
@cindex @code{--jobserver-style} @cindex @code{--jobserver-style}
@ -12645,6 +12651,13 @@ Be aware that the @code{MAKEFLAGS} variable may contain multiple instances of
the @code{--jobserver-auth=} option. Only the @emph{last} instance is the @code{--jobserver-auth=} option. Only the @emph{last} instance is
relevant. relevant.
Remember that GNU Make does not create a jobserver at all if the user requests
``infinite parallelism'' by using @samp{-j} with no argument limiting the
number of job slots. In this situation there's no maximum number of job
slots, thus no need for a jobserver, and there will be no
@samp{--jobserver-auth} option in @code{MAKEFLAGS}: every instance of
@code{make} can invoke as many jobs in parallel as it likes.
Second, every command @code{make} starts has one implicit job slot Second, every command @code{make} starts has one implicit job slot
reserved for it before it starts. Any tool which wants to participate reserved for it before it starts. Any tool which wants to participate
in the jobserver protocol should assume it can always run one job in the jobserver protocol should assume it can always run one job