Compiling (Speeding up): Difference between revisions

From FreeCAD Documentation
No edit summary
mNo edit summary
(3 intermediate revisions by one other user not shown)
Line 1: Line 1:
<languages/>
<languages/>
<translate>
<translate>

<!--T:8-->
<!--T:8-->
{{Docnav
{{Docnav
|[[Compile_on_Docker|Compile on Docker]]
|[[FreeCAD_Docker_CLI_mode|FreeCAD Docker CLI mode]]
|[[Third Party Libraries|Third-party libraries]]
|[[Third_Party_Libraries|Third Party Libraries]]
}}
}}

</translate>
</translate>
{{TOCright}}
{{TOCright}}
<translate>
<translate>

==Overview== <!--T:1-->
==Overview== <!--T:15-->

<!--T:1-->
FreeCAD is a large application that may take from 10 minutes to one hour to compile completely from source. This depends primarily on the CPU that you have, and the number of cores that are used in the compilation process. Here are some tips to shorten that process and make build times shorter.
FreeCAD is a large application that may take from 10 minutes to one hour to compile completely from source. This depends primarily on the CPU that you have, and the number of cores that are used in the compilation process. Here are some tips to shorten that process and make build times shorter.


Line 62: Line 67:
== distcc == <!--T:7-->
== distcc == <!--T:7-->
The {{incode|distcc}} program can be used to perform distributed compilation of C and C++ code across several machines in a network.
The {{incode|distcc}} program can be used to perform distributed compilation of C and C++ code across several machines in a network.



<!--T:5-->
<!--T:5-->
{{Docnav
{{Docnav
|[[Compile_on_Docker|Compile on Docker]]
|[[FreeCAD_Docker_CLI_mode|FreeCAD Docker CLI mode]]
|[[Third Party Libraries|Third-party libraries]]
|[[Third_Party_Libraries|Third Party Libraries]]
}}
}}



Revision as of 21:24, 29 December 2020

Overview

FreeCAD is a large application that may take from 10 minutes to one hour to compile completely from source. This depends primarily on the CPU that you have, and the number of cores that are used in the compilation process. Here are some tips to shorten that process and make build times shorter.

CCache

Install ccache to cache builds.

Disable modules

When using cmake to configure the build, you can disable the compilation of certain workbenches that you may not need at the moment. This is useful if you only need to test a few workbenches.

For example, to avoid building the FEM and Mesh workbenches:

cmake -DBUILD_FEM=OFF -DBUILD_MESH=OFF ../freecad-source

Use cmake-gui, cmake-curses-gui, or cmake-qt-gui to display all the possible variables that can be edited in the configuration; using these interfaces you can easily switch on or off different workbenches.

Number of jobs in parallel

After configuring with cmake, the make program launches the actual C++ compiler to work on the source code files. You can speed up compilation by working on various files at the same time. This is achieved with the -j option of make, which denotes the number of "jobs" or compilation commands that are run simultaneously. This option is an integer number.

Run four compilation commands in parallel:

make -j4

Compile as many files in parallel as the number of CPU cores in your system. This is useful if you have many cores and want to use them all to compile the software.

make -j$(nproc)

Compile as many files in parallel as the number of CPU cores in your system, minus two. Use this so that your system is still responsive to do some other task; for example, two cores will allow you to use a browser, while the rest of the cores keep compiling the software on the background.

make -j$(nproc --ignore=2)

distcc

The distcc program can be used to perform distributed compilation of C and C++ code across several machines in a network.