Installation

With mamba or conda

Note that Jumble was built in conda-forge with gfortran version 14, so you cannot use another compiler than gfortran for your consumer program, and it must be version \(\ge\) 8 of gfortran. Note also that Jumble was built in single-precision, with optimization options. If you want another compiler or other compilation options, you need to compile from source.

See instructions in Jumble feedstock.

Compilation from source

Dependencies

CMake (version ≥ 3.16)¹.
A Fortran compiler.

Instructions

Download Jumble.
Unzip it:
```
unzip Jumble-master.zip
```
Decide in which directory you want to install Jumble after compilation and type the command below with your choice of directory after -DCMAKE_INSTALL_PREFIX= (enter an absolute path). For example:
```
cmake -B build -S Jumble-master -DCMAKE_INSTALL_PREFIX=~/.local
```
The command above is the configuration command. It will create a directory named build.
Compile:
```
cmake --build build
```
Install:
```
cmake --install build
```

You do not need to keep the downloaded directory Jumble-master nor the build directory after installation.

Advanced instructions

Most users should not need these advanded instructions.

You can choose any name and any location for the build directory at the configuration step. You have to refer to the build directory at the compilation and installation steps:
```
cmake -B /wherever/any/name -S Jumble-master -DCMAKE_INSTALL_PREFIX=~/.local
cmake --build /wherever/any/name
cmake --install /wherever/any/name
```
Optionally, in the configuration command, you can change the precision for real numbers of procedures zroots_unity, outersum, cumprod, outerdiv and vabs. By default, the precision for real numbers in Jumble is set to be the default kind of the real type of the Fortran compiler you will use. You can change the value for the procedures listed above to any real kind you want. For example:
```
cmake -B build -S Jumble-master -DJumble_CPP_WP='kind(0d0)'
```
or you could choose selected_real_kind(10) for example, etc.

On Mac OS, after downloading the application from the CMake web site, run it, then click on "How to Install For Command Line Use" in the Tools menu. ↩