This page provides details on several ways to build Conduit from source.

For the shortest path from zero to Conduit, see Quick Start.

If you are building features that depend on third party libraries we recommend using uberenv which leverages Spack or Spack directly. We also provide info about building for known HPC clusters using uberenv. and a Docker example that leverages Spack.

Obtain the Conduit source

Clone the Conduit repo from Github:

git clone --recursive

--recursive is necessary because we are using a git submodule to pull in BLT ( If you cloned without --recursive, you can checkout this submodule using:

cd conduit
git submodule init
git submodule update

Configure a build

Conduit uses CMake for its build system. These instructions assume cmake is in your path. We recommend CMake 3.21 or newer, for more details see Supported CMake Versions. is a simple wrapper for the cmake call to configure conduit. This creates a new out-of-source build directory build-debug and a directory for the install install-debug. It optionally includes a host-config.cmake file with detailed configuration options.

cd conduit

Build, test, and install Conduit:

cd build-debug
make -j 8
make test
make install

Build Options

The core Conduit library has no dependencies outside of the repo, however Conduit provides optional support for I/O and Communication (MPI) features that require externally built third party libraries.

Conduit’s build system supports the following CMake options:

Main CMake Options





Controls if shared (ON) or static (OFF) libraries are built.

(default = ON)


Controls if Fortran components of Conduit are built.

(default = ON)


Controls if the Conduit Python module and related tests are built.

(default = OFF)


Controls if Conduit MPI features are built.

(default = ON)


Controls if Conduit OpenMP features are built.

(default = OFF)


Controls if unit tests are built.

(default = ON)


Controls if Conduit examples are built.

(default = ON)


Controls if Conduit utilities are built.

(default = ON)


Controls if the Conduit documentation is built (when sphinx is available).

(default = ON)


Controls if code coverage compiler flags are used to build Conduit.

(default = OFF)


Extra control for if Conduit unit tests are built. Useful for in cases where Conduit is pulled into a larger CMake project

(default = ON)

CMake Options for Third-party Library Paths




Path to a HDF5 install (optional). Controls if HDF5 I/O support is built into conduit_relay.


Path to a Zlib install (optional). (Needed for HDF5 support)


Path to a Silo install (optional). Controls if Silo I/O support is built into conduit_relay. Requires HDF5.


Path to a H5ZZFP install (optional). Controls if HDF5 ZFP support is built into conduit_relay.


Path to a ZFP install (optional). Controls if ZFP support is built into conduit_relay.


Path to a Parmetis install (optional). Controls if Parmetis support for conduit_blueprint partitioning features is built.


Path to a Caliiper install (optional). Controls if Caliper performance annotation support.


Path to a Adiak install (optional). (Caliper support depends on Adiak)


Path to a BLT install (default = blt)

Installation Path Options

Conduit’s build system provides an install target that installs the Conduit libraires, headers, python modules, and documentation. These CMake options allow you to control install destination paths:




Standard CMake install path option (optional).


Path to install Python modules into (optional).

Additional Build Notes

  • Python - The Conduit Python module builds for both Python 2 and Python 3. To select a specific Python, set the CMake variable PYTHON_EXECUTABLE to path of the desired python binary. When PYTHON_MODULE_INSTALL_PREFIX is set and ENABLE_PYTHON=ON, Conduit’s Python modules will be installed to ${PYTHON_MODULE_INSTALL_PREFIX} directory instead of ${CMAKE_INSTALL_PREFIX}/python-modules.

  • MPI - We use CMake’s standard FindMPI logic. To select a specific MPI set the CMake variables MPI_C_COMPILER and MPI_CXX_COMPILER, or the other FindMPI options for MPI include paths and MPI libraries. To run the mpi unit tests, you may also need change the CMake variables MPIEXEC_EXECUTABLE and MPIEXEC_NUMPROC_FLAG, so you can use a different launcher, such as srun and set number of MPI tasks used.

  • BLT - Conduit uses BLT ( as the foundation of its CMake-based build system. It is included as a submodule in Conduit’s git repo, and also available in our release tarballs. The BLT_SOURCE_DIR CMake option defaults to src/blt, where we expect the blt submodule. The most compelling reason to override is to share a single instance of BLT across multiple projects.

Host Config Files

To handle build options, third party library paths, etc we rely on CMake’s initial-cache file mechanism.

cmake -C config_file.cmake

We call these initial-cache files host-config files, since we typically create a file for each platform or specific hosts if necessary.

These files use standard CMake commands. To properly seed the cache, CMake set commands need to specify CACHE as follows:


Building Conduit and Third Party Dependencies

We use Spack ( to help build Conduit’s third party dependencies on OSX and Linux. Conduit builds on Windows as well, but there is no automated process to build dependencies necessary to support Conduit’s optional features.

Uberenv (scripts/uberenv/ automates fetching spack, building and installing third party dependencies, and can optionally install Conduit as well. To automate the full install process, Uberenv uses the Conduit Spack package along with extra settings such as Spack compiler and external third party package details for common HPC platforms.

Building Third Party Dependencies for Development


Conduit developers use scripts/uberenv/ to setup third party libraries for Conduit development. For info on how to use the Conduit Spack package see Building Conduit and its Dependencies with Spack.

On OSX and Linux, you can use scripts/uberenv/ to help setup your development environment. This script leverages Spack to build all of the external third party libraries and tools used by Conduit. Fortran support is optional and all dependencies should build without a fortran compiler. After building these libraries and tools, it writes an initial host-config file and adds the Spack built CMake binary to your PATH so can immediately call the helper script to configure a conduit build.

#build third party libs using spack
python3 scripts/uberenv/

# run the configure helper script and give it the
# path to a host-config file
./ uberenv_libs/`hostname`*.cmake

Uberenv Options for Building Third Party Dependencies has a few options that allow you to control how dependencies are built:





Destination directory



Spack spec

linux: %gcc osx: %clang


Spack environment file (spack.yaml)

linux: (empty) osx: (empty)


Ignore SSL Errors



Fully install conduit, not just dependencies



Invoke tests during build and against install


The -k option exists for sites where SSL certificate interception undermines fetching from github and https hosted source tarballs. When enabled, clones spack using:

git -c http.sslVerify=false clone

And passes -k to any spack commands that may fetch via https.

Default invocation on Linux:

python3 scripts/uberenv/ --prefix uberenv_libs \
                                   --spec %gcc

Suggested invocation on macOS:

python3 scripts/uberenv/ --prefix uberenv_libs \
                                   --spec %clang \
                                   --spack-config-dir scripts/uberenv_configs/spack_configs/envs/darwin/spack.yaml

The uberenv –install installs conduit@develop (not just the development dependencies):

python3 scripts/uberenv/ --install

To run tests during the build process to validate the build and install, you can use the --run_tests option:

python3 scripts/uberenv/ --install \

For details on Spack’s spec syntax, see the Spack Specs & dependencies documentation.

You use the –spack-env-file option to specify a Spack environment file. See the Spack Environments and Spack System Packages documentation for details.

For macOS, the enries in scripts/uberenv_configs/spack_configs/envs/darwin/spack.yaml are X-Code’s clang and gfortran from


The bootstrapping process ignores ~/.spack/compilers.yaml to avoid conflicts and surprises from a user’s specific Spack settings on HPC platforms.

When run, checkouts a specific version of Spack from github as {prefix}/spack. It then uses Spack to build and install Conduit’s dependencies to create a Spack environment at {prefix}/spack_env. Finally, it generates a host-config file at {prefix}/{hostname}.cmake that specifies the compiler settings and paths to all of the dependencies.

Building Conduit and its Dependencies with Spack

As of 1/4/2017, Spack’s develop branch includes a recipe to build and install Conduit.

To install the latest released version of Conduit with all options (and also build all of its dependencies as necessary) run:

spack install conduit

To build and install Conduit’s github develop branch run:

spack install conduit@develop

The Conduit Spack package provides several variants that customize the options and dependencies used to build Conduit. To see these variants, please rune

spack info conduit

Variants are enabled using + and disabled using ~. For example, to build Conduit with the minimum set of options (and dependencies) run:

spack install conduit~python~mpi~hdf5~silo~docs

You can specify specific versions of a dependency using ^. For Example, to build Conduit with Python 3,19,19:

spack install conduit+python ^python@3.10.10

Supported CMake Versions

We recommend CMake 3.21 or newer. We test building Conduit with CMake 3.21 and newer 3.2x variants.

Using Conduit in Another Project

Under src/examples there are examples demonstrating how to use Conduit in a CMake-based build system (using-with-cmake) and via a Makefile (using-with-make).

Building Conduit in a Docker Container

Under src/examples/docker/ubuntu there is an example Dockerfile which can be used to create an ubuntu-based docker image with a build of the Conduit. There is also a script that demonstrates how to build a Docker image from the Dockerfile ( and a script that runs this image in a Docker container ( The Conduit repo is cloned into the image’s file system at /conduit, the build directory is /conduit/build-debug, and the install directory is /conduit/install-debug.

Building Conduit with pip

Conduit provides a that allows pip to use CMake to build and install Conduit and the Conduit Python module. This script assumes that CMake is in your path.

Example Basic Build:

pip install . --user

Or for those with certificate woes:

pip install  --trusted-host --trusted-host . --user

You can enable Conduit features using the following environment variables:





Path to HDF5 install for HDF5 Support



Build Conduit with MPI Support


Example Build with MPI and HDF5 Support:

env ENABLE_MPI=ON HDF5_DIR={path/to/hdf5/install} pip install . --user

Notes for Cray systems

HDF5 and gtest use runtime features such as dlopen. Because of this, building static on Cray systems commonly yields the following flavor of compiler warning:

Using 'zzz' in statically linked applications requires at runtime the shared libraries from the glibc version used for linking

You can avoid related linking warnings by adding the -dynamic compiler flag, or by setting the CRAYPE_LINK_TYPE environment variable:

export CRAYPE_LINK_TYPE=dynamic

Shared Memory Maps are read only on Cray systems, so updates to data using Node::mmap will not be seen between processes.

Notes for using OpenMPI in a container as root

By default OpenMPI prevents the root user from launching MPI jobs. If you are running as root in a container you can use the following env vars to turn off this restriction: