| Issues | Pull Requests | CI | Conda | Documentation | Docker |
|---|---|---|---|---|---|
|
|
|
|
|
This website is intended to provide some guidance on how to get and install the bml library. LA-UR number LA-UR-17-27373.
The basic matrix library (bml) is a collection of various matrix data formats (for dense and sparse) and their associated algorithms for basic matrix operations. Application programming interfaces (API) are available for both C and FORTRAN. The current status of this library allows us to use two different formats for representing matrix data. Currently these formats are: dense, ELLPACK-R, ELLBLOCK, ELLSORT, and CSR. For information on how to use the BML library can be find in BML-API.
We are running the following mailing list for discussions on usage and features of the bml library:
The bml library supports the following matrix formats:
- dense
- ELLPACK-R
- ELLSORT
- ELLBLOCK
- CSR
We offer binary packages of the bml library in RPM format thanks to SUSE's OpenBuild Service and for Ubuntu in DEB format.
We are switching our CI tests from Travis-CI to GitHub Actions because Travis-CI
is limiting the number of builds for open source projects. Our workflow
uses a custom Docker image which
comes with the necessary compiler tool chain to build and test the bml
library. Using docker is a convenient and quick way to develop, build,
and test the bml library.
$ ./scripts/run-local-docker-container.sh
latest: Pulling from nicolasbock/bml
2f94e549220a: Already exists
8d8ab0ffcd5e: Pull complete
3fa4d3b6f5b4: Pull complete
4f4fb700ef54: Pull complete
Digest: sha256:18237f909f19896a57c658c93af5e8ed91c9fa596f15021be777a97444a3eaaf
Status: Downloaded newer image for nicolasbock/bml:latest
docker.io/nicolasbock/bml:latest
groups: cannot find name for group ID 1000
I have no name!@3a4ae718ba4f:/bml$Inside the container:
I have no name!@6ea3f4937c0d:/bml$ ./build.sh compile
Writing output to /bml/build.log
Running command compile
mkdir: created directory '/bml/build'
mkdir: created directory '/bml/install'
-- CMake version 3.12.1
-- The C compiler identification is GNU 7.5.0
-- The CXX compiler identification is GNU 7.5.0
-- The Fortran compiler identification is GNU 7.5.0
-- Check for working C compiler: /usr/bin/gcc
-- Check for working C compiler: /usr/bin/gcc -- worksAlternatively, you can run one of the CI tests by executing e.g.
I have no name!@6ea3f4937c0d:/bml$ ./scripts/ci-gcc-10-C-single-real.sh
+++ dirname ./scripts/ci-gcc-10-C-single-real.sh
++ readlink --canonicalize ./scripts/..
+ basedir=/bml
+ export CC=gcc-10
+ CC=gcc-10
+ export CXX=g++-11
+ CXX=g++-11
+ export FC=gfortran-11
+ FC=gfortran-11The bml library is built with CMake. For convenience, we provide a shell script
which goes through the necessary motions and builds the library, runs the tests,
and installs it (in the install directory).
We suggest to take a look at the example_build.sh script that sets the
most important environmental variables needed by build.sh script. Change
the Variables according to the compilers and architecture. The script can be run
just by doing:
$ ./scripts/example_build.sh
Writing output to /bml/build.log
Running command configure
mkdir: created directory '/bml/build'
mkdir: created directory '/bml/install'
-- CMake version 3.12.1
-- The C compiler identification is GNU 7.5.0
-- The CXX compiler identification is GNU 7.5.0
-- The Fortran compiler identification is GNU 7.5.0By running:
$ ./build.sh installthe library will be built in the build directory and installed in the
install directory. In case you change any sources and simply want to
rebuild the library, you don't have to run build.sh again, but rather
$ make -C buildThe compiled library can be installed by running
$ make -C build installThe install directory can be modified by running
$ CMAKE_INSTALL_PREFIX=/some/path ./build.sh install(which assumes that you are using the bash shell).
To build with GNU compilers, OpenMP, and Intel MKL do the following.
$ CC=gcc FC=gfortran \
BLAS_VENDOR=Intel CMAKE_BUILD_TYPE=Release \
BML_OPENMP=yes CMAKE_INSTALL_PREFIX=/some/path \
./build.sh installTo build with MPI, OpenMP, and use Intel MKL do the following.
$ CC=mpicc FC=mpif90 \
BLAS_VENDOR=Intel CMAKE_BUILD_TYPE=Release \
BML_OPENMP=yes BML_MPI=yes CMAKE_INSTALL_PREFIX=/some/path \
./build.sh installIn order to build the library, the following tools need to be installed:
gccwith Fortran support>=cmake-2.8.8>=python-2.7>=OpenMP-3.1(i.e.>=gcc-4.7)
In case the build fails for some reason, please contact the developers by opening an issue on GitHub (https://github.com/lanl/bml/issues) and attach the files
build/CMakeFiles/CMakeOutput.log
build/CMakeFiles/CMakeError.logOur main development happens on the master branch and is continuously
verified for correctness. If you would like to contribute with your work to the
bml project, please follow the instructions at the GitHub help page "About pull
requests". To
summarize:
- Fork the project on github
- Clone that forked repository
- Create a branch in it
- Commit any changes to the branch
- Push the branch to your forked repository
- Go to https://github.com/lanl/bml and click on 'Create Pull Request'
During the review process you might want to update your pull request. Please add
commits or amend your existing commits as necessary. If you amend any
commits you need to add the --force-with-lease option to the
git push command. Please make sure that your pull request contains only
one logical change (see "Structural split of change"
for further details.
Please indent your C code using
$ indent -gnu -nut -i4 -bli0 -cli4 -ppi0 -cbi0 -npcs -bfdaYou can use the script indent.sh to indent all C code.
For low level optimization work it is useful to understand what assembly code the compiler generates. For example, to verify that the compiler vectorizes the loop in the following example:
void double_array(float a[8]) {
a = __builtin_assume_aligned(a, 64);
for (int i = 0; i < 8; i++) {
a[i] *= 2;
}
}we can build the source with
gcc -S -O3 -fverbose-asm test.cand analyze the generated assembly code,
# test.c:8: a[i] *= 2;
movaps (%rdi), %xmm0 # MEM <vector(4) float> [(float *)a_9], vect__5.8
addps %xmm0, %xmm0 #, vect__5.8
movaps %xmm0, (%rdi) # vect__5.8, MEM <vector(4) float> [(float *)a_9]
movaps 16(%rdi), %xmm0 # MEM <vector(4) float> [(float *)a_9 + 16B], vect__5.8
addps %xmm0, %xmm0 #, vect__5.8
movaps %xmm0, 16(%rdi) # vect__5.8, MEM <vector(4) float> [(float *)a_9 + 16B]The aligned memory access, movaps, moving 4 (aligned packed single-precision) float values into %xmm0, and the subsequent addps instruction show that the compiler fully vectorized the loop.
Note that the Compiler Explorer provides an alternative that does not require local compilations, see https://godbolt.org/z/ejEdqKa6Y.
If you find this library useful, we encourage you to cite us. Our project has a citable DOI:
with the following bibtex snipped:
@misc{bml,
author = {Nicolas Bock and
Susan Mniszewski and
Bálint Aradi and
Michael Wall and
Christian F. A. Negre
Jamal Mohd-Yusof and
Anders N. M. Niklasson},
title = {qmmd/bml v2.1.2},
month = feb,
year = 2022,
doi = {10.5281/zenodo.5570404},
url = {https://doi.org/10.5281/zenodo.5570404}
}Another citation source is the following journal article (DOI: 10.1007/s11227-018-2533-0):
@article{bock2018basic,
title = {The basic matrix library (BML) for quantum chemistry},
author = {Bock, Nicolas and
Negre, Christian FA and
Mniszewski, Susan M and
Mohd-Yusof, Jamaludin and
Aradi, B{\'a}lint and
Fattebert, Jean-Luc and
Osei-Kuffuor, Daniel and
Germann, Timothy C and
Niklasson, Anders MN},
journal = {The Journal of Supercomputing},
volume = {74},
number = {11},
pages = {6201--6219},
year = {2018},
publisher = {Springer}
}The core developers of the bml in alphabetical order:
- Christian Negre <[email protected]>
- Nicolas Bock <[email protected]>
- Susan M. Mniszewski <[email protected]>
- Adedoyin Adetokunbo <[email protected]>
- Bálint Aradi <[email protected]>
- Daniel Osei-Kuffuor <[email protected]>
- Jamaludin Mohd-Yusof <[email protected]>
- Jean-Luc Fattebert <[email protected]>
- Mike Wall <[email protected]>
The bml library is licensed under the BSD 3-clause license.
Copyright 2015. Los Alamos National Security, LLC. This software was produced under U.S. Government contract DE-AC52-06NA25396 for Los Alamos National Laboratory (LANL), which is operated by Los Alamos National Security, LLC for the U.S. Department of Energy. The U.S. Government has rights to use, reproduce, and distribute this software. NEITHER THE GOVERNMENT NOR LOS ALAMOS NATIONAL SECURITY, LLC MAKES ANY WARRANTY, EXPRESS OR IMPLIED, OR ASSUMES ANY LIABILITY FOR THE USE OF THIS SOFTWARE. If software is modified to produce derivative works, such modified software should be clearly marked, so as not to confuse it with the version available from LANL.
Additionally, redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
- Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
- Neither the name of Los Alamos National Security, LLC, Los Alamos National Laboratory, LANL, the U.S. Government, nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission
THIS SOFTWARE IS PROVIDED BY LOS ALAMOS NATIONAL SECURITY, LLC AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL LOS ALAMOS NATIONAL SECURITY, LLC OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
NOTICE OF OSS COPYRIGHT ASSERTION:
LANS has asserted copyright on the software package entitled Basic Matrix Library (bml), Version 0.x (C16006).
The basic matrix library (bml) is a collection of various matrix data formats (for dense and sparse) and their associated algorithms for basic matrix operations.
This code is unclassified and has been assigned LA-CC-15-093. Los Alamos National Laboratory’s Export Control Team made an in-house determination that this software is controlled under Department of Commerce regulations and the Export Control Classification Number (ECCN) EAR99. The export control review is attached.
The developers intend to distribute this software package under the OSI Certified BSD 3-Clause License (http://www.opensource.org/licenses/BSD-3-Clause)
This code was developed using funding from:
- Basic Energy Sciences (LANL2014E8AN) and the Laboratory Directed Research and Development Program of Los Alamos National Laboratory. To tests these developments we used resources provided by the Los Alamos National Laboratory Institutional Computing Program, which is supported by the U.S. Department of Energy National Nuclear Security Administration
- Exascale Computing Project (17-SC-20-SC), a collaborative effort of two U.S. Department of Energy organizations (Office of Science and the National Nuclear Security Administration) responsible for the planning and preparation of a capable exascale ecosystem, including software, applications, hardware, advanced system engineering, and early testbed platforms, in support of the nation’s exascale computing imperative.
Larry Kwei, LAFO Program Manager, has granted his concurrence to asserting copyright and then distributing the Basic Matrix Library (bml), Version 0.x code using an open source software license. See attached memo.
LANS acknowledges that it will comply with the DOE OSS policy as follows:
- submit form DOE F 241.4 to the Energy Science and Technology Software Center (ESTSC),
- provide the unique URL on the form for ESTSC to distribute, and
- maintain an OSS Record available for inspection by DOE.
Following is a table briefly summarizes information for this software package:
| CODE NAME | Basic Matrix Library (bml), Version 0.x (C16006) |
| Classification Review Number | LA-CC-15-093 |
| Export Control Classification Number (ECCN) | EAR99 |
| B&R Code | YN0100000 |