1# libCEED: the CEED API Library 2 3[](https://travis-ci.org/CEED/libCEED) 4[](https://opensource.org/licenses/BSD-2-Clause) 5[](https://codedocs.xyz/CEED/libCEED/) 6 7## Code for Efficient Extensible Discretization 8 9This repository contains an initial low-level API library for the efficient 10high-order discretization methods developed by the ECP co-design [Center for 11Efficient Exascale Discretizations (CEED)](http://ceed.exascaleproject.org). 12While our focus is on high-order finite elements, the approach is mostly 13algebraic and thus applicable to other discretizations in factored form, see the 14[API documentation](doc/libCEED.md). 15 16One of the challenges with high-order methods is that a global sparse matrix is 17no longer a good representation of a high-order linear operator, both with 18respect to the FLOPs needed for its evaluation, as well as the memory transfer 19needed for a matvec. Thus, high-order methods require a new "format" that still 20represents a linear (or more generally non-linear) operator, but not through a 21sparse matrix. 22 23The goal of libCEED is to propose such a format, as well as supporting 24implementations and data structures, that enable efficient operator evaluation 25on a variety of computational device types (CPUs, GPUs, etc.). This new operator 26description is based on algebraically [factored form](doc/libCEED.md), which is 27easy to incorporate in a wide variety of applications, without significant 28refactoring of their own discretization infrastructure. 29 30The repository is part of the [CEED software suite][ceed-soft], a collection of 31software benchmarks, miniapps, libraries and APIs for efficient exascale 32discretizations based on high-order finite element and spectral element methods. 33See http://github.com/ceed for more information and source code availability. 34 35The CEED research is supported by the [Exascale Computing Project][ecp] 36(17-SC-20-SC), a collaborative effort of two U.S. Department of Energy 37organizations (Office of Science and the National Nuclear Security 38Administration) responsible for the planning and preparation of a [capable 39exascale ecosystem](https://exascaleproject.org/what-is-exascale), including 40software, applications, hardware, advanced system engineering and early testbed 41platforms, in support of the nation’s exascale computing imperative. 42 43For more details on the CEED API see http://ceed.exascaleproject.org/ceed-code/. 44 45## Building 46 47The CEED library, `libceed`, is a C99 library with no external dependencies. 48It can be built using 49 50 make 51 52## Testing 53 54The test suite produces [TAP](https://testanything.org) output and is run by: 55 56 make test 57 58or, using the `prove` tool distributed with Perl (recommended) 59 60 make prove 61 62## Examples 63 64libCEED comes with several examples of its usage, ranging from standalone C 65codes in the `/examples/ceed` directory to examples based on external packages, 66such as MFEM, PETSc and Nek5000. 67 68To build the examples, set the `MFEM_DIR`, `PETSC_DIR` and `NEK5K_DIR` variables 69and run: 70 71```console 72# libCEED examples on CPU and GPU 73cd examples/ceed 74make 75./ex1 -ceed /cpu/self 76./ex1 -ceed /gpu/occa 77cd ../.. 78 79# MFEM+libCEED examples on CPU and GPU 80cd examples/mfem 81make 82./bp1 -ceed /cpu/self -no-vis 83./bp1 -ceed /gpu/occa -no-vis 84cd ../.. 85 86# PETSc+libCEED examples on CPU and GPU 87cd examples/petsc 88make 89./bp1 -ceed /cpu/self 90./bp1 -ceed /gpu/occa 91cd ../.. 92 93# Nek+libCEED examples on CPU and GPU 94cd examples/nek5000 95./make-nek-examples.sh 96./run-nek-example.sh -ceed /cpu/self -b 3 97./run-nek-example.sh -ceed /gpu/occa -b 3 98cd ../.. 99``` 100 101The above code assumes a GPU-capable machine enabled in the OCCA 102backend. Depending on the availabl backends, other Ceed resource specifiers can 103be provided with the `-ceed` option, for example: 104 105CEED resource (`-ceed`) | Backend 106----------------------- | --------------------------------- 107`/cpu/self` | Serial reference implementation 108`/cpu/occa` | Serial OCCA kernels 109`/gpu/occa` | CUDA OCCA kernels 110`/omp/occa` | OpenMP OCCA kernels 111`/ocl/occa` | OpenCL OCCA kernels 112`/gpu/magma` | CUDA MAGMA kernels 113 114## Install 115 116To install libCEED, run 117 118 make install prefix=/usr/local 119 120or (e.g., if creating packages), 121 122 make install prefix=/usr DESTDIR=/packaging/path 123 124Note that along with the library, libCEED installs kernel sources, e.g. OCCA 125kernels are installed in `$prefix/lib/okl`. This allows the OCCA backend to 126build specialized kernels at run-time. In a normal setting, the kernel sources 127will be found automatically (relative to the library file `libceed.so`). 128However, if that fails (e.g. if `libceed.so` is moved), one can copy (cache) the 129kernel sources inside the user OCCA directory, `~/.occa` using 130 131 $(OCCA_DIR)/bin/occa cache ceed $(CEED_DIR)/lib/okl/*.okl 132 133This will allow OCCA to find the sources regardless of the location of the CEED 134library. One may occasionally need to clear the OCCA cache, which can be accomplished 135by removing the `~/.occa` directory or by calling `$(OCCA_DIR)/bin/occa clear -a`. 136 137### pkg-config 138 139In addition to library and header, libCEED provides a [pkg-config][pkg-config1] 140file that can be used to easily compile and link. [For example][pkg-config2], if 141`$prefix` is a standard location or you set the environment variable 142`PKG_CONFIG_PATH`, 143 144 cc `pkg-config --cflags --libs ceed` -o myapp myapp.c 145 146will build `myapp` with libCEED. This can be used with the source or 147installed directories. Most build systems have support for pkg-config. 148 149## Contact 150 151You can reach the libCEED team by emailing [ceed-users@llnl.gov](mailto:ceed-users@llnl.gov) 152or by leaving a comment in the [issue tracker](https://github.com/CEED/libCEED/issues). 153 154## Copyright 155 156The following copyright applies to each file in the CEED software suite, unless 157otherwise stated in the file: 158 159> Copyright (c) 2017, Lawrence Livermore National Security, LLC. Produced at the 160> Lawrence Livermore National Laboratory. LLNL-CODE-734707. All Rights reserved. 161 162See files LICENSE and NOTICE for details. 163 164[ceed-soft]: http://ceed.exascaleproject.org/software/ 165[ecp]: https://exascaleproject.org/exascale-computing-project 166[pkg-config1]: https://en.wikipedia.org/wiki/Pkg-config 167[pkg-config2]: https://people.freedesktop.org/~dbn/pkg-config-guide.html#faq 168
){kind=link}
){kind=link}
){kind=link}