Awesome
QUDA 1.1.0
Overview
QUDA is a library for performing calculations in lattice QCD on graphics processing units (GPUs), leveraging NVIDIA's CUDA platform. The current release includes optimized Dirac operators and solvers for the following fermion actions:
- Wilson
- Clover-improved Wilson
- Twisted mass (including non-degenerate pairs)
- Twisted mass with a clover term
- Staggered fermions
- Improved staggered (asqtad or HISQ)
- Domain wall (4-d or 5-d preconditioned)
- Möbius fermion
Implementations of CG, multi-shift CG, BiCGStab, BiCGStab(l), and DD-preconditioned GCR are provided, including robust mixed-precision variants supporting combinations of double, single, half and quarter precisions (where the latter two are 16-bit and 8-bit "block floating point", respectively). The library also includes auxiliary routines necessary for Hybrid Monte Carlo, such as HISQ link fattening, force terms and clover- field construction. Use of many GPUs in parallel is supported throughout, with communication handled by QMP or MPI.
QUDA includes an implementations of adaptive multigrid for the Wilson, clover-improved, twisted-mass and twisted-clover fermion actions. We note however that this is undergoing continued evolution and improvement and we highly recommend using adaptive multigrid use the latest develop branch. More details can be found [here] (https://github.com/lattice/quda/wiki/Multigrid-Solver).
Support for eigen-vector deflation solvers is also included through the Thick Restarted Lanczos Method (TRLM), and we offer an Implicitly Restarted Arnoldi for observing non-hermitian operator spectra. For more details we refer the user to the wiki: [QUDA's eigensolvers] (https://github.com/lattice/quda/wiki/QUDA%27s-eigensolvers) [Deflating coarse grid solves in Multigrid] (https://github.com/lattice/quda/wiki/Multigrid-Solver#multigrid-inverter--lanczos)
Software Compatibility:
The library has been tested under Linux (CentOS 7 and Ubuntu 18.04) using releases 10.1 through 11.4 of the CUDA toolkit. Earlier versions of the CUDA toolkit will not work, and we highly recommend the use of 11.x. QUDA has been tested in conjunction with x86-64, IBM POWER8/POWER9 and ARM CPUs. Both GCC and Clang host compilers are supported, with the mininum recommended versions being 7.x and 6, respectively. CMake 3.15 or greater to required to build QUDA.
See also Known Issues below.
Hardware Compatibility:
For a list of supported devices, see
http://developer.nvidia.com/cuda-gpus
Before building the library, you should determine the "compute
capability" of your card, either from NVIDIA's documentation or by
running the deviceQuery example in the CUDA SDK, and pass the
appropriate value to the QUDA_GPU_ARCH
variable in cmake.
QUDA 1.1.0, supports devices of compute capability 3.0 or greater. QUDA is no longer supported on the older Tesla (1.x) and Fermi (2.x) architectures.
See also "Known Issues" below.
Installation:
It is recommended to build QUDA in a separate directory from the source directory. For instructions on how to build QUDA using cmake see this page https://github.com/lattice/quda/wiki/Building-QUDA-with-cmake. Note that this requires cmake version 3.15 or later. You can obtain cmake from https://cmake.org/download/. On Linux the binary tar.gz archives unpack into a cmake directory and usually run fine from that directory.
The basic steps for building with cmake are:
- Create a build dir, outside of the quda source directory.
- In your build-dir run
cmake <path-to-quda-src>
- It is recommended to set options by calling
ccmake
in your build dir. Alternatively you can use the-DVARIABLE=value
syntax in the previous step. - run 'make -j <N>' to build with N parallel jobs.
- Now is a good time to get a coffee.
You are most likely to want to specify the GPU architecture of the machine you are building for. Either configure QUDA_GPU_ARCH in step 3 or specify e.g. -DQUDA_GPU_ARCH=sm_60 for a Pascal GPU in step 2.
Multi-GPU support
QUDA supports using multiple GPUs through MPI and QMP, together with
the optional use of NVSHMEM GPU-initiated communication for improved
strong scaling of the Dirac operators. To enable multi-GPU support
either set QUDA_MPI
or QUDA_QMP
to ON when configuring QUDA
through cmake.
Note that in any case cmake will automatically try to detect your MPI
installation. If you need to specify a particular MPI please set
MPI_C_COMPILER
and MPI_CXX_COMPILER
in cmake. See also
https://cmake.org/cmake/help/v3.9/module/FindMPI.html for more help.
For QMP please set QUDA_QMP_HOME
to the installation directory of QMP.
For more details see https://github.com/lattice/quda/wiki/Multi-GPU-Support
To enable NVSHMEM support set QUDA_NVSHMEM
to ON, and set the
location of the local NVSHMEM installation with QUDA_NVSHMEM_HOME
.
For more details see
https://github.com/lattice/quda/wiki/Multi-GPU-with-NVSHMEM
External dependencies
The eigen-vector solvers (eigCG and incremental eigCG) by default will
use Eigen, however, QUDA can be configured to use MAGMA if available
(see https://github.com/lattice/quda/wiki/Deflated-Solvers for more
details). MAGMA is available from
http://icl.cs.utk.edu/magma/index.html. MAGMA is enabled using the
cmake option QUDA_MAGMA=ON
.
Version 1.1.0 of QUDA includes interface for the external (P)ARPACK
library for eigenvector computing. (P)ARPACK is available, e.g., from
https://github.com/opencollab/arpack-ng. (P)ARPACK is enabled using
CMake option QUDA_ARPACK=ON
. Note that with a multi-GPU option, the
build system will automatically use PARPACK library.
Automatic download and installation of Eigen, (P)ARPACK, QMP and QIO is supported in QUDA through the CMake options QUDA_DOWNLOAD_EIGEN, QUDA_DOWNLOAD_ARPACK, and QUDA_DOWNLOAD_USQCD.
Application Interfaces
By default only the QDP and MILC interfaces are enabled. For
interfacing support with QDPJIT, BQCD, CPS or TIFR; this should be
enabled at by setting the corresponding QUDA_INTERFACE_<application>
variable e.g., QUDA_INTERFACE_BQCD=ON
. To keep compilation time to
a minimum it is recommended to only enable those interfaces that are
used by a given application.
Tuning
Throughout the library, auto-tuning is used to select optimal launch
parameters for most performance-critical kernels. This tuning process
takes some time and will generally slow things down the first time a
given kernel is called during a run. To avoid this one-time overhead in
subsequent runs (using the same action, solver, lattice volume, etc.),
the optimal parameters are cached to disk. For this to work, the
QUDA_RESOURCE_PATH
environment variable must be set, pointing to a
writable directory. Note that since the tuned parameters are hardware-
specific, this "resource directory" should not be shared between jobs
running on different systems (e.g., two clusters with different GPUs
installed). Attempting to use parameters tuned for one card on a
different card may lead to unexpected errors.
This autotuning information can also be used to build up a first-order
kernel profile: since the autotuner measures how long a kernel takes
to run, if we simply keep track of the number of kernel calls, from
the product of these two quantities we have a time profile of a given
job run. If QUDA_RESOURCE_PATH
is set, then this profiling
information is output to the file "profile.tsv" in this specified
directory. Optionally, the output filename can be specified using the
QUDA_PROFILE_OUTPUT
environment variable, to avoid overwriting
previously generated profile outputs. In addition to the kernel
profile, a policy profile, e.g., collections of kernels and/or other
algorithms that are auto-tuned, is also output to the file
"profile_async.tsv". The policy profile for example includes
the entire multi-GPU dslash, whose style and order of communication is
autotuned. Hence while the dslash kernel entries appearing the kernel
profile do include communication time, the entries in the policy
profile include all constituent parts (halo packing, interior update,
communication and exterior update).
Using the Library:
Include the header file include/quda.h in your application, link against lib/libquda.so, and study tests/invert_test.cpp (for Wilson, clover, twisted-mass, or domain wall fermions) or tests/staggered_invert_test.cpp (for asqtad/HISQ fermions) for examples of the solver interface. The various solver options are enumerated in include/enum_quda.h.
Known Issues:
- When the auto-tuner is active in a multi-GPU run it may cause issues with binary reproducibility of this run if domain-decomposition preconditioning is used. This is caused by the possibility of different launch configurations being used on different GPUs in the tuning run simultaneously. If binary reproducibility is strictly required make sure that a run with active tuning has completed. This will ensure that the same launch configurations for a given kernel is used on all GPUs and binary reproducibility.
Getting Help:
Please visit http://lattice.github.io/quda for contact information. Bug reports are especially welcome.
Acknowledging QUDA:
If you find this software useful in your work, please cite:
M. A. Clark, R. Babich, K. Barros, R. Brower, and C. Rebbi, "Solving Lattice QCD systems of equations using mixed precision solvers on GPUs," Comput. Phys. Commun. 181, 1517 (2010) [arXiv:0911.3191 [hep-lat]].
When taking advantage of multi-GPU support, please also cite:
R. Babich, M. A. Clark, B. Joo, G. Shi, R. C. Brower, and S. Gottlieb, "Scaling lattice QCD beyond 100 GPUs," International Conference for High Performance Computing, Networking, Storage and Analysis (SC), 2011 [arXiv:1109.2935 [hep-lat]].
When taking advantage of adaptive multigrid, please also cite:
M. A. Clark, B. Joo, A. Strelchenko, M. Cheng, A. Gambhir, and R. Brower, "Accelerating Lattice QCD Multigrid on GPUs Using Fine-Grained Parallelization," International Conference for High Performance Computing, Networking, Storage and Analysis (SC), 2016 [arXiv:1612.07873 [hep-lat]].
When taking advantage of block CG, please also cite:
M. A. Clark, A. Strelchenko, A. Vaquero, M. Wagner, and E. Weinberg, "Pushing Memory Bandwidth Limitations Through Efficient Implementations of Block-Krylov Space Solvers on GPUs," Comput. Phys. Commun. 233 (2018), 29-40 [arXiv:1710.09745 [hep-lat]].
When taking advantage of the Möbius MSPCG solver, please also cite:
Jiqun Tu, M. A. Clark, Chulwoo Jung, Robert Mawhinney, "Solving DWF Dirac Equation Using Multi-splitting Preconditioned Conjugate Gradient with Tensor Cores on NVIDIA GPUs," published in the Platform of Advanced Scientific Computing (PASC21) [arXiv:2104.05615[hep-lat]].
Authors:
- Ronald Babich (NVIDIA)
- Simone Bacchio (Cyprus)
- Michael Baldhauf (Regensburg)
- Kipton Barros (Los Alamos National Laboratory)
- Richard Brower (Boston University)
- Nuno Cardoso (NCSA)
- Kate Clark (NVIDIA)
- Michael Cheng (Boston University)
- Carleton DeTar (Utah University)
- Justin Foley (NIH)
- Arjun Gambhir (William and Mary)
- Joel Giedt (Rensselaer Polytechnic Institute)
- Steven Gottlieb (Indiana University)
- Kyriakos Hadjiyiannakou (Cyprus)
- Dean Howarth (Lawrence Livermore Lab, Lawrence Berkeley Lab)
- Balint Joo (OLCF, Oak Ridge National Laboratory, formerly Jefferson Lab)
- Hyung-Jin Kim (Samsung Advanced Institute of Technology)
- Bartek Kostrzewa (Bonn)
- James Osborn (Argonne National Laboratory)
- Claudio Rebbi (Boston University)
- Eloy Romero (William and Mary)
- Hauke Sandmeyer (Bielefeld)
- Mario Schröck (INFN)
- Guochun Shi (NCSA)
- Alexei Strelchenko (Fermi National Accelerator Laboratory)
- Jiqun Tu (NVIDIA)
- Alejandro Vaquero (Utah University)
- Mathias Wagner (NVIDIA)
- Andre Walker-Loud (Lawrence Berkley Laboratory)
- Evan Weinberg (NVIDIA)
- Frank Winter (Jefferson Lab)
- Yi-Bo Yang (Chinese Academy of Sciences)
Portions of this software were developed at the Innovative Systems Lab, National Center for Supercomputing Applications http://www.ncsa.uiuc.edu/AboutUs/Directorates/ISL.html
Development was supported in part by the U.S. Department of Energy under grants DE-FC02-06ER41440, DE-FC02-06ER41449, and DE-AC05-06OR23177; the National Science Foundation under grants DGE-0221680, PHY-0427646, PHY-0835713, OCI-0946441, and OCI-1060067; as well as the PRACE project funded in part by the EUs 7th Framework Programme (FP7/2007-2013) under grants RI-211528 and FP7-261557. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the Department of Energy, the National Science Foundation, or the PRACE project.