5.1. Introduction
CVODES [96] is part of a software family called SUNDIALS: SUite of Nonlinear and DIfferential/ALgebraic equation Solvers [64]. This suite consists of CVODE, ARKODE, KINSOL, and IDA, and variants of these with sensitivity analysis capabilities. CVODES is a solver for stiff and nonstiff initial value problems (IVPs) for systems of ordinary differential equation (ODEs). In addition to solving stiff and nonstiff ODE systems, CVODES has sensitivity analysis capabilities, using either the forward or the adjoint methods.
5.1.1. Historical Background
Fortran solvers for ODE initial value problems are widespread and heavily used. Two solvers that have been written at LLNL in the past are VODE [21] and VODPK [27]. VODE is a general purpose solver that includes methods for both stiff and nonstiff systems, and in the stiff case uses direct methods (full or banded) for the solution of the linear systems that arise at each implicit step. Externally, VODE is very similar to the well known solver LSODE [87]. VODPK is a variant of VODE that uses a preconditioned Krylov (iterative) method, namely GMRES, for the solution of the linear systems. VODPK is a powerful tool for large stiff systems because it combines established methods for stiff integration, nonlinear iteration, and Krylov (linear) iteration with a problem-specific treatment of the dominant source of stiffness, in the form of the user-supplied preconditioner matrix [22]. The capabilities of both VODE and VODPK have been combined in the C-language package CVODE [37].
At present, CVODE may utilize a variety of Krylov methods provided in SUNDIALS that can be used in conjuction with Newton iteration: these include the GMRES (Generalized Minimal RESidual) [90], FGMRES (Flexible Generalized Minimum RESidual) [89], Bi-CGStab (Bi-Conjugate Gradient Stabilized) [107], TFQMR (Transpose-Free Quasi-Minimal Residual) [52], and PCG (Preconditioned Conjugate Gradient) [59] linear iterative methods. As Krylov methods, these require almost no matrix storage for solving the Newton equations as compared to direct methods. However, the algorithms allow for a user-supplied preconditioner matrix, and for most problems preconditioning is essential for an efficient solution. For very large stiff ODE systems, the Krylov methods are preferable over direct linear solver methods, and are often the only feasible choice. Among the Krylov methods in SUNDIALS, we recommend GMRES as the best overall choice. However, users are encouraged to compare all options, especially if encountering convergence failures with GMRES. Bi-CGStab and TFQMR have an advantage in storage requirements, in that the number of workspace vectors they require is fixed, while that number for GMRES depends on the desired Krylov subspace size. FGMRES has an advantage in that it is designed to support preconditioners that vary between iterations (e.g. iterative methods). PCG exhibits rapid convergence and minimal workspace vectors, but only works for symmetric linear systems.
In the process of translating the VODE and VODPK algorithms into C, the overall CVODE organization has been changed considerably. One key feature of the CVODE organization is that the linear system solvers comprise a layer of code modules that is separated from the integration algorithm, allowing for easy modification and expansion of the linear solver array. A second key feature is a separate module devoted to vector operations; this facilitated the extension to multiprosessor environments with minimal impacts on the rest of the solver, resulting in PVODE [30], the parallel variant of CVODE.
CVODES is written with a functionality that is a superset of that of the pair CVODE/PVODE. Sensitivity analysis capabilities, both forward and adjoint, have been added to the main integrator. Enabling forward sensititivity computations in CVODES will result in the code integrating the so-called sensitivity equations simultaneously with the original IVP, yielding both the solution and its sensitivity with respect to parameters in the model. Adjoint sensitivity analysis, most useful when the gradients of relatively few functionals of the solution with respect to many parameters are sought, involves integration of the original IVP forward in time followed by the integration of the so-called adjoint equations backward in time. CVODES provides the infrastructure needed to integrate any final-condition ODE dependent on the solution of the original IVP (in particular the adjoint system).
Development of CVODES was concurrent with a redesign of the vector operations
module across the SUNDIALS suite. The key feature of the N_Vector module is
that it is written in terms of abstract vector operations with the actual vector
functions attached by a particular implementation (such as serial or parallel)
of N_Vector. This allows writing the SUNDIALS solvers in a manner
independent of the actual N_Vector implementation (which can be
user-supplied), as well as allowing more than one N_Vector module to be
linked into an executable file. SUNDIALS (and thus CVODES) is supplied with
serial, MPI-parallel, and both OpenMP and Pthreads thread-parallel N_Vector
implementations.
There were several motivations for choosing the C language for CVODE, and later for CVODES. First, a general movement away from Fortran and toward C in scientific computing was apparent. Second, the pointer, structure, and dynamic memory allocation features in C are extremely useful in software of this complexity. Finally, we prefer C over C++ for CVODES because of the wider availability of C compilers, the potentially greater efficiency of C, and the greater ease of interfacing the solver to applications written in extended Fortran.
5.1.2. Changes from previous versions
5.1.2.1. Changes in v6.5.0
Added the functions CVodeGetJac(), CVodeGetJacTime(),
CVodeGetJacNumSteps() to assist in debugging simulations utilizing
a matrix-based linear solver.
Added support for the SYCL backend with RAJA 2022.x.y.
Fixed an underflow bug during root finding.
A new capability to keep track of memory allocations made through the SUNMemoryHelper
classes has been added. Memory allocation stats can be accessed through the
SUNMemoryHelper_GetAllocStats() function. See the documentation for
the SUNMemoryHelper classes for more details.
Added support for CUDA v12.
Fixed an issue with finding oneMKL when using the icpx compiler with the
-fsycl flag as the C++ compiler instead of dpcpp.
Fixed the shape of the arrays returned by FN_VGetArrayPointer functions as well
as the FSUNDenseMatrix_Data, FSUNBandMatrix_Data, FSUNSparseMatrix_Data,
FSUNSparseMatrix_IndexValues, and FSUNSparseMatrix_IndexPointers functions.
Compiling and running code that uses the SUNDIALS Fortran interfaces with
bounds checking will now work.
5.1.2.2. Changes in v6.4.1
Fixed a bug with the Kokkos interfaces that would arise when using clang.
Fixed a compilation error with the Intel oneAPI 2022.2 Fortran compiler in the
Fortran 2003 interface test for the serial N_Vector.
Fixed a bug in the SUNLINSOL_LAPACKBAND and SUNLINSOL_LAPACKDENSE modules which would cause the tests to fail on some platforms.
5.1.2.3. Changes in v6.4.0
CMake 3.18.0 or newer is now required for CUDA support.
A C++14 compliant compiler is now required for C++ based features and examples e.g., CUDA, HIP, RAJA, Trilinos, SuperLU_DIST, MAGMA, GINKGO, and KOKKOS.
Added support for GPU enabled SuperLU_DIST and SuperLU_DIST v8.x.x. Removed support for SuperLU_DIST v6.x.x or older. Fix mismatched definition and declaration bug in SuperLU_DIST matrix constructor.
Added support for the Ginkgo linear
algebra library. This support includes new SUNMatrix and SUNLinearSolver
implementations, see the sections §10.16 and
§11.23.
Added new NVector, dense SUNMatrix, and dense SUNLinearSolver
implementations utilizing the Kokkos Ecosystem for
performance portability, see sections §9.19,
§10.17, and §11.24 for more information.
Fixed a bug in the CUDA and HIP vectors where N_VMaxNorm() would return
the minimum positive floating-point value for the zero vector.
Fixed a memory leak where the projection memory would not be deallocated when
calling CVodeFree().
5.1.2.4. Changes in v6.3.0
Added the function CVodeGetUserData() to retrieve the user data pointer
provided to CVodeSetUserData().
Fixed the unituitive behavior of the USE_GENERIC_MATH CMake option which
caused the double precision math functions to be used regardless of the value of
SUNDIALS_PRECISION. Now, SUNDIALS will use precision appropriate math
functions when they are available and the user may provide the math library to
link to via the advanced CMake option SUNDIALS_MATH_LIBRARY.
Changed SUNDIALS_LOGGING_ENABLE_MPI CMake option default to be ‘OFF’.
5.1.2.5. Changes in v6.2.0
Added the SUNLogger API which provides a SUNDIALS-wide
mechanism for logging of errors, warnings, informational output,
and debugging output.
Deprecated SUNNonlinSolSetPrintLevel_Newton(),
SUNNonlinSolSetInfoFile_Newton(),
SUNNonlinSolSetPrintLevel_FixedPoint(),
SUNNonlinSolSetInfoFile_FixedPoint(),
SUNLinSolSetInfoFile_PCG(), SUNLinSolSetPrintLevel_PCG(),
SUNLinSolSetInfoFile_SPGMR(), SUNLinSolSetPrintLevel_SPGMR(),
SUNLinSolSetInfoFile_SPFGMR(), SUNLinSolSetPrintLevel_SPFGMR(),
SUNLinSolSetInfoFile_SPTFQM(), SUNLinSolSetPrintLevel_SPTFQMR(),
SUNLinSolSetInfoFile_SPBCGS(), SUNLinSolSetPrintLevel_SPBCGS()
it is recommended to use the SUNLogger API instead. The SUNLinSolSetInfoFile_**
and SUNNonlinSolSetInfoFile_* family of functions are now enabled
by setting the CMake option SUNDIALS_LOGGING_LEVEL to a value >= 3.
Added the function SUNProfiler_Reset() to reset the region timings and
counters to zero.
Added the function CVodePrintAllStats() to output all of the integrator,
nonlinear solver, linear solver, and other statistics in one call. The file
scripts/sundials_csv.py contains functions for parsing the comma-separated
value output files.
Added support for integrating IVPs with constraints using BDF methods
and projecting the solution onto the constraint manifold with a user
defined projection function. This implementation is accompanied by
additions to user documentation and CVODES examples. See
CVodeSetProjFn() for more information.
Added the functions
CVodeSetEtaFixedStepBounds(),
CVodeSetEtaMaxFirstStep(),
CVodeSetEtaMaxEarlyStep(),
CVodeSetNumStepsEtaMaxEarlyStep(),
CVodeSetEtaMax(),
CVodeSetEtaMin(),
CVodeSetEtaMinErrFail(),
CVodeSetEtaMaxErrFail(),
CVodeSetNumFailsEtaMaxErrFail(), and
CVodeSetEtaConvFail() to adjust various parameters controlling changes
in step size.
Added the functions CVodeSetDeltaGammaMaxLSetup() and
CVodeSetDeltaGammaMaxBadJac() to adjust the \(\gamma\) change
thresholds to require a linear solver setup or Jacobian/precondition update,
respectively.
The behavior of N_VSetKernelExecPolicy_Sycl() has been updated to be
consistent with the CUDA and HIP vectors. The input execution policies are now
cloned and may be freed after calling N_VSetKernelExecPolicy_Sycl().
Additionally, NULL inputs are now allowed and, if provided, will reset the
vector execution policies to the defaults.
Fixed the SUNContext convenience class for C++ users to disallow copy
construction and allow move construction.
A memory leak in the SYCL vector was fixed where the execution policies were not freed when the vector was destroyed.
The include guard in nvector_mpimanyvector.h has been corrected to enable
using both the ManyVector and MPIManyVector NVector implementations in the same
simulation.
Changed exported SUNDIALS PETSc CMake targets to be INTERFACE IMPORTED instead of UNKNOWN IMPORTED.
A bug was fixed in the functions
CVodeGetNumNonlinSolvConvFails(),
CVodeGetNonlinSolvStats(),
CVodeGetSensNumNonlinSolvConvFails(),
CVodeGetSensNonlinSolvStats(),
CVodeGetStgrSensNumNonlinSolvConvFails(), and
CVodeGetStgrSensNonlinSolvStats()
where the number of nonlinear solver failures returned was the number of failed
steps due to a nonlinear solver failure i.e., if a nonlinear solve failed
with a stale Jacobian or preconditioner but succeeded after updating the
Jacobian or preconditioner, the initial failure was not included in the
nonlinear solver failure count. These functions have been updated to return the
total number of nonlinear solver failures. As such users may see an increase in
the number of failures reported.
The functions
CVodeGetNumStepSolveFails(),
CVodeGetNumStepSensSolveFails(), and
CVodeGetNumStepStgrSensSolveFails()
have been added to retrieve the number of failed steps due to a nonlinear solver
failure. The counts returned from these functions will match those previously
returned by
CVodeGetNumNonlinSolvConvFails(),
CVodeGetNonlinSolvStats(),
CVodeGetSensNumNonlinSolvConvFails(),
CVodeGetSensNonlinSolvStats(),
CVodeGetStgrSensNumNonlinSolvConvFails(), and
CVodeGetStgrSensNonlinSolvStats().
5.1.2.6. Changes in v6.1.1
Fixed exported SUNDIALSConfig.cmake.
5.1.2.7. Changes in v6.1.0
Added new reduction implementations for the CUDA and HIP NVECTORs that use
shared memory (local data storage) instead of atomics. These new implementations
are recommended when the target hardware does not provide atomic support for the
floating point precision that SUNDIALS is being built with. The HIP vector uses
these by default, but the N_VSetKernelExecPolicy_Cuda() and
N_VSetKernelExecPolicy_Hip() functions can be used to choose between
different reduction implementations.
SUNDIALS::<lib> targets with no static/shared suffix have been added for use
within the build directory (this mirrors the targets exported on installation).
CMAKE_C_STANDARD is now set to 99 by default.
Fixed exported SUNDIALSConfig.cmake when profiling is enabled without Caliper.
Fixed sundials_export.h include in sundials_config.h.
Fixed memory leaks in the SUNLINSOL_SUPERLUMT linear solver.
5.1.2.8. Changes in v6.0.0
SUNContext
SUNDIALS v6.0.0 introduces a new SUNContext object on which all other
SUNDIALS objects depend. As such, the constructors for all SUNDIALS packages,
vectors, matrices, linear solvers, nonlinear solvers, and memory helpers have
been updated to accept a context as the last input. Users upgrading to SUNDIALS
v6.0.0 will need to call SUNContext_Create() to create a context object
with before calling any other SUNDIALS library function, and then provide this
object to other SUNDIALS constructors. The context object has been introduced to
allow SUNDIALS to provide new features, such as the profiling/instrumentation
also introduced in this release, while maintaining thread-safety. See the
documentation section on the SUNContext for more details.
A script upgrade-to-sundials-6-from-5.sh has been provided with the release
(obtainable from the GitHub release page) to help ease the transition to
SUNDIALS v6.0.0. The script will add a SUNCTX_PLACEHOLDER argument to all of
the calls to SUNDIALS constructors that now require a SUNContext object. It
can also update deprecated SUNDIALS constants/types to the new names. It can be
run like this:
> ./upgrade-to-sundials-6-from-5.sh <files to update>
SUNProfiler
A capability to profile/instrument SUNDIALS library code has been added. This
can be enabled with the CMake option SUNDIALS_BUILD_WITH_PROFILING. A
built-in profiler will be used by default, but the Caliper library can also be used instead with the
CMake option ENABLE_CALIPER. See the documentation section on
profiling for more details. WARNING: Profiling will impact performance, and
should be enabled judiciously.
SUNMemoryHelper
The SUNMemoryHelper functions SUNMemoryHelper_Alloc(),
SUNMemoryHelper_Dealloc(), and SUNMemoryHelper_Copy() have been
updated to accept an opaque handle as the last input. At a minimum, user-defined
SUNMemoryHelper implementations will need to update these functions to
accept the additional argument. Typically, this handle is the execution stream
(e.g., a CUDA/HIP stream or SYCL queue) for the operation. The CUDA, HIP, and SYCL
implementations have been updated accordingly. Additionally, the constructor
SUNMemoryHelper_Sycl() has been updated to remove the SYCL queue as an
input.
NVector
Two new optional vector operations, N_VDotProdMultiLocal() and
N_VDotProdMultiAllReduce(), have been added to support
low-synchronization methods for Anderson acceleration.
The CUDA, HIP, and SYCL execution policies have been moved from the sundials
namespace to the sundials::cuda, sundials::hip, and sundials::sycl
namespaces respectively. Accordingly, the prefixes “Cuda”, “Hip”, and “Sycl”
have been removed from the execution policy classes and methods.
The Sundials namespace used by the Trilinos Tpetra NVector has been replaced
with the sundials::trilinos::nvector_tpetra namespace.
The serial, PThreads, PETSc, hypre, Parallel, OpenMP_DEV, and OpenMP vector
functions N_VCloneVectorArray_* and N_VDestroyVectorArray_* have been
deprecated. The generic N_VCloneVectorArray() and
N_VDestroyVectorArray() functions should be used instead.
The previously deprecated constructor N_VMakeWithManagedAllocator_Cuda and
the function N_VSetCudaStream_Cuda have been removed and replaced with
N_VNewWithMemHelp_Cuda() and N_VSetKerrnelExecPolicy_Cuda()
respectively.
The previously deprecated macros PVEC_REAL_MPI_TYPE and
PVEC_INTEGER_MPI_TYPE have been removed and replaced with
MPI_SUNREALTYPE and MPI_SUNINDEXTYPE respectively.
SUNLinearSolver
The following previously deprecated functions have been removed:
Removed |
Replacement |
|---|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
CVODES
Added a new function CVodeGetLinSolveStats() to get the CVODES linear
solver statistics as a group.
Added a new function, CVodeSetMonitorFn(), that takes a user-function
to be called by CVODES after every nst successfully completed time-steps.
This is intended to provide a way of monitoring the CVODES statistics
throughout the simulation.
The previously deprecated function CVodeSetMaxStepsBetweenJac has been
removed and replaced with CVodeSetJacEvalFrequency().
Deprecations
In addition to the deprecations noted elsewhere, many constants, types, and functions have been renamed so that they are properly namespaced. The old names have been deprecated and will be removed in SUNDIALS v7.0.0.
The following constants, macros, and typedefs are now deprecated:
Deprecated Name |
New Name |
|---|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
In addition, the following functions are now deprecated (compile-time warnings will be thrown if supported by the compiler):
Deprecated Name |
New Name |
|---|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
In addition, the entire sundials_lapack.h header file is now deprecated for
removal in SUNDIALS v7.0.0. Note, this header file is not needed to use the
SUNDIALS LAPACK linear solvers.
5.1.2.9. Changes in v5.8.0
The RAJA N_Vector implementation has been updated to support the SYCL
backend in addition to the CUDA and HIP backend. Users can choose the backend
when configuring SUNDIALS by using the SUNDIALS_RAJA_BACKENDS CMake
variable. This module remains experimental and is subject to change from version
to version.
A new SUNMatrix and SUNLinearSolver implementation were added to
interface with the Intel oneAPI Math Kernel Library (oneMKL). Both the matrix
and the linear solver support general dense linear systems as well as block
diagonal linear systems. See Chapter §11.14 for more
details. This module is experimental and is subject to change from version to
version.
Added a new optional function to the SUNLinearSolver API,
SUNLinSolSetZeroGuess, to indicate that the next call to SUNlinSolSolve
will be made with a zero initial guess. SUNLinearSolver implementations that do
not use the SUNLinSolNewEmpty constructor will, at a minimum, need set the
setzeroguess function pointer in the linear solver ops structure to
NULL. The SUNDIALS iterative linear solver implementations have been updated
to leverage this new set function to remove one dot product per solve.
CVODES now supports a new “matrix-embedded” SUNLinearSolver type. This type
supports user-supplied SUNLinearSolver implementations that set up and solve
the specified linear system at each linear solve call. Any matrix-related data
structures are held internally to the linear solver itself, and are not provided
by the SUNDIALS package.
Added the function CVodeSetNlsRhsFn to supply an alternative right-hand side
function for use within nonlinear system function evaluations.
The installed SUNDIALSConfig.cmake file now supports the COMPONENTS option
to find_package. The exported targets no longer have IMPORTED_GLOBAL
set.
A bug was fixed in SUNMatCopyOps where the matrix-vector product setup
function pointer was not copied.
A bug was fixed in the SPBCGS and SPTFQMR solvers for the case where a non-zero initial guess and a solution scaling vector are provided. This fix only impacts codes using SPBCGS or SPTFQMR as standalone solvers as all SUNDIALS packages utilize a zero initial guess.
5.1.2.10. Changes in v5.7.0
A new N_Vector implementation based on the SYCL abstraction layer has been added targeting Intel GPUs. At
present the only SYCL compiler supported is the DPC++ (Intel oneAPI) compiler. See Section
§9.17 for more details. This module is considered experimental and is subject to major
changes even in minor releases.
A new SUNMatrix and SUNLinearSolver implementation were added to interface with the MAGMA linear algebra library.
Both the matrix and the linear solver support general dense linear systems as well as block diagonal linear systems, and
both are targeted at GPUs (AMD or NVIDIA). See Section §11.13 for more
details.
5.1.2.11. Changes in v5.6.1
Fixed a bug in the SUNDIALS CMake which caused an error if the CMAKE_CXX_STANDARD and SUNDIALS_RAJA_BACKENDS options were not provided.
Fixed some compiler warnings when using the IBM XL compilers.
5.1.2.12. Changes in v5.6.0
A new N_Vector implementation based on the AMD ROCm HIP platform has been added. This vector can target NVIDIA or
AMD GPUs. See §9.16 for more details. This module is considered experimental and is subject
to change from version to version.
The RAJA N_Vector implementation has been updated to support the HIP backend in addition to the CUDA backend. Users
can choose the backend when configuring SUNDIALS by using the SUNDIALS_RAJA_BACKENDS CMake variable. This module
remains experimental and is subject to change from version to version.
A new optional operation, N_VGetDeviceArrayPointer, was added to the N_Vector API. This operation is useful for
N_Vectors that utilize dual memory spaces, e.g. the native SUNDIALS CUDA N_Vector.
The SUNMATRIX_CUSPARSE and SUNLINEARSOLVER_CUSOLVERSP_BATCHQR implementations no longer require the SUNDIALS CUDA
N_Vector. Instead, they require that the vector utilized provides the N_VGetDeviceArrayPointer operation, and that
the pointer returned by N_VGetDeviceArrayPointer is a valid CUDA device pointer.
5.1.2.13. Changes in v5.5.0
Refactored the SUNDIALS build system. CMake 3.12.0 or newer is now required. Users will likely see deprecation warnings, but otherwise the changes should be fully backwards compatible for almost all users. SUNDIALS now exports CMake targets and installs a SUNDIALSConfig.cmake file.
Added support for SuperLU DIST 6.3.0 or newer.
5.1.2.14. Changes in v5.4.0
Added the function CVodeSetLSNormFactor to specify the factor for converting between integrator tolerances (WRMS
norm) and linear solver tolerances (L2 norm) i.e., tol_L2 = nrmfac * tol_WRMS.
Added new functions CVodeComputeState, and CVodeGetNonlinearSystemData which advanced users might find useful if
providing a custom SUNNonlinSolSysFn.
This change may cause an error in existing user code. The CVodeF function for forward integration with
checkpointing is now subject to a restriction on the number of time steps allowed to reach the output time. This is the
same restriction applied to the CVode function. The default maximum number of steps is 500, but this may be changed
using the CVodeSetMaxNumSteps function. This change fixes a bug that could cause an infinite loop in the CVodeF
function.
The expected behavior of SUNNonlinSolGetNumIters and SUNNonlinSolGetNumConvFails in the SUNNonlinearSolver API
have been updated to specify that they should return the number of nonlinear solver iterations and convergence failures
in the most recent solve respectively rather than the cumulative number of iterations and failures across all solves
respectively. The API documentation and SUNDIALS provided SUNNonlinearSolver implementations have been updated
accordingly. As before, the cumulative number of nonlinear iterations may be retreived by calling
CVodeGetNumNonlinSolvIters, CVodeGetSensNumNonlinSolvIters, CVodeGetStgrSensNumNonlinSolvIters, the
cumulative number of failures with CVodeGetNumNonlinSolvConvFails, CVodeGetSensNumNonlinSolvConvFails,
CVodeGetStgrSensNumNonlinSolvConvFails, or both with CVodeGetNonlinSolvStats, CVodeGetSensNonlinSolvStats,
CVodeGetStgrSensNonlinSolvStats.
A minor inconsistency in checking the Jacobian evaluation frequency has been fixed. As a result codes using using a
non-default Jacobian update frequency through a call to CVodeSetMaxStepsBetweenJac will need to increase the
provided value by 1 to achieve the same behavior as before. For greater clarity the function
CVodeSetMaxStepsBetweenJac has been deprecated and replaced with CVodeSetJacEvalFrequency. Additionally, the
function CVodeSetLSetupFrequency has been added to set the frequency of calls to the linear solver setup function.
A new API, SUNMemoryHelper, was added to support GPU users who have complex memory management needs such as
using memory pools. This is paired with new constructors for the NVECTOR_CUDA and NVECTOR_RAJA modules that accept a
SUNMemoryHelper object. Refer to §2.6.1, §13,
§9.15 and §9.18 for more information.
The NVECTOR_RAJA module has been updated to mirror the NVECTOR_CUDA module. Notably, the update adds managed
memory support to the NVECTOR_RAJA module. Users of the module will need to update any calls to the N_VMake_Raja
function because that signature was changed. This module remains experimental and is subject to change from version to
version.
The NVECTOR_TRILINOS module has been updated to work with Trilinos 12.18+. This update changes the local ordinal
type to always be an int.
Added support for CUDA v11.
5.1.2.15. Changes in v5.3.0
Fixed a bug in the iterative linear solver modules where an error is not returned if the Atimes function is NULL or,
if preconditioning is enabled, the PSolve function is NULL.
Added the ability to control the CUDA kernel launch parameters for the NVECTOR_CUDA and SUNMATRIX_CUSPARSE
modules. These modules remain experimental and are subject to change from version to version. In addition, the
NVECTOR_CUDA kernels were rewritten to be more flexible. Most users should see equivalent performance or some
improvement, but a select few may observe minor performance degradation with the default settings. Users are encouraged
to contact the SUNDIALS team about any perfomance changes that they notice.
Added new capabilities for monitoring the solve phase in the SUNNONLINSOL_NEWTON and SUNNONLINSOL_FIXEDPOINT
modules, and the SUNDIALS iterative linear solver modules. SUNDIALS must be built with the CMake option
SUNDIALS_BUILD_WITH_MONITORING to use these capabilties.
Added the optional functions CVodeSetJacTimesRhsFn and CVodeSetJacTimesRhsFnB to specify an alternative
right-hand side function for computing Jacobian-vector products with the internal difference quotient approximation.
5.1.2.16. Changes in v5.2.0
Fixed a build system bug related to the Fortran 2003 interfaces when using the IBM XL compiler. When building the
Fortran 2003 interfaces with an XL compiler it is recommended to set CMAKE_Fortran_COMPILER to f2003,
xlf2003, or xlf2003_r.
Fixed a linkage bug affecting Windows users that stemmed from dllimport/dllexport attributes missing on some SUNDIALS API functions.
Fixed a memory leak from not deallocating the atolSmin0 and atolQSmin0 arrays.
Added a new SUNMatrix implementation, SUNMATRIX_CUSPARSE, that interfaces to the sparse matrix implementation
from the NVIDIA cuSPARSE library. In addition, the SUNLINSOL_CUSOLVER_BATCHQR linear solver has been updated to use
this matrix, therefore, users of this module will need to update their code. These modules are still considered to be
experimental, thus they are subject to breaking changes even in minor releases.
The functions CVodeSetLinearSolutionScaling and CVodeSetLinearSolutionScalingB were added to enable or disable
the scaling applied to linear system solutions with matrix-based linear solvers to account for a lagged value of
\(\gamma\) in the linear system matrix \(I - \gamma J\). Scaling is enabled by default when using a matrix-based
linear solver with BDF methods.
5.1.2.17. Changes in v5.1.0
Fixed a build system bug related to finding LAPACK/BLAS.
Fixed a build system bug related to checking if the KLU library works.
Fixed a build system bug related to finding PETSc when using the CMake
variables PETSC_INCLUDES and PETSC_LIBRARIES instead of PETSC_DIR.
Added a new build system option, CUDA_ARCH, that can be used to specify the CUDA
architecture to compile for.
Added two utility functions, SUNDIALSFileOpen() and
SUNDIALSFileClose() for creating/destroying file pointers that are
useful when using the Fortran 2003 interfaces.
Added support for constant damping to the SUNNonlinearSolver_FixedPoint module when using Anderson acceleration.
5.1.2.18. Changes in v5.0.0
Build system changes
Increased the minimum required CMake version to 3.5 for most SUNDIALS configurations, and 3.10 when CUDA or OpenMP with device offloading are enabled.
The CMake option
BLAS_ENABLEand the variableBLAS_LIBRARIEShave been removed to simplify builds as SUNDIALS packages do not use BLAS directly. For third party libraries that require linking to BLAS, the path to the BLAS library should be included in the variable for the third party library e.g.,SUPERLUDIST_LIBRARIESwhen enabling SuperLU_DIST.Fixed a bug in the build system that prevented the
NVECTOR_PTHREADSmodule from being built.
NVECTOR module changes
Two new functions were added to aid in creating custom
N_Vectorobjects. The constructorN_VNewEmpty()allocates an “empty” genericN_Vectorwith the object’s content pointer and the function pointers in the operations structure initialized toNULL. When used in the constructor for custom objects this function will ease the introduction of any new optional operations to theN_VectorAPI by ensuring only required operations need to be set. Additionally, the functionN_VCopyOps()has been added to copy the operation function pointers between vector objects. When used in clone routines for custom vector objects these functions also will ease the introduction of any new optional operations to theN_VectorAPI by ensuring all operations are copied when cloning objects. See §9.1.2 for more details.Two new
N_Vectorimplementations,NVECTOR_MANYVECTORandNVECTOR_MPIMANYVECTOR, have been created to support flexible partitioning of solution data among different processing elements (e.g., CPU + GPU) or for multi-physics problems that couple distinct MPI-based simulations together. This implementation is accompanied by additions to user documentation and SUNDIALS examples. See §9.22 and §9.23 for more details.One new required vector operation and ten new optional vector operations have been added to the
N_VectorAPI. The new required operation, , returns the global length of an . The optional operations have been added to support the newNVECTOR_MPIMANYVECTORimplementation. The operation must be implemented by subvectors that are combined to create anNVECTOR_MPIMANYVECTOR, but is not used outside of this context. The remaining nine operations are optional local reduction operations intended to eliminate unnecessary latency when performing vector reduction operations (norms, etc.) on distributed memory systems. The optional local reduction vector operations areN_VDotProdLocal(),N_VMaxNormLocal(),N_VL1NormLocal(),N_VWSqrSumLocal(),N_VWSqrSumMaskLocal(),N_VInvTestLocal(),N_VConstrMaskLocal(),N_VMinLocal(), andN_VMinQuotientLocal(). If anN_Vectorimplementation defines any of the local operations as , then theNVECTOR_MPIMANYVECTORwill call standardN_Vectoroperations to complete the computation.An additional
N_Vectorimplementation,NVECTOR_MPIPLUSX, has been created to support the MPI+X paradigm where X is a type of on-node parallelism (e.g., OpenMP, CUDA). The implementation is accompanied by additions to user documentation and SUNDIALS examples. See §9.24 for more details.The and functions have been removed from the
NVECTOR_CUDAandNVECTOR_RAJAimplementations respectively. Accordingly, thenvector_mpicuda.h,libsundials_nvecmpicuda.lib,libsundials_nvecmpicudaraja.lib, and files have been removed. Users should use theNVECTOR_MPIPLUSXmodule coupled in conjunction with theNVECTOR_CUDAorNVECTOR_RAJAmodules to replace the functionality. The necessary changes are minimal and should require few code modifications. See the programs in and for examples of how to use theNVECTOR_MPIPLUSXmodule with theNVECTOR_CUDAandNVECTOR_RAJAmodules respectively.Fixed a memory leak in the
NVECTOR_PETSCmodule clone function.Made performance improvements to the
NVECTOR_CUDAmodule. Users who utilize a non-default stream should no longer see default stream synchronizations after memory transfers.Added a new constructor to the
NVECTOR_CUDAmodule that allows a user to provide custom allocate and free functions for the vector data array and internal reduction buffer. See §9.15 for more details.Added new Fortran 2003 interfaces for most
N_Vectormodules. See §9 for more details on how to use the interfaces.Added three new
N_Vectorutility functionsN_VGetVecAtIndexVectorArray(),N_VSetVecAtIndexVectorArray(), andN_VNewVectorArray()for working with arrays when using the Fortran 2003 interfaces.
SUNMatrix module changes
Two new functions were added to aid in creating custom
SUNMatrixobjects. The constructorSUNMatNewEmpty()allocates an “empty” genericSUNMatrixwith the object’s content pointer and the function pointers in the operations structure initialized to . When used in the constructor for custom objects this function will ease the introduction of any new optional operations to theSUNMatrixAPI by ensuring only required operations need to be set. Additionally, the functionSUNMatCopyOps()has been added to copy the operation function pointers between matrix objects. When used in clone routines for custom matrix objects these functions also will ease the introduction of any new optional operations to theSUNMatrixAPI by ensuring all operations are copied when cloning objects. See §10 for more details.A new operation,
SUNMatMatvecSetup(), was added to theSUNMatrixAPI to perform any setup necessary for computing a matrix-vector product. This operation is useful forSUNMatriximplementations which need to prepare the matrix itself, or communication structures before performing the matrix-vector product. Users who have implemented customSUNMatrixmodules will need to at least update their code to set the corresponding structure member toNULL. See §10.2 for more details.The generic
SUNMatrixAPI now defines error codes to be returned bySUNMatrixoperations. Operations which return an integer flag indiciating success/failure may return different values than previously. See §10.2.1 for more details.A new
SUNMatrix(andSUNLinearSolver) implementation was added to facilitate the use of the SuperLU_DIST library with SUNDIALS. See §10.15 for more details.Added new Fortran 2003 interfaces for most
SUNMatrixmodules. See §10 for more details on how to use the interfaces.
SUNLinearSolver module changes
A new function was added to aid in creating custom
SUNLinearSolverobjects. The constructor allocates an “empty” genericSUNLinearSolverwith the object’s content pointer and the function pointers in the operations structure initialized to . When used in the constructor for custom objects this function will ease the introduction of any new optional operations to theSUNLinearSolverAPI by ensuring only required operations need to be set. See §11.1.8 for more details.The return type of the
SUNLinearSolverAPI function has changed from to to be consistent with the type used to store row indices in dense and banded linear solver modules.Added a new optional operation to the
SUNLinearSolverAPI,SUNLinSolLastFlag(), that returns a for identifying the linear solver module.The
SUNLinearSolverAPI has been updated to make the initialize and setup functions optional.A new
SUNLinearSolver(andSUNMatrix) implementation was added to facilitate the use of the SuperLU_DIST library with SUNDIALS. See §11.20 for more details.Added a new
SUNLinearSolverimplementation, SUNLINEARSOLVER_CUSOLVERSP, which leverages the NVIDIA cuSOLVER sparse batched QR method for efficiently solving block diagonal linear systems on NVIDIA GPUs.Added three new accessor functions to the
SUNLINSOL_KLUmodule,SUNLinSol_KLUGetSymbolic(), ,SUNLinSol_KLUGetNumeric()andSUNLinSol_KLUGetCommon(), to provide user access to the underlying KLU solver structures. See §11.10 for more details.Added new Fortran 2003 interfaces for most
SUNLinearSolvermodules. See §11 for more details on how to use the interfaces.
SUNNonlinearSolver module changes
A new function was added to aid in creating custom
SUNNonlinearSolverobjects. The constructorSUNNonlinSolSetConvTestFN()allocates an “empty” genericSUNNonlinearSolverwith the object’s content pointer and the function pointers in the operations structure initialized to . When used in the constructor for custom objects this function will ease the introduction of any new optional operations to theSUNNonlinearSolverAPI by ensuring only required operations need to be set. See §12.1.7 for more details.To facilitate the use of user supplied nonlinear solver convergence test functions the function in the
SUNNonlinearSolverAPI has been updated to take a data pointer as input. The supplied data pointer will be passed to the nonlinear solver convergence test function on each call.The inputs values passed to the first two inputs of the function
SUNNonlinSolSolve()in theSUNNonlinearSolverhave been changed to be the predicted state and the initial guess for the correction to that state. Additionally, the definitions ofSUNNonlinSolLSetupFn()andSUNNonlinSolLSolveFn()in theSUNNonlinearSolverAPI have been updated to remove unused input parameters. For more information on the nonlinear system formulation see §12.4 and for more details on the API functions see §12.Added a new
SUNNonlinearSolverimplementation,SUNNONLINSOL_PETSC, which interfaces to the PETSc SNES nonlinear solver API. See §12.9 for more details.Added new Fortran 2003 interfaces for most
SUNNonlinearSolvermodules. See §2.5 for more details on how to use the interfaces.
5.1.2.18.1. CVODES changes
Fixed a bug in the CVODES constraint handling where the step size could be set below the minimum step size.
Fixed a bug in the CVODES nonlinear solver interface where the norm of the accumulated correction was not updated when using a non-default convergence test function.
Fixed a bug in the CVODES
cvRescalefunction where the loops to compute the array of scalars for the fused vector scale operation stopped one iteration early.Fixed a bug where the
CVodeFfunction would return the wrong flag under certrain cirumstances.Fixed a bug where the
CVodeFfunction would not return a root inCV_NORMAL_STEPmode if the root occurred after the desired output time.Removed extraneous calls to
N_VMinfor simulations where the scalar valued absolute tolerance, or all entries of the vector-valued absolute tolerance array, are strictly positive. In this scenario, CVODES will remove at least one global reduction per time step.The CVLS interface has been updated to only zero the Jacobian matrix before calling a user-supplied Jacobian evaluation function when the attached linear solver has type
SUNLINEARSOLVER_DIRECT.A new linear solver interface function
CVLsLinSysFnwas added as an alternative method for evaluating the linear system \(M = I - \gamma J\).Added new functions,
CVodeGetCurrentGamma,CVodeGetCurrentState,CVodeGetCurrentStateSens, andCVodeGetCurrentSensSolveIndexwhich may be useful to users who choose to provide their own nonlinear solver implementations.Added a Fortran 2003 interface to CVODES. See Chapter §2.5 for more details.
5.1.2.19. Changes in v4.1.0
An additional N_Vector implementation was added for the TPETRA vector from
the Trilinos library to facilitate interoperability between SUNDIALS and
Trilinos. This implementation is accompanied by additions to user documentation
and SUNDIALS examples.
A bug was fixed where a nonlinear solver object could be freed twice in some use cases.
The EXAMPLES_ENABLE_RAJA CMake option has been removed. The option
EXAMPLES_ENABLE_CUDA enables all examples that use CUDA including the RAJA
examples with a CUDA back end (if the RAJA N_Vector is enabled).
The implementation header file cvodes_impl.h is no longer installed. This
means users who are directly manipulating the CVodeMem structure will need
to update their code to use CVODES’s public API.
Python is no longer required to run make test and make test_install.
5.1.2.20. Changes in v4.0.2
Added information on how to contribute to SUNDIALS and a contributing agreement.
Moved definitions of DLS and SPILS backwards compatibility functions to a source
file. The symbols are now included in the CVODES library,
libsundials_cvodes.
5.1.2.21. Changes in v4.0.1
No changes were made in this release.
5.1.2.22. Changes in v4.0.0
CVODES’ previous direct and iterative linear solver interfaces, CVDLS and
CVSPILS, have been merged into a single unified linear solver interface,
CVLS, to support any valid SUNLinearSolver module. This includes the
“DIRECT” and “ITERATIVE” types as well as the new “MATRIX_ITERATIVE” type.
Details regarding how CVLS utilizes linear solvers of each type as well as
discussion regarding intended use cases for user-supplied SUNLinearSolver
implementations are included in Chapter §11. All CVODES example
programs and the standalone linear solver examples have been updated to use the
unified linear solver interface.
The unified interface for the new CVLS module is very similar to the previous CVDLS and CVSPILS interfaces. To minimize challenges in user migration to the new names, the previous C routine names may still be used; these will be deprecated in future releases, so we recommend that users migrate to the new names soon.
The names of all constructor routines for SUNDIALS-provided SUNLinearSolver
implementations have been updated to follow the naming convention
SUNLinSol_* where * is the name of the linear solver. The new names are
SUNLinSol_Band, SUNLinSol_Dense, SUNLinSol_KLU,
SUNLinSol_LapackBand, SUNLinSol_LapackDense, SUNLinSol_PCG,
SUNLinSol_SPBCGS, SUNLinSol_SPFGMR, SUNLinSol_SPGMR,
SUNLinSol_SPTFQMR, and SUNLinSol_SuperLUMT. Solver-specific “set”
routine names have been similarly standardized. To minimize challenges in user
migration to the new names, the previous routine names may still be used; these
will be deprecated in future releases, so we recommend that users migrate to the
new names soon. All CVODES example programs and the standalone linear solver
examples have been updated to use the new naming convention.
The SUNBandMatrix constructor has been simplified to remove the storage
upper bandwidth argument.
SUNDIALS integrators have been updated to utilize generic nonlinear solver
modules defined through the SUNNonlinearSolver API. This API will ease the
addition of new nonlinear solver options and allow for external or user-supplied
nonlinear solvers. The SUNNonlinearSolver API and SUNDIALS provided modules
are described in Chapter §12 and follow the same object
oriented design and implementation used by the N_Vector, SUNMatrix, and
SUNLinearSolver modules. Currently two SUNNonlinearSolver
implementations are provided, SUNNONLINSOL_NEWTON and
SUNNONLINSOL_FIXEDPOINT. These replicate the previous integrator specific
implementations of a Newton iteration and a fixed-point iteration (previously
referred to as a functional iteration), respectively. Note the
SUNNONLINSOL_FIXEDPOINT module can optionally utilize Anderson’s method to
accelerate convergence. Example programs using each of these nonlinear solver
modules in a standalone manner have been added and all CVODES example programs
have been updated to use generic SUNNonlinearSolver modules.
With the introduction of SUNNonlinearSolver modules, the input parameter
iter to CVodeCreate has been removed along with the function
CVodeSetIterType and the constants CV_NEWTON and CV_FUNCTIONAL.
Instead of specifying the nonlinear iteration type when creating the CVODES
memory structure, CVODES uses the SUNNONLINSOL_NEWTON module implementation
of a Newton iteration by default. For details on using a non-default or
user-supplied nonlinear solver see Chapters §5.4.1,
§5.4.3, and §5.4.4. CVODES functions for
setting the nonlinear solver options (e.g., CVodeSetMaxNonlinIters) or
getting nonlinear solver statistics (e.g., CVodeGetNumNonlinSolvIters)
remain unchanged and internally call generic SUNNonlinearSolver functions as
needed.
Three fused vector operations and seven vector array operations have been added
to the N_Vector API. These optional operations are disabled by default and
may be activated by calling vector specific routines after creating an
N_Vector (see Chapter §9 for more details). The
new operations are intended to increase data reuse in vector operations, reduce
parallel communication on distributed memory systems, and lower the number of
kernel launches on systems with accelerators. The fused operations are
N_VLinearCombination, N_VScaleAddMulti, and N_VDotProdMulti and the
vector array operations are N_VLinearCombinationVectorArray,
N_VScaleVectorArray, N_VConstVectorArray, N_VWrmsNormVectorArray,
N_VWrmsNormMaskVectorArray, N_VScaleAddMultiVectorArray, and
N_VLinearCombinationVectorArray. If an N_Vector implementation defines
any of these operations as NULL, then standard N_Vector operations will
automatically be called as necessary to complete the computation. Multiple
updates to NVECTOR_CUDA were made:
Changed
N_VGetLength_Cudato return the global vector length instead of the local vector length.Added
N_VGetLocalLength_Cudato return the local vector length.Added
N_VGetMPIComm_Cudato return the MPI communicator used.Removed the accessor functions in the namespace suncudavec.
Changed the
N_VMake_Cudafunction to take a host data pointer and a device data pointer instead of anN_VectorContent_Cudaobject.Added the ability to set the
cudaStream_tused for execution of theNVECTOR_CUDAkernels. See the functionN_VSetCudaStreams_Cuda.Added
N_VNewManaged_Cuda,N_VMakeManaged_Cuda, andN_VIsManagedMemory_Cudafunctions to accommodate using managed memory with theNVECTOR_CUDA.
Multiple changes to NVECTOR_RAJA were made:
Changed
N_VGetLength_Rajato return the global vector length instead of the local vector length.Added
N_VGetLocalLength_Rajato return the local vector length.Added
N_VGetMPIComm_Rajato return the MPI communicator used.Removed the accessor functions in the namespace suncudavec.
A new N_Vector implementation for leveraging OpenMP 4.5+ device offloading
has been added, NVECTOR_OPENMPDEV. See §9.20 for more
details. Two changes were made in the CVODE/CVODES/ARKODE initial step size
algorithm:
Fixed an efficiency bug where an extra call to the right hand side function was made.
Changed the behavior of the algorithm if the max-iterations case is hit. Before the algorithm would exit with the step size calculated on the penultimate iteration. Now it will exit with the step size calculated on the final iteration.
5.1.2.23. Changes in v3.2.1
The changes in this minor release include the following:
Fixed a bug in the CUDA
N_Vectorwhere theN_VInvTestoperation could write beyond the allocated vector data.Fixed library installation path for multiarch systems. This fix changes the default library installation path to
CMAKE_INSTALL_PREFIX/CMAKE_INSTALL_LIBDIRfromCMAKE_INSTALL_PREFIX/lib.CMAKE_INSTALL_LIBDIRis automatically set, but is available as a CMake option that can modified.
5.1.2.24. Changes in v3.2.0
Support for optional inequality constraints on individual components of the
solution vector has been added to CVODE and CVODES. See Chapter
§5.2 and the description of
CVodeSetConstraints() for more details. Use of CVodeSetConstraints
requires the N_Vector operations N_MinQuotient, N_VConstrMask, and
N_VCompare that were not previously required by CVODE and CVODES.
Fixed a thread-safety issue when using ajdoint sensitivity analysis.
Fixed a problem with setting sunindextype which would occur with some
compilers (e.g. armclang) that did not define __STDC_VERSION__.
Added hybrid MPI/CUDA and MPI/RAJA vectors to allow use of more than one MPI rank when using a GPU system. The vectors assume one GPU device per MPI rank.
Changed the name of the RAJA N_Vector library to
libsundials_nveccudaraja.lib from libsundials_nvecraja.lib to better
reflect that we only support CUDA as a backend for RAJA currently.
Several changes were made to the build system:
CMake 3.1.3 is now the minimum required CMake version.
Deprecate the behavior of the
SUNDIALS_INDEX_TYPECMake option and added theSUNDIALS_INDEX_SIZECMake option to select thesunindextypeinteger size.The native CMake FindMPI module is now used to locate an MPI installation.
If MPI is enabled and MPI compiler wrappers are not set, the build system will check if
CMAKE_<language>_COMPILERcan compile MPI programs before trying to locate and use an MPI installation.The previous options for setting MPI compiler wrappers and the executable for running MPI programs have been have been depreated. The new options that align with those used in native CMake FindMPI module are
MPI_C_COMPILER,MPI_CXX_COMPILER,MPI_Fortran_COMPILER, andMPIEXEC_EXECUTABLE.When a Fortran name-mangling scheme is needed (e.g.,
ENABLE_LAPACKisON) the build system will infer the scheme from the Fortran compiler. If a Fortran compiler is not available or the inferred or default scheme needs to be overridden, the advanced optionsSUNDIALS_F77_FUNC_CASEandSUNDIALS_F77_FUNC_UNDERSCOREScan be used to manually set the name-mangling scheme and bypass trying to infer the scheme.Parts of the main CMakeLists.txt file were moved to new files in the
srcandexampledirectories to make the CMake configuration file structure more modular.
5.1.2.25. Changes in v3.1.2
The changes in this minor release include the following:
Updated the minimum required version of CMake to 2.8.12 and enabled using rpath by default to locate shared libraries on OSX.
Fixed Windows specific problem where
sunindextypewas not correctly defined when using 64-bit integers for the SUNDIALS index type. On Windowssunindextypeis now defined as the MSVC basic type__int64.Added sparse SUNMatrix “Reallocate” routine to allow specification of the nonzero storage.
Updated the KLU SUNLinearSolver module to set constants for the two reinitialization types, and fixed a bug in the full reinitialization approach where the sparse SUNMatrix pointer would go out of scope on some architectures.
Updated the “ScaleAdd” and “ScaleAddI” implementations in the sparse SUNMatrix module to more optimally handle the case where the target matrix contained sufficient storage for the sum, but had the wrong sparsity pattern. The sum now occurs in-place, by performing the sum backwards in the existing storage. However, it is still more efficient if the user-supplied Jacobian routine allocates storage for the sum \(I+\gamma J\) manually (with zero entries if needed).
Added new example,
cvRoberts_FSA_dns_Switch.c, which demonstrates switching on/off forward sensitivity computations. This example came from the usage notes page of the SUNDIALS website.The misnamed function
CVSpilsSetJacTimesSetupFnBShas been deprecated and replaced byCVSpilsSetJacTimesBS. The deprecated functionCVSpilsSetJacTimesSetupFnBSwill be removed in the next major release.Changed the LICENSE install path to
instdir/include/sundials.
5.1.2.26. Changes in v3.1.1
The changes in this minor release include the following:
Fixed a minor bug in the cvSLdet routine, where a return was missing in the error check for three inconsistent roots.
Fixed a potential memory leak in the SPGMR and SPFGMR linear solvers: if “Initialize” was called multiple times then the solver memory was reallocated (without being freed).
Updated KLU
SUNLinearSolvermodule to use atypedeffor the precision-specific solve function to be used (to avoid compiler warnings).Added missing typecasts for some
(void*)pointers (again, to avoid compiler warnings).Bugfix in
sunmatrix_sparse.cwhere we had usedintinstead ofsunindextypein one location.Added missing
#include <stdio.h>inN_VectorandSUNMatrixheader files.Fixed an indexing bug in the CUDA
N_Vectorimplementation ofN_VWrmsNormMaskand revised the RAJAN_Vectorimplementation ofN_VWrmsNormMaskto work with mask arrays using values other than zero or one. Replaceddoublewithrealtypein the RAJA vector test functions.
In addition to the changes above, minor corrections were also made to the example programs, build system, and user documentation.
5.1.2.27. Changes in v3.1.0
Added N_Vector print functions that write vector data to a specified file
(e.g., N_VPrintFile_Serial).
Added make test and make test_install options to the build system for
testing SUNDIALS after building with make and installing with make
install respectively.
5.1.2.28. Changes in v3.0.0
All interfaces to matrix structures and linear solvers have been reworked, and all example programs have been updated. The goal of the redesign of these interfaces was to provide more encapsulation and ease in interfacing custom linear solvers and interoperability with linear solver libraries. Specific changes include:
Added generic SUNMATRIX module with three provided implementations: dense, banded and sparse. These replicate previous SUNDIALS Dls and Sls matrix structures in a single object-oriented API.
Added example problems demonstrating use of generic SUNMATRIX modules.
Added generic SUNLINEARSOLVER module with eleven provided implementations: dense, banded, LAPACK dense, LAPACK band, KLU, SuperLU_MT, SPGMR, SPBCGS, SPTFQMR, SPFGMR, PCG. These replicate previous SUNDIALS generic linear solvers in a single object-oriented API.
Added example problems demonstrating use of generic SUNLINEARSOLVER modules.
Expanded package-provided direct linear solver (Dls) interfaces and scaled, preconditioned, iterative linear solver (Spils) interfaces to utilize generic SUNMATRIX and SUNLINEARSOLVER objects.
Removed package-specific, linear solver-specific, solver modules (e.g. CVDENSE, KINBAND, IDAKLU, ARKSPGMR) since their functionality is entirely replicated by the generic Dls/Spils interfaces and SUNLINEARSOLVER/SUNMATRIX modules. The exception is CVDIAG, a diagonal approximate Jacobian solver available to CVODE and CVODES.
Converted all SUNDIALS example problems to utilize new generic SUNMATRIX and SUNLINEARSOLVER objects, along with updated Dls and Spils linear solver interfaces.
Added Spils interface routines to ARKode, CVODE, CVODES, IDA and IDAS to allow specification of a user-provided “JTSetup” routine. This change supports users who wish to set up data structures for the user-provided Jacobian-times-vector (“JTimes”) routine, and where the cost of one JTSetup setup per Newton iteration can be amortized between multiple JTimes calls.
Two additional N_Vector implementations were added – one for CUDA and one
for RAJA vectors. These vectors are supplied to provide very basic support for
running on GPU architectures. Users are advised that these vectors both move all
data to the GPU device upon construction, and speedup will only be realized if
the user also conducts the right-hand-side function evaluation on the device. In
addition, these vectors assume the problem fits on one GPU. Further information
about RAJA, users are referred to th web site, https://software.llnl.gov/RAJA/.
These additions are accompanied by additions to various interface functions and
to user documentation.
All indices for data structures were updated to a new sunindextype that can
be configured to be a 32- or 64-bit integer data index type. sunindextype is
defined to be int32_t or int64_t when portable types are supported,
otherwise it is defined as int or long int. The Fortran interfaces
continue to use long int for indices, except for their sparse matrix
interface that now uses the new sunindextype. This new flexible capability
for index types includes interfaces to PETSc, hypre, SuperLU_MT, and KLU with
either 32-bit or 64-bit capabilities depending how the user configures SUNDIALS.
To avoid potential namespace conflicts, the macros defining booleantype
values TRUE and FALSE have been changed to SUNTRUE and SUNFALSE
respectively.
Temporary vectors were removed from preconditioner setup and solve routines for all packages. It is assumed that all necessary data for user-provided preconditioner operations will be allocated and stored in user-provided data structures.
The file include/sundials_fconfig.h was added. This file contains SUNDIALS
type information for use in Fortran programs.
Added functions SUNDIALSGetVersion and SUNDIALSGetVersionNumber to get
SUNDIALS release version information at runtime.
The build system was expanded to support many of the xSDK-compliant keys. The xSDK is a movement in scientific software to provide a foundation for the rapid and efficient production of high-quality, sustainable extreme-scale scientific applications. More information can be found at, https://xsdk.info.
In addition, numerous changes were made to the build system. These include the
addition of separate BLAS_ENABLE and BLAS_LIBRARIES CMake variables,
additional error checking during CMake configuration, minor bug fixes, and
renaming CMake options to enable/disable examples for greater clarity and an
added option to enable/disable Fortran 77 examples. These changes included
changing EXAMPLES_ENABLE to EXAMPLES_ENABLE_C, changing CXX_ENABLE
to EXAMPLES_ENABLE_CXX, changing F90_ENABLE to EXAMPLES_ENABLE_F90,
and adding an EXAMPLES_ENABLE_F77 option.
A bug fix was made in CVodeFree to call lfree unconditionally (if
non-NULL).
Corrections and additions were made to the examples, to installation-related files, and to the user documentation.
5.1.2.29. Changes in v2.9.0
Two additional N_Vector implementations were added – one for Hypre
(parallel) ParVector vectors, and one for PETSc vectors. These additions are
accompanied by additions to various interface functions and to user
documentation.
Each N_Vector module now includes a function, N_VGetVectorID, that
returns the N_Vector module name.
A bug was fixed in the interpolation functions used in solving backward problems for adjoint sensitivity analysis.
For each linear solver, the various solver performance counters are now
initialized to 0 in both the solver specification function and in solver
linit function. This ensures that these solver counters are initialized upon
linear solver instantiation as well as at the beginning of the problem solution.
A memory leak was fixed in the banded preconditioner interface. In addition, updates were done to return integers from linear solver and preconditioner ’free’ functions.
The Krylov linear solver Bi-CGstab was enhanced by removing a redundant dot product. Various additions and corrections were made to the interfaces to the sparse solvers KLU and SuperLU_MT, including support for CSR format when using KLU.
In interpolation routines for backward problems, added logic to bypass sensitivity interpolation if input sensitivity argument is NULL.
New examples were added for use of sparse direct solvers within sensitivity integrations and for use of OpenMP.
Minor corrections and additions were made to the CVODES solver, to the examples, to installation-related files, and to the user documentation.
5.1.2.30. Changes in v2.8.0
Two major additions were made to the linear system solvers that are available
for use with the CVODES solver. First, in the serial case, an interface to the
sparse direct solver KLU was added. Second, an interface to SuperLU_MT, the
multi-threaded version of SuperLU, was added as a thread-parallel sparse direct
solver option, to be used with the serial version of the N_Vector module. As
part of these additions, a sparse matrix (CSC format) structure was added to
CVODES.
Otherwise, only relatively minor modifications were made to the CVODES solver:
In cvRootfind, a minor bug was corrected, where the input array rootdir
was ignored, and a line was added to break out of root-search loop if the
initial interval size is below the tolerance ttol.
In CVLapackBand, the line smu = MIN(N-1,mu+ml) was changed to smu = mu
+ ml to correct an illegal input error for DGBTRF/DGBTRS.
Some minor changes were made in order to minimize the differences between the sources for private functions in CVODES and CVODE.
An option was added in the case of Adjoint Sensitivity Analysis with dense or
banded Jacobian: With a call to CVDlsSetDenseJacFnBS or
CVDlsSetBandJacFnBS, the user can specify a user-supplied Jacobian function
of type CVDls***JacFnBS, for the case where the backward problem depends on
the forward sensitivities.
In CVodeQuadSensInit, the line cv_mem->cv_fQS_data = ... was corrected
(missing Q).
In the User Guide, a paragraph was added in Section 6.2.1 on CVodeAdjReInit,
and a paragraph was added in Section 6.2.9 on CVodeGetAdjY. In the example
cvsRoberts_ASAi_dns, the output was revised to include the use of
CVodeGetAdjY.
Two minor bugs were fixed regarding the testing of input on the first call to
CVode – one involving tstop and one involving the initialization of
*tret.
For the Adjoint Sensitivity Analysis case in which the backward problem depends
on the forward sensitivities, options have been added to allow for user-supplied
pset, psolve, and jtimes functions.
In order to avoid possible name conflicts, the mathematical macro and function
names MIN, MAX, SQR, RAbs, RSqrt, RExp, RPowerI, and
RPowerR were changed to SUNMIN, SUNMAX, SUNSQR, SUNRabs,
SUNRsqrt, SUNRexp, SRpowerI, and SUNRpowerR, respectively. These
names occur in both the solver and example programs.
In the example cvsHessian_ASA_FSA, an error was corrected in the function
fB2: y2 in place of y3 in the third term of Ith(yBdot,6).
Two new N_Vector modules have been added for thread-parallel computing
environments — one for OpenMP, denoted NVECTOR_OPENMP, and one for Pthreads,
denoted NVECTOR_PTHREADS.
With this version of SUNDIALS, support and documentation of the Autotools mode of installation is being dropped, in favor of the CMake mode, which is considered more widely portable.
5.1.2.31. Changes in v2.7.0
One significant design change was made with this release: The problem size and
its relatives, bandwidth parameters, related internal indices, pivot arrays, and
the optional output lsflag have all been changed from type int to type
long int, except for the problem size and bandwidths in user calls to
routines specifying BLAS/LAPACK routines for the dense/band linear solvers. The
function NewIntArray is replaced by a pair NewIntArray / NewLintArray,
for int and long int arrays, respectively. In a minor change to the user
interface, the type of the index which in CVODES was changed from long
int to int.
Errors in the logic for the integration of backward problems were identified and fixed.
A large number of minor errors have been fixed. Among these are the following:
In CVSetTqBDF, the logic was changed to avoid a divide by zero. After the
solver memory is created, it is set to zero before being filled. In each linear
solver interface function, the linear solver memory is freed on an error return,
and the **Free function now includes a line setting to NULL the main memory
pointer to the linear solver memory. In the rootfinding functions
CVRcheck1 / CVRcheck2, when an exact zero is found, the array glo of
\(g\) values at the left endpoint is adjusted, instead of shifting the
\(t\) location tlo slightly. In the installation files, we modified the
treatment of the macro SUNDIALS_USE_GENERIC_MATH, so that the parameter
GENERIC_MATH_LIB is either defined (with no value) or not defined.
5.1.2.32. Changes in v2.6.0
Two new features related to the integration of ODE IVP problems were added in this release: (a) a new linear solver module, based on BLAS and LAPACK for both dense and banded matrices, and (b) an option to specify which direction of zero-crossing is to be monitored while performing rootfinding.
This version also includes several new features related to sensitivity analysis, among which are: (a) support for integration of quadrature equations depending on both the states and forward sensitivity (and thus support for forward sensitivity analysis of quadrature equations), (b) support for simultaneous integration of multiple backward problems based on the same underlying ODE (e.g., for use in an forward-over-adjoint method for computing second order derivative information), (c) support for backward integration of ODEs and quadratures depending on both forward states and sensitivities (e.g., for use in computing second-order derivative information), and (d) support for reinitialization of the adjoint module.
The user interface has been further refined. Some of the API changes involve:
(a) a reorganization of all linear solver modules into two families (besides the
existing family of scaled preconditioned iterative linear solvers, the direct
solvers, including the new LAPACK-based ones, were also organized into a
direct family); (b) maintaining a single pointer to user data, optionally
specified through a Set-type function; and (c) a general streamlining of the
preconditioner modules distributed with the solver. Moreover, the prototypes of
all functions related to integration of backward problems were modified to
support the simultaneous integration of multiple problems. All backward problems
defined by the user are internally managed through a linked list and identified
in the user interface through a unique identifier.
5.1.2.33. Changes in v2.5.0
The main changes in this release involve a rearrangement of the entire SUNDIALS
source tree (see §5.3). At the user interface level,
the main impact is in the mechanism of including SUNDIALS header files which
must now include the relative path (e.g. #include <cvode/cvode.h>).
Additional changes were made to the build system: all exported header files are
now installed in separate subdirectories of the instaltion include directory.
In the adjoint solver module, the following two bugs were fixed: in CVodeF
the solver was sometimes incorrectly taking an additional step before returning
control to the user (in CV_NORMAL mode) thus leading to a failure in the
interpolated output function; in CVodeB, while searching for the current
check point, the solver was sometimes reaching outside the integration interval
resulting in a segmentation fault.
The functions in the generic dense linear solver (sundials_dense and
sundials_smalldense) were modified to work for rectangular \(m \times
n\) matrices (\(m \le n\)), while the factorization and solution functions
were renamed to DenseGETRF / denGETRF and DenseGETRS / denGETRS,
respectively. The factorization and solution functions in the generic band
linear solver were renamed BandGBTRF and BandGBTRS, respectively.
5.1.2.34. Changes in v2.4.0
CVSPBCG and CVSPTFQMR modules have been added to interface with the Scaled Preconditioned Bi-CGstab (SPBCG) and Scaled Preconditioned Transpose-Free Quasi-Minimal Residual (SPTFQMR) linear solver modules, respectively (for details see Chapter §5.4.1). At the same time, function type names for Scaled Preconditioned Iterative Linear Solvers were added for the user-supplied Jacobian-times-vector and preconditioner setup and solve functions.
A new interpolation method was added to the CVODES adjoint module. The function
CVadjMalloc has an additional argument which can be used to select the
desired interpolation scheme.
The deallocation functions now take as arguments the address of the respective memory block pointer.
To reduce the possibility of conflicts, the names of all header files have been
changed by adding unique prefixes (cvodes_ and sundials_). When using
the default installation procedure, the header files are exported under various
subdirectories of the target include directory. For more details see
Appendix §14.
5.1.2.35. Changes in v2.3.0
A minor bug was fixed in the interpolation functions of the adjoint CVODES module.
5.1.2.36. Changes in v2.2.0
The user interface has been further refined. Several functions used for setting optional inputs were combined into a single one. An optional user-supplied routine for setting the error weight vector was added. Additionally, to resolve potential variable scope issues, all SUNDIALS solvers release user data right after its use. The build systems has been further improved to make it more robust.
5.1.2.37. Changes in v2.1.2
A bug was fixed in the CVode function that was potentially leading to
erroneous behaviour of the rootfinding procedure on the integration first step.
5.1.2.38. Changes in v2.1.1
This CVODES release includes bug fixes related to forward sensitivity computations (possible loss of accuray on a BDF order increase and incorrect logic in testing user-supplied absolute tolerances). In addition, we have added the option of activating and deactivating forward sensitivity calculations on successive CVODES runs without memory allocation/deallocation.
Other changes in this minor SUNDIALS release affect the build system.
5.1.2.39. Changes in v2.1.0
The major changes from the previous version involve a redesign of the user
interface across the entire SUNDIALS suite. We have eliminated the mechanism of
providing optional inputs and extracting optional statistics from the solver
through the iopt and ropt arrays. Instead, CVODES now provides a set of
routines (with prefix CVodeSet) to change the default values for various
quantities controlling the solver and a set of extraction routines (with prefix
CVodeGet) to extract statistics after return from the main solver routine.
Similarly, each linear solver module provides its own set of Set- and
Get-type routines. For more details see
§5.4.1.5.10 and
§5.4.1.5.12.
Additionally, the interfaces to several user-supplied routines (such as those
providing Jacobians, preconditioner information, and sensitivity right hand
sides) were simplified by reducing the number of arguments. The same information
that was previously accessible through such arguments can now be obtained
through Get-type functions.
The rootfinding feature was added, whereby the roots of a set of given functions may be computed during the integration of the ODE system.
Installation of CVODES (and all of SUNDIALS) has been completely redesigned and is now based on configure scripts.
5.1.3. Reading this User Guide
This user guide is a combination of general usage instructions. Specific example programs are provided as a separate document. We expect that some readers will want to concentrate on the general instructions, while others will refer mostly to the examples, and the organization is intended to accommodate both styles.
There are different possible levels of usage of CVODES. The most casual user, with a small IVP problem only, can get by with reading §5.2.1, then Chapter §5.4.1 up to §5.4.2 only, and looking at examples in [97]. In addition, to solve a forward sensitivity problem the user should read §5.2.7, followed by Chapter §5.4.3 and look at examples in [97].
In a different direction, a more expert user with an IVP problem may want to (a)
use a package preconditioner (§5.4.2.7), (b) supply
his/her own Jacobian or preconditioner routines
(§5.4.1.6), (c) do multiple runs of problems of
the same size (CVodeReInit()), (d) supply a new N_Vector module
(§9), or even (e) supply new SUNLinearSolver and/or
SUNMatrix modules (Chapters §10 and §11). An
advanced user with a forward sensitivity problem may also want to (a) provide
his/her own sensitivity equations right-hand side routine
§5.4.3.3, (b) perform multiple runs with the
same number of sensitivity parameters
(§5.4.3.2.1, or (c) extract additional
diagnostic information
(§5.4.3.2.7). A user with an
adjoint sensitivity problem needs to understand the IVP solution approach at the
desired level and also go through §5.2.8 for a short
mathematical description of the adjoint approach, Chapter
§5.4.4 for the usage of the adjoint module in CVODES, and
the examples in [97].
The structure of this document is as follows:
In Chapter §5.2, we give short descriptions of the numerical methods implemented by CVODES for the solution of initial value problems for systems of ODEs, continue with short descriptions of preconditioning §5.2.3, stability limit detection (§5.2.4), and rootfinding (§5.2.5), and conclude with an overview of the mathematical aspects of sensitivity analysis, both forward (§5.2.7) and adjoint (§5.2.8).
The following chapter describes the structure of the SUNDIALS suite of solvers (§5.3) and the software organization of the CVODES solver (§5.3.1).
Chapter §5.4.1 is the main usage document for CVODES for simulation applications. It includes a complete description of the user interface for the integration of ODE initial value problems. Readers that are not interested in using CVODES for sensitivity analysis can then skip the next two chapters.
Chapter §5.4.3 describes the usage of CVODES for forward sensitivity analysis as an extension of its IVP integration capabilities. We begin with a skeleton of the user main program, with emphasis on the steps that are required in addition to those already described in Chapter §5.4.1. Following that we provide detailed descriptions of the user-callable interface routines specific to forward sensitivity analysis and of the additonal optional user-defined routines.
Chapter §5.4.4 describes the usage of CVODES for adjoint sensitivity analysis. We begin by describing the CVODES checkpointing implementation for interpolation of the original IVP solution during integration of the adjoint system backward in time, and with an overview of a user’s main program. Following that we provide complete descriptions of the user-callable interface routines for adjoint sensitivity analysis as well as descriptions of the required additional user-defined routines.
Chapter §9 gives a brief overview of the generic
N_Vectormodule shared among the various components of SUNDIALS, and details on theN_Vectorimplementations provided with SUNDIALS.Chapter §10 gives a brief overview of the generic
SUNMatrixmodule shared among the various components of SUNDIALS, and details on theSUNMatriximplementations provided with SUNDIALS: a dense implementation (§§10.9), a banded implementation (§§10.12) and a sparse implementation (§§10.14).Chapter §11 gives a brief overview of the generic
SUNLinearSolvermodule shared among the various components of SUNDIALS. This chapter contains details on theSUNLinearSolverimplementations provided with SUNDIALS. The chapter also contains details on theSUNLinearSolverimplementations provided with SUNDIALS that interface with external linear solver libraries.Finally, in the appendices, we provide detailed instructions for the installation of CVODES, within the structure of SUNDIALS (Appendix §14), as well as a list of all the constants used for input to and output from CVODES functions (Appendix §5.5).
Finally, the reader should be aware of the following notational conventions in
this user guide: program listings and identifiers (such as CVodeInit) within
textual explanations appear in typewriter type style; fields in C structures
(such as content) appear in italics; and packages or modules, such as CVDLS,
are written in all capitals.
Warning
Usage and installation instructions that constitute important warnings are marked in yellow boxes like this one.
5.1.4. SUNDIALS License and Notices
All SUNDIALS packages are released open source, under the BSD 3-Clause license for more details see the LICENSE and NOTICE files provided with all SUNDIALS packages.