Gromacs  2020.4
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Macros Groups Pages
Classes | Typedefs | Enumerations | Functions | Variables
gmx::test Namespace Reference

Description

Testing utilities namespace.

This namespace contains utilities for writing unit tests, mostly from the Testing Utilities (testutils) module.

Classes

class  AnalysisDataTestInputPointSet
 Represents a single set of points in AnalysisDataTestInputFrame structure. More...
 
class  AnalysisDataTestInputFrame
 Represents a single frame in AnalysisDataTestInput structure. More...
 
class  AnalysisDataTestInput
 Represents static input data for AbstractAnalysisData tests. More...
 
class  AnalysisDataTestFixture
 Test fixture for AbstractAnalysisData testing. More...
 
class  ModuleSelection
 Helper class for tests that need an initialized selection. More...
 
class  ModuleTest
 Test fixture to test matching file types for modules. More...
 
class  SetAtomsSupportedFiles
 Helper to test supported file names. More...
 
class  SetAtomsUnSupportedFiles
 Helper to test supported file names. More...
 
class  AnyOutputSupportedFiles
 Helper to test supported file names. More...
 
class  NoOptionalOutput
 Helper to test support for disabling all output options. More...
 
class  OutputSelectorDeathTest
 Helper to test that invalid selection is rejected. More...
 
class  SetVelocitySupportedFiles
 Helper to test supported file names. More...
 
class  SetVelocityUnSupportedFiles
 Helper to test supported file names. More...
 
class  SetForceSupportedFiles
 Helper to test supported file names. More...
 
class  SetForceUnSupportedFiles
 Helper to test supported file names. More...
 
class  SetPrecisionSupportedFiles
 Helper to test supported file names. More...
 
class  SetPrecisionUnSupportedFiles
 Helper to test supported file names. More...
 
class  FlagTest
 Test fixture to test user inputs. More...
 
class  AllocatorTest
 Templated test fixture. More...
 
class  CommandLine
 Helper class for tests that need to construct command lines. More...
 
class  CommandLineTestHelper
 Helper class for tests that construct command lines that need to reference existing files. More...
 
class  CommandLineTestBase
 Test fixture for tests that call a single command-line program with input/output files. More...
 
class  ConfMatch
 Match the contents as an gro file. More...
 
class  IFileMatcher
 Represents a file matcher, matching file contents against reference (or other) data. More...
 
class  IFileMatcherSettings
 Represents a factory for creating a file matcher. More...
 
class  TextFileMatch
 Use a ITextBlockMatcher for matching the contents. More...
 
class  NoContentsMatch
 Do not check the contents of the file. More...
 
class  InteractiveTestHelper
 Helper class for testing interactive sessions. More...
 
class  LoggerTestHelper
 Helper class for tests to check output written to a logger. More...
 
class  TestReferenceChecker
 Handles comparison to test reference data. More...
 
class  TestReferenceData
 Handles creation of and comparison to test reference data. More...
 
class  StdioTestHelper
 Helper class for tests where code reads directly from stdin. More...
 
class  StringTestBase
 Test fixture for tests that check string formatting. More...
 
class  FloatingPointDifference
 Computes and represents a floating-point difference value. More...
 
class  FloatingPointTolerance
 Specifies a floating-point comparison tolerance and checks whether a difference is within the tolerance. More...
 
class  TestException
 Exception class for reporting errors in tests. More...
 
class  TestFileManager
 Helper for tests that need input and output files. More...
 
class  TestFileInputRedirector
 In-memory implementation for IFileInputRedirector for tests. More...
 
class  TestFileOutputRedirector
 In-memory implementation of IFileOutputRedirector for tests. More...
 
class  TestOptionsProvider
 Provides additional options for the test executable. More...
 
class  ITextBlockMatcher
 Represents a text matcher, matching text stream contents against reference data. More...
 
class  ITextBlockMatcherSettings
 Represents a factory for creating a text matcher. More...
 
class  ExactTextMatch
 Use an exact text match (the contents should be exactly equal). More...
 
class  NoTextMatch
 Do not match the text (the contents are ignored). More...
 
class  FilteringExactTextMatch
 Use an exact text match after scrubbing lines of the text that match the supplied regular expressions. More...
 
class  TprAndFileManager
 Helper to bundle generated TPR and the file manager to clean it up. More...
 
class  XvgMatch
 Match the contents as an xvg file. More...
 

Typedefs

typedef std::shared_ptr
< MockAnalysisDataModule > 
MockAnalysisDataModulePointer
 Smart pointer to manage an MockAnalysisDataModule object.
 
typedef std::tuple< int, int,
BiasParams::DisableUpdateSkips > 
BiasTestParameters
 Convenience typedef: growth type enum, potential type enum, disable update skips.
 
typedef gmx::unique_cptr
< gmx_pme_t, gmx_pme_destroy
PmeSafePointer
 A safe pointer type for PME.
 
typedef ArrayRef< const realChargesVector
 Charges.
 
typedef std::vector< RVecCoordinatesVector
 Coordinates.
 
typedef ArrayRef< RVecForcesVector
 Forces.
 
typedef ArrayRef< const IVecGridLineIndicesVector
 Gridline indices.
 
typedef ArrayRef< const realSplineParamsDimVector
 Spline parameters (theta or dtheta). A reference to a single dimension's spline data; this means (atomCount * pmeOrder) values or derivatives.
 
typedef std::array
< SplineParamsDimVector, 3 > 
SplineParamsVector
 Spline parameters (theta or dtheta) in all 3 dimensions.
 
template<typename ValueType >
using SparseGridValuesInput = std::map< IVec, ValueType >
 Non-zero grid values for test input; keys are 3d indices (IVec)
 
typedef SparseGridValuesInput
< real
SparseRealGridValuesInput
 Non-zero real grid values.
 
typedef SparseGridValuesInput
< t_complex > 
SparseComplexGridValuesInput
 Non-zero complex grid values.
 
template<typename ValueType >
using SparseGridValuesOutput = std::map< std::string, ValueType >
 Non-zero grid values for test output; keys are string representations of the cells' 3d indices (IVec); this allows for better sorting.
 
typedef SparseGridValuesOutput
< real
SparseRealGridValuesOutput
 Non-zero real grid values.
 
typedef SparseGridValuesOutput
< t_complex > 
SparseComplexGridValuesOutput
 Non-zero complex grid values.
 
typedef std::array< real, 3 *3 > Matrix3x3
 TODO: make proper C++ matrix for the whole Gromacs, get rid of this.
 
typedef std::vector
< std::unique_ptr
< TestHardwareContext > > 
TestHardwareContexts
 A container of handles to hardware contexts.
 
typedef ::testing::Types
< int32_t, real, RVec,
test::MoveOnly > 
TestTypes
 The types used in testing of all operations.
 
using TestTypesCopyable = ::testing::Types< int32_t, real, RVec >
 The types used in testing minus move only types.
 
template<typename T >
using HostAllocatorTestCopyable = HostAllocatorTest< T >
 Typed test fixture for tests requiring a copyable type.
 
using AllocatorTypesToTest = ::testing::Types< HostAllocator< real >, HostAllocator< int32_t >, HostAllocator< RVec >, HostAllocator< MoveOnly >>
 Declare allocator types to test.
 
using Implementations = ::testing::Types< std::allocator< int32_t >, std::allocator< float >, std::allocator< double >, std::allocator< BasicVector< float >>, std::allocator< BasicVector< double >>, AlignedAllocator< int32_t >, AlignedAllocator< float >, AlignedAllocator< double >, AlignedAllocator< BasicVector< float >>, AlignedAllocator< BasicVector< double >>>
 The types used in testing.
 
typedef
gmx::test::DispersionCorrectionTestFixture 
DispersionCorrectionTest
 Test fixture for mdrun with dispersion correction.
 
using EnergyTermsToCompare = std::unordered_map< std::string, FloatingPointTolerance >
 Convenience type.
 
typedef std::unique_ptr
< EnergyFrameReader > 
EnergyFrameReaderPtr
 Convenience smart pointer typedef.
 
typedef unique_cptr< ener_file,
done_ener_file > 
ener_file_ptr
 Convenience smart pointer typedef.
 
typedef unique_cptr
< t_enxframe, done_enxframe
enxframe_ptr
 Convenience smart pointer typedef.
 
typedef gmx::test::ImdTestFixture ImdTest
 Test fixture for mdrun with IMD settings.
 
typedef MultiSimTest MultiSimTerminationTest
 Convenience typedef.
 
typedef std::unique_ptr
< CommandLine
CommandLinePointer
 Convenience typedef.
 
typedef
gmx::test::OriresTestFixture 
OriresTest
 Test fixture for mdrun with orires.
 
typedef MultiSimTest ReplicaExchangeEnsembleTest
 Convenience typedef.
 
typedef MultiSimTest ReplicaExchangeTerminationTest
 Convenience typedef.
 
typedef gmx::test::SwapTestFixture CompelTest
 Test fixture for mdrun with "Computational Electrophysiology" settings,.
 
typedef MdrunTestFixture MdrunTerminationTest
 Convenience typedef.
 
typedef unique_cptr
< gmx_output_env_t,
output_env_done > 
oenv_ptr
 Convenience smart pointer typedef.
 
typedef unique_cptr
< t_trxstatus, close_trx > 
trxstatus_file_ptr
 Convenience smart pointer typedef.
 
typedef unique_cptr
< t_trxframe, done_trxframe
trxframe_ptr
 Convenience smart pointer typedef.
 
typedef std::unique_ptr
< IFileMatcher
FileMatcherPointer
 Smart pointer for managing a IFileMatcher.
 
using MdpFieldValues = std::map< std::string, std::string >
 Helper typedef.
 
typedef std::unique_ptr
< ITextBlockMatcher
TextBlockMatcherPointer
 Smart pointer for managing a ITextBlockMatcher.
 

Enumerations

enum  TestEnums { efTestString, efTestInt, efTestFloat }
 Test enums to decide on how to use options framework.
 
enum  PmeSolveAlgorithm { Coulomb, LennardJones }
 PME solver type.
 
enum  CodePath { CPU, GPU }
 Hardware code path being tested.
 
enum  ComparisonConditions : int {
  CompareIfBothFound, NoComparison, MustCompare, CompareIfReferenceFound,
  CompareIfTestFound, Count
}
 Enumeration controlling how data within trajectory frames are compared.
 
enum  ReferenceDataMode { erefdataCompare, erefdataCreateMissing, erefdataUpdateChanged, erefdataUpdateAll }
 Mode of operation for reference data handling. More...
 

Functions

static AwhTestParameters getAwhTestParameters (int eawhgrowth, int eawhpotential)
 Helper function to set up the C-style AWH parameters for the test.
 
static AwhTestParameters getAwhTestParameters ()
 Helper function to set up the C-style AWH parameters for the test.
 
TrajectoryFrameWriterPointer createMinimalTrajectoryFrameWriter (const std::string &filename, const TopologyInformation &topology, const Selection &selection, OutputRequirements requirements)
 Create minimal TrajectoryFrameWriter using the provided builder. More...
 
static void splitAndAddValues (OptionsAssigner *assigner, const std::string &optionValues, TestEnums type)
 Helper function to split string with multiple values into option settings. More...
 
bool pmeSupportsInputForMode (const gmx_hw_info_t &hwinfo, const t_inputrec *inputRec, CodePath mode)
 Tells if this generally valid PME input is supported for this mode.
 
uint64_t getSplineModuliDoublePrecisionUlps (int splineOrder)
 For double precision checks, the recursive interpolation and use of trig functions in make_dft_mod require a lot more flops, and thus opportunity for deviation between implementations.
 
PmeSafePointer pmeInitWrapper (const t_inputrec *inputRec, const CodePath mode, const gmx_device_info_t *gpuInfo, PmeGpuProgramHandle pmeGpuProgram, const Matrix3x3 &box, const real ewaldCoeff_q, const real ewaldCoeff_lj)
 PME initialization.
 
PmeSafePointer pmeInitEmpty (const t_inputrec *inputRec, const CodePath mode, const gmx_device_info_t *gpuInfo, PmeGpuProgramHandle pmeGpuProgram, const Matrix3x3 &box, real ewaldCoeff_q, real ewaldCoeff_lj)
 Simple PME initialization based on input, no atom data. More...
 
std::unique_ptr
< StatePropagatorDataGpu > 
makeStatePropagatorDataGpu (const gmx_pme_t &pme)
 Make a GPU state-propagator manager.
 
void pmeInitAtoms (gmx_pme_t *pme, StatePropagatorDataGpu *stateGpu, const CodePath mode, const CoordinatesVector &coordinates, const ChargesVector &charges)
 PME initialization with atom data. More...
 
static realpmeGetRealGridInternal (const gmx_pme_t *pme)
 Getting local PME real grid pointer for test I/O.
 
static void pmeGetRealGridSizesInternal (const gmx_pme_t *pme, CodePath mode, IVec &gridSize, IVec &paddedGridSize)
 Getting local PME real grid dimensions.
 
static t_complex * pmeGetComplexGridInternal (const gmx_pme_t *pme)
 Getting local PME complex grid pointer for test I/O.
 
static void pmeGetComplexGridSizesInternal (const gmx_pme_t *pme, IVec &gridSize, IVec &paddedGridSize)
 Getting local PME complex grid dimensions.
 
template<typename ValueType >
static void pmeGetGridAndSizesInternal (const gmx_pme_t *, CodePath, ValueType *&, IVec &, IVec &)
 Getting the PME grid memory buffer and its sizes - template definition.
 
template<>
void pmeGetGridAndSizesInternal< real > (const gmx_pme_t *pme, CodePath mode, real *&grid, IVec &gridSize, IVec &paddedGridSize)
 Getting the PME real grid memory buffer and its sizes.
 
template<>
void pmeGetGridAndSizesInternal< t_complex > (const gmx_pme_t *pme, CodePath, t_complex *&grid, IVec &gridSize, IVec &paddedGridSize)
 Getting the PME complex grid memory buffer and its sizes.
 
void pmePerformSplineAndSpread (gmx_pme_t *pme, CodePath mode, bool computeSplines, bool spreadCharges)
 PME spline calculation and charge spreading. More...
 
static realpmeGetSplineDataInternal (const gmx_pme_t *pme, PmeSplineDataType type, int dimIndex)
 Getting the internal spline data buffer pointer.
 
void pmePerformSolve (const gmx_pme_t *pme, CodePath mode, PmeSolveAlgorithm method, real cellVolume, GridOrdering gridOrdering, bool computeEnergyAndVirial)
 PME solving.
 
void pmePerformGather (gmx_pme_t *pme, CodePath mode, PmeForceOutputHandling inputTreatment, ForcesVector &forces)
 PME force gathering.
 
void pmeFinalizeTest (const gmx_pme_t *pme, CodePath mode)
 PME test finalization before fetching the outputs.
 
void pmeSetSplineData (const gmx_pme_t *pme, CodePath mode, const SplineParamsDimVector &splineValues, PmeSplineDataType type, int dimIndex)
 Setting atom spline values/derivatives to be used in spread/gather. More...
 
void pmeSetGridLineIndices (gmx_pme_t *pme, CodePath mode, const GridLineIndicesVector &gridLineIndices)
 Setting gridline indices to be used in spread/gather. More...
 
size_t pmeGetGridPlainIndexInternal (const IVec &index, const IVec &paddedGridSize, GridOrdering gridOrdering)
 Getting plain index into the complex 3d grid.
 
template<typename ValueType >
static void pmeSetGridInternal (const gmx_pme_t *pme, CodePath mode, GridOrdering gridOrdering, const SparseGridValuesInput< ValueType > &gridValues)
 Setting real or complex grid.
 
void pmeSetRealGrid (const gmx_pme_t *pme, CodePath mode, const SparseRealGridValuesInput &gridValues)
 Setting real grid to be used in gather.
 
void pmeSetComplexGrid (const gmx_pme_t *pme, CodePath mode, GridOrdering gridOrdering, const SparseComplexGridValuesInput &gridValues)
 Setting complex grid to be used in solve.
 
SplineParamsDimVector pmeGetSplineData (const gmx_pme_t *pme, CodePath mode, PmeSplineDataType type, int dimIndex)
 Getting the single dimension's spline values or derivatives.
 
GridLineIndicesVector pmeGetGridlineIndices (const gmx_pme_t *pme, CodePath mode)
 Getting the gridline indices.
 
template<typename ValueType >
static SparseGridValuesOutput
< ValueType > 
pmeGetGridInternal (const gmx_pme_t *pme, CodePath mode, GridOrdering gridOrdering)
 Getting real or complex grid - only non zero values.
 
SparseRealGridValuesOutput pmeGetRealGrid (const gmx_pme_t *pme, CodePath mode)
 Getting the real grid (spreading output of pmePerformSplineAndSpread())
 
SparseComplexGridValuesOutput pmeGetComplexGrid (const gmx_pme_t *pme, CodePath mode, GridOrdering gridOrdering)
 Getting the complex grid output of pmePerformSolve()
 
PmeOutput pmeGetReciprocalEnergyAndVirial (const gmx_pme_t *pme, CodePath mode, PmeSolveAlgorithm method)
 Getting the reciprocal energy and virial.
 
const char * codePathToString (CodePath codePath)
 Return a string useful for human-readable messages describing a codePath.
 
const PmeTestEnvironment * getPmeTestEnv ()
 Get the test environment.
 
void callAddGlobalTestEnvironment ()
 This constructs the test environment during setup of the unit test so that they can use the hardware context. More...
 
static gmx_hw_info_t * hardwareInit ()
 Simple hardware initialization.
 
template<typename T >
ArrayRef< char > charArrayRefFromArray (T *data, size_t size)
 Convenience function to transform a view into one with base type of (non-const) char. More...
 
template<typename T >
void runTest (const gmx_gpu_info_t &gpuInfo, ArrayRef< T > input, ArrayRef< T > output)
 Does a device transfer of input to the device in gpuInfo, and back to output.
 
template<typename VectorType >
bool isPinned (const VectorType &v)
 Helper function for wrapping a call to isHostMemoryPinned.
 
template<typename T >
void fillInputContents (ArrayRef< T > inputRef, int scaleFactor)
 Initialization overload for non-BasicVector.
 
template<typename T >
void fillInputContents (ArrayRef< BasicVector< T >> inputRef, int scaleFactor)
 Initialization overload for BasicVector.
 
template<typename PaddedVectorOfT >
void fillInput (PaddedVectorOfT *input, int scaleFactor)
 Dispatcher function for filling.
 
template<typename T >
void compareViews (ArrayRef< T > input, ArrayRef< T > output)
 Comparison overload for non-BasicVector.
 
template<typename T >
void compareViews (ArrayRef< BasicVector< T >> input, ArrayRef< BasicVector< T >> output)
 Comparison overload for BasicVector<T>
 
void applyShake (ConstraintsTestData *testData, t_pbc pbc)
 Initialize and apply SHAKE constraints. More...
 
void applyLincs (ConstraintsTestData *testData, t_pbc pbc)
 Initialize and apply LINCS constraints. More...
 
void applyLincsCuda (ConstraintsTestData *testData, t_pbc pbc)
 Apply CUDA version of LINCS constraints to the test data. More...
 
void integrateLeapFrogSimple (LeapFrogTestData *testData, int numSteps)
 Integrate using CPU version of Leap-Frog. More...
 
void integrateLeapFrogGpu (LeapFrogTestData *testData, int numSteps)
 Integrate using CUDA version of Leap-Frog. More...
 
void applySettle (SettleTestData *testData, t_pbc pbc, bool updateVelocities, bool calcVirial, const std::string &testDescription)
 Apply SETTLE using CPU version of the algorithm. More...
 
void applySettleGpu (SettleTestData *testData, t_pbc pbc, bool updateVelocities, bool calcVirial, const std::string &testDescription)
 Apply SETTLE using GPU version of the algorithm. More...
 
static const std::array
< gmx::RVec, 51 > 
c_waterPositions ({{{.130,-.041,-.291},{.120,-.056,-.192},{.044,-.005,-.327},{-.854,-.406,.477},{-.900,-.334,.425},{-.858,-.386,.575},{.351,-.061,.853},{.401,-.147,.859},{.416,.016,.850},{-.067,-.796,.873},{-.129,-.811,.797},{-.119,-.785,.958},{-.635,-.312,-.356},{-.629,-.389,-.292},{-.687,-.338,-.436},{.321,-.919,.242},{.403,-.880,.200},{.294,-1.001,.193},{-.404,.735,.728},{-.409,.670,.803},{-.324,.794,.741},{.461,-.596,-.135},{.411,-.595,-.221},{.398,-.614,-.059},{-.751,-.086,.237},{-.811,-.148,.287},{-.720,-.130,.152},{.202,.285,-.364},{.122,.345,-.377},{.192,.236,-.278},{-.230,-.485,.081},{-.262,-.391,.071},{-.306,-.548,.069},{.464,-.119,.323},{.497,-.080,.409},{.540,-.126,.258},{-.462,.107,.426},{-.486,.070,.336},{-.363,.123,.430},{.249,-.077,-.621},{.306,-.142,-.571},{.233,-.110,-.714},{-.922,-.164,.904},{-.842,-.221,.925},{-.971,-.204,.827},{.382,.700,.480},{.427,.610,.477},{.288,.689,.513},{.781,.264,-.113},{.848,.203,-.070},{.708,.283,-.048}}})
 Database of 51 water atom input positions (taken from spc216.gro) for use as test inputs.
 
static void readTprInput (const TprAndFileManager *tprHandle, gmx_mtop_t *mtop, t_inputrec *ir)
 Reads a tpr for the test. More...
 
void checkEnergiesAgainstReferenceData (const std::string &energyFilename, const EnergyTermsToCompare &energyTermsToCompare, TestReferenceChecker *checker)
 Check a subset of the energies found in an energy file against reference data. More...
 
EnergyFrameReaderPtr openEnergyFileToReadTerms (const std::string &filename, const std::vector< std::string > &requiredEnergyTermNames)
 Open the file and return an object that can read the required terms from frames of an .edr file. More...
 
static t_enxframe * make_enxframe ()
 Helper function to obtain resources.
 
void done_enxframe (t_enxframe *fr)
 Helper function to clean up resources. More...
 
template<typename... Args>
void executeSimulatorComparisonTest (const std::string &environmentVariable, Args &&...args)
 Run and compare a simulator run with and without an environment variable. More...
 
template<typename... Args>
void executeRerunTest (Args &&...args)
 Run and compare a simulator run to its rerun. More...
 
static void organizeMdpFile (SimulationRunner *runner, int nsteps=2)
 Build a simple .mdp file.
 
static void compareBox (const TrajectoryFrame &reference, const TrajectoryFrame &test, const TrajectoryFrameMatchSettings &matchSettings, const FloatingPointTolerance tolerance)
 Compares the box from reference and test according to the matchSettings and tolerance. More...
 
static std::vector< RVecputAtomsInBox (const TrajectoryFrame &frame)
 Help put all atom coordinates in frame into its box. More...
 
static bool shouldDoComparison (const ComparisonConditions comparisonConditions, const bool referenceIsEmpty, const bool testIsEmpty)
 Return whether the comparisonConditions and emptiness of reference and test frames means that a comparison should be attempted. More...
 
static void compareCoordinates (const TrajectoryFrame &reference, const TrajectoryFrame &test, const TrajectoryFrameMatchSettings &matchSettings, const FloatingPointTolerance tolerance)
 Compares the position coordinates from reference and test according to the matchSettings and tolerance.
 
static void compareVelocities (const TrajectoryFrame &reference, const TrajectoryFrame &test, const TrajectoryFrameMatchSettings &matchSettings, const FloatingPointTolerance tolerance)
 Compares the velocities from reference and test according to the matchSettings and tolerance.
 
static void compareForces (const TrajectoryFrame &reference, const TrajectoryFrame &test, const TrajectoryFrameMatchSettings &matchSettings, const FloatingPointTolerance tolerance)
 Compares the forces from reference and test according to the matchSettings and tolerance.
 
static t_trxframe * make_trxframe ()
 Helper function to obtain resources.
 
void done_trxframe (t_trxframe *fr)
 Helper function to clean up resources. More...
 
void checkConfFile (TextInputStream *input, TestReferenceChecker *checker, const ConfMatchSettings &settings)
 Adds content of a gro file to TestReferenceChecker object. More...
 
int getNumberOfTestMpiRanks ()
 Returns the number of MPI ranks to use for an MPI test. More...
 
void initReferenceData (IOptionsContainer *options)
 Initializes reference data handling. More...
 
static void throwIfNonEmptyAndOnlyWhitespace (const std::string &s, const char *id)
 Throw a TestException if the caller tries to write particular refdata that can't work. More...
 
static double convertDoubleReferenceValue (const std::string &value)
 Helper function to parse a floating-point reference data value.
 
int gmxSetenv (const char *name, const char *value, int overwrite)
 Workaround to make setenv work on Windows.
 
int gmxUnsetenv (const char *name)
 Workaround to make unsetenv work on Windows.
 
bool isNumberOfPpRanksSupported (const std::string &simulationName, int possibleNumberOfPpRanks)
 Return whether the number of ranks is supported by the simulation simulationName in the database. More...
 
std::string reportNumbersOfPpRanksSupported (const std::string &simulationName)
 Return a string describing the numbers of ranks supported for the simulation simulationName in the database.
 
MdpFieldValues prepareMdpFieldValues (const std::string &simulationName, const std::string &integrator, const std::string &tcoupl, const std::string &pcoupl)
 Set up values for an .mdp file that permits a highly reproducible simulation. More...
 
MdpFieldValues prepareMdpFieldValues (const char *simulationName, const char *integrator, const char *tcoupl, const char *pcoupl)
 Set up values for an .mdp file that permits a highly reproducible simulation. More...
 
std::string prepareMdpFileContents (const MdpFieldValues &mdpFieldValues)
 Make a string containing an .mdp file from the mdpFieldValues. More...
 
testing::Matcher< std::tuple
< float, float > > 
FloatEq (const FloatingPointTolerance &tolerance)
 Make matcher for floats for use with GoogleMock that compare equal when tolerance is satisifed. More...
 
testing::Matcher< std::tuple
< double, double > > 
DoubleEq (const FloatingPointTolerance &tolerance)
 Make matcher for doubles for use with GoogleMock that compare equal when tolerance is satisifed. More...
 
testing::Matcher< std::tuple
< real, real > > 
RealEq (const FloatingPointTolerance &tolerance)
 Make matcher for reals for use with GoogleMock that compare equal when tolerance is satisifed. More...
 
testing::Matcher< std::tuple
< RVec, RVec > > 
RVecEq (const FloatingPointTolerance &tolerance)
 Make matcher for RVecs for use with GoogleMock that compare equal when tolerance is satisifed. More...
 
void registerTestOptions (const char *name, TestOptionsProvider *provider)
 Registers a test option provider with the test framework. More...
 
void initTestOptions (IOptionsContainer *options)
 Initializes the options from all registered test providers. More...
 
void checkXvgFile (TextInputStream *input, TestReferenceChecker *checker, const XvgMatchSettings &settings)
 Adds content of xvg file to TestReferenceChecker object. More...
 

Variables

const double g_coords []
 Database of 21 test coordinates that represent a trajectory */. More...
 
const char *const trajectoryFileNames []
 Character array of different file names to test. More...
 
const char *const setAtomsSupported []
 Names here work for setAtoms module. More...
 
const char *const setAtomsUnSupported [] = { "spc2-traj.trr", "spc2-traj.xtc", "spc2-traj.g96" }
 Names here don't work for setAtoms module.
 
const char *const anySupported []
 Names here work for stuff that has no specific requirements. More...
 
const char *const setVelocitySupported []
 Names here work for setVelocity module. More...
 
const char *const setVelocityUnSupported [] = { "spc2-traj.xtc", "spc2-traj.pdb", "spc2-traj.g96" }
 Names here don't work for setVelocity module.
 
const char *const setForceSupported []
 Names here work for setForce module. More...
 
const char *const setForceUnSupported []
 Names here don't work for setForce module. More...
 
const char *const setPrecisionSupported []
 Names here work for setPrecision module. More...
 
const char *const setPrecisionUnSupported []
 Names here don't work for setPrecision module. More...
 
constexpr int64_t c_splineModuliSinglePrecisionUlps = 1
 Spline moduli are computed in double precision, so they're very good in single precision.
 
static const char * g_butaneTopFileFormatString
 Format string for building a configurable .top file.
 
static bool g_bDeleteFilesAfterTest = true
 Controls whether TestFileManager should delete temporary files after the test finishes. More...
 

Function Documentation

void gmx::test::applyLincs ( ConstraintsTestData *  testData,
t_pbc  pbc 
)

Initialize and apply LINCS constraints.

Apply LINCS constraints to the test data.

Parameters
[in]testDataTest data structure.
[in]pbcPeriodic boundary data.
void gmx::test::applyLincsCuda ( ConstraintsTestData *  testData,
t_pbc  pbc 
)

Apply CUDA version of LINCS constraints to the test data.

All the data is copied to the GPU device, then LINCS is applied and the resulting coordinates are copied back.

void gmx::test::applySettle ( SettleTestData *  testData,
t_pbc  pbc,
bool  updateVelocities,
bool  calcVirial,
const std::string &  testDescription 
)

Apply SETTLE using CPU version of the algorithm.

Initializes SETTLE object, applies algorithm, destroys the object. The coordinates, velocities and virial are updated in the testData object.

Parameters
[in,out]testDataAn object, containing all the data structures needed by SETTLE.
[in]pbcPeriodic boundary setup.
[in]updateVelocitiesIf the velocities should be updated.
[in]calcVirialIf the virial should be computed.
[in]testDescriptionBrief description that will be printed in case of test failure.
void gmx::test::applySettleGpu ( SettleTestData *  testData,
t_pbc  pbc,
bool  updateVelocities,
bool  calcVirial,
const std::string &  testDescription 
)

Apply SETTLE using GPU version of the algorithm.

Initializes SETTLE object, copied data to the GPU, applies algorithm, copies the data back, destroys the object. The coordinates, velocities and virial are updated in the testData object.

Parameters
[in,out]testDataAn object, containing all the data structures needed by SETTLE.
[in]pbcPeriodic boundary setup.
[in]updateVelocitiesIf the velocities should be updated.
[in]calcVirialIf the virial should be computed.
[in]testDescriptionBrief description that will be printed in case of test failure.
void gmx::test::applyShake ( ConstraintsTestData *  testData,
t_pbc  pbc 
)

Initialize and apply SHAKE constraints.

Apply SHAKE constraints to the test data.

Parameters
[in]testDataTest data structure.
[in]pbcPeriodic boundary data.
void gmx::test::callAddGlobalTestEnvironment ( )

This constructs the test environment during setup of the unit test so that they can use the hardware context.

Implement a stub definition for tests that don't ask for a real one.

Declare a function that all unit test implementations can use to set up any environment that they need.

When registering the unit test in CMake, the HARDWARE_DETECTION flag requires that the code for that unit test implements this function. Otherwise, a default stub implementation is provided.

This approach conforms to the recommendation by GoogleTest to arrange for the code that sets up the global test environment to be called from main, rather than potentially rely on brittle static initialization order.

template<typename T >
ArrayRef<char> gmx::test::charArrayRefFromArray ( T *  data,
size_t  size 
)

Convenience function to transform a view into one with base type of (non-const) char.

This transformation is useful for using containers with C APIs where the function signature is not declared const even where the semantics of the usage actually are const.

Parameters
[in]dataThe data pointer.
[in]sizeThe size of the data pointer (in T).
Template Parameters
TThe base type of the container
void gmx::test::checkConfFile ( TextInputStream *  input,
TestReferenceChecker *  checker,
const ConfMatchSettings &  settings 
)

Adds content of a gro file to TestReferenceChecker object.

Parameters
[in]inputStream that provides the gro content.
[in,out]checkerChecker to use.
[in]settingsSettings to use for matching.

Parses a gro file from the input stream, and checks the contents against reference data (only first two lines for now).

See Also
ConfMatch
void gmx::test::checkEnergiesAgainstReferenceData ( const std::string &  energyFilename,
const EnergyTermsToCompare &  energyTermsToCompare,
TestReferenceChecker *  checker 
)

Check a subset of the energies found in an energy file against reference data.

Opens the energy file, loops over all frames, matching the indicated energies against refdata at the given tolerance.

Parameters
[in]energyFilenameThe name of an energy file.
[in]energyTermsToCompareSet of energies to match at given tolerances.
[in]checkerRoot checker for reference data.
Todo:
This is quite similar to the functionality used in PmeTest, and we should consider reducing the duplication.
void gmx::test::checkXvgFile ( TextInputStream *  input,
TestReferenceChecker *  checker,
const XvgMatchSettings &  settings 
)

Adds content of xvg file to TestReferenceChecker object.

Parameters
[in]inputStream that provides the xvg content.
[in,out]checkerChecker to use.
[in]settingsSettings to use for matching.

Parses an xvg file from the input stream, and checks the contents against reference data. settings can be used to customize the matching. Only a single data set is supported (but multiple columns work). A subset of xmgrace formatting is also checked; static content that is nearly always the same is skipped.

See Also
XvgMatch
static void gmx::test::compareBox ( const TrajectoryFrame &  reference,
const TrajectoryFrame &  test,
const TrajectoryFrameMatchSettings &  matchSettings,
const FloatingPointTolerance  tolerance 
)
static

Compares the box from reference and test according to the matchSettings and tolerance.

Todo:
This could be streamlined when we have a proper 3D matrix class and view.
TrajectoryFrameWriterPointer gmx::test::createMinimalTrajectoryFrameWriter ( const std::string &  filename,
const TopologyInformation &  topology,
const Selection &  selection,
OutputRequirements  requirements 
)
inline

Create minimal TrajectoryFrameWriter using the provided builder.

Parameters
[in]filenameName of file to create object for.
[in]topologyReference to input top.
[in]selectionReference to input selection.
[in]requirementsRequirements for constructing OutputManagar.
Exceptions
InconsistentInputErrorWhen builder can not create the CoordinateFile.
Returns
unique_ptr to new CoordinateFile object.
void gmx::test::done_enxframe ( t_enxframe *  fr)

Helper function to clean up resources.

Helper function to free resources (NB free_enxframe only frees the contents, not the pointer itself)

void gmx::test::done_trxframe ( t_trxframe *  fr)

Helper function to clean up resources.

Helper function to free all resources.

testing::Matcher< std::tuple< double, double > > gmx::test::DoubleEq ( const FloatingPointTolerance &  tolerance)

Make matcher for doubles for use with GoogleMock that compare equal when tolerance is satisifed.

Used like

EXPECT_THAT(testDoubles, Pointwise(DoubleEq(tolerance), referenceDoubles));

template<typename... Args>
void gmx::test::executeRerunTest ( Args &&...  args)

Run and compare a simulator run to its rerun.

Run grompp, run mdrun and rerun the resulting trajectory. Compare energies (via EnergyComparator) and trajectories.

template<typename... Args>
void gmx::test::executeSimulatorComparisonTest ( const std::string &  environmentVariable,
Args &&...  args 
)

Run and compare a simulator run with and without an environment variable.

Run grompp, and repeat mdrun with and without the environment variable set. Compare energies (via EnergyComparator) and trajectories.

testing::Matcher< std::tuple< float, float > > gmx::test::FloatEq ( const FloatingPointTolerance &  tolerance)

Make matcher for floats for use with GoogleMock that compare equal when tolerance is satisifed.

Used like

EXPECT_THAT(testFloats, Pointwise(FloatEq(tolerance), referenceFloats));

void gmx::test::integrateLeapFrogGpu ( LeapFrogTestData *  testData,
int  numSteps 
)

Integrate using CUDA version of Leap-Frog.

Copies data from CPU to GPU, integrates the equation of motion for requested number of steps using Leap-Frog algorithm, copies the result back.

Parameters
[in]testDataData needed for the integrator
[in]numStepsTotal number of steps to run integration for.
void gmx::test::integrateLeapFrogSimple ( LeapFrogTestData *  testData,
int  numSteps 
)

Integrate using CPU version of Leap-Frog.

Parameters
[in]testDataData needed for the integrator
[in]numStepsTotal number of steps to run integration for.
bool gmx::test::isNumberOfPpRanksSupported ( const std::string &  simulationName,
int  possibleNumberOfPpRanks 
)

Return whether the number of ranks is supported by the simulation simulationName in the database.

This method lets test runners understand when end-to-end tests should be expected to work.

EnergyFrameReaderPtr gmx::test::openEnergyFileToReadTerms ( const std::string &  filename,
const std::vector< std::string > &  requiredEnergyTermNames 
)

Open the file and return an object that can read the required terms from frames of an .edr file.

Parameters
[in]filenameName of the energy file to use
[in]requiredEnergyTermNamesNames of the energy terms that the caller requires to be present for an .edr file frame to be considered valid
Exceptions
FileIOErrorIf the .edr file cannot be opened
APIErrorIf any required energy term is not present in the file
std::bad_allocWhen out of memory

This function is intended to have the main responsibility for making EnergyFrameReader objects.

void gmx::test::pmeInitAtoms ( gmx_pme_t *  pme,
StatePropagatorDataGpu *  stateGpu,
const CodePath  mode,
const CoordinatesVector &  coordinates,
const ChargesVector &  charges 
)

PME initialization with atom data.

PME initialization with atom data and system box.

PmeSafePointer gmx::test::pmeInitEmpty ( const t_inputrec *  inputRec,
const CodePath  mode,
const gmx_device_info_t *  gpuInfo,
PmeGpuProgramHandle  pmeGpuProgram,
const Matrix3x3 &  box,
real  ewaldCoeff_q,
real  ewaldCoeff_lj 
)

Simple PME initialization based on input, no atom data.

Simple PME initialization (no atom data)

void gmx::test::pmePerformSplineAndSpread ( gmx_pme_t *  pme,
CodePath  mode,
bool  computeSplines,
bool  spreadCharges 
)

PME spline calculation and charge spreading.

PME spline computation and charge spreading.

void gmx::test::pmeSetGridLineIndices ( gmx_pme_t *  pme,
CodePath  mode,
const GridLineIndicesVector &  gridLineIndices 
)

Setting gridline indices to be used in spread/gather.

Setting gridline indices be used in spread/gather.

void gmx::test::pmeSetSplineData ( const gmx_pme_t *  pme,
CodePath  mode,
const SplineParamsDimVector &  splineValues,
PmeSplineDataType  type,
int  dimIndex 
)

Setting atom spline values/derivatives to be used in spread/gather.

Setting atom spline values or derivatives to be used in spread/gather.

MdpFieldValues gmx::test::prepareMdpFieldValues ( const std::string &  simulationName,
const std::string &  integrator,
const std::string &  tcoupl,
const std::string &  pcoupl 
)

Set up values for an .mdp file that permits a highly reproducible simulation.

An internal database of several kinds of simulation useful for such comparisons is available, whose simulationName keys are

  • argon12
  • argon5832
  • tip3p5
  • spc2
  • spc216
  • alanine_vsite_vacuo
  • alanine_vsite_solvated
  • glycine_vacuo
  • glycine_no_constraints_vacuo
  • nonanol_vacuo

Some of these systems are pretty minimal, because having few atoms means few interactions, highly reproducible forces, and allows tests to focus on the correctness of the implementation of high-level mdrun features. The boxes are of a reasonable size so that domain decomposition is possible. The pressure-coupling parameters are isotropic, and set up so that there will not be dramatic collapse of volume over the handful of MD steps that will be run. A single temperature-coupling group is used.

Parameters
[in]simulationNameThe name of the simulation, which indexes the database
[in]integratorThe integrator to use
[in]tcouplThe temperature-coupling algorithm to use
[in]pcouplThe pressure-coupling algorithm to use
Returns
Mdp file values
Exceptions
std::bad_allocif out of memory std::out_of_range if simulationName is not in the database
MdpFieldValues gmx::test::prepareMdpFieldValues ( const char *  simulationName,
const char *  integrator,
const char *  tcoupl,
const char *  pcoupl 
)

Set up values for an .mdp file that permits a highly reproducible simulation.

An internal database of several kinds of simulation useful for such comparisons is available, whose simulationName keys are

  • argon12
  • argon5832
  • tip3p5
  • spc2
  • spc216
  • alanine_vsite_vacuo
  • alanine_vsite_solvated
  • glycine_vacuo
  • glycine_no_constraints_vacuo
  • nonanol_vacuo

Some of these systems are pretty minimal, because having few atoms means few interactions, highly reproducible forces, and allows tests to focus on the correctness of the implementation of high-level mdrun features. The boxes are of a reasonable size so that domain decomposition is possible. The pressure-coupling parameters are isotropic, and set up so that there will not be dramatic collapse of volume over the handful of MD steps that will be run. A single temperature-coupling group is used.

Parameters
[in]simulationNameThe name of the simulation, which indexes the database
[in]integratorThe integrator to use
[in]tcouplThe temperature-coupling algorithm to use
[in]pcouplThe pressure-coupling algorithm to use
Returns
Mdp file values
Exceptions
std::bad_allocif out of memory std::out_of_range if simulationName is not in the database
std::string gmx::test::prepareMdpFileContents ( const MdpFieldValues &  mdpFieldValues)

Make a string containing an .mdp file from the mdpFieldValues.

Exceptions
std::bad_allocif out of memory
static std::vector<RVec> gmx::test::putAtomsInBox ( const TrajectoryFrame &  frame)
static

Help put all atom coordinates in frame into its box.

This can perhaps go away when frame->x is a container.

static void gmx::test::readTprInput ( const TprAndFileManager *  tprHandle,
gmx_mtop_t *  mtop,
t_inputrec *  ir 
)
static

Reads a tpr for the test.

Reads a tpr to have access to the system information for print out.

Parameters
[in]tprHandleHandle to the tpr to red in.
[in]mtopPointer to topology datastructure to populate.
[in]irPointer to inputrec to populate.
testing::Matcher< std::tuple< real, real > > gmx::test::RealEq ( const FloatingPointTolerance &  tolerance)

Make matcher for reals for use with GoogleMock that compare equal when tolerance is satisifed.

Used like

EXPECT_THAT(testReals, Pointwise(RealEq(tolerance), referenceReals));

testing::Matcher< std::tuple< RVec, RVec > > gmx::test::RVecEq ( const FloatingPointTolerance &  tolerance)

Make matcher for RVecs for use with GoogleMock that compare equal when tolerance is satisifed.

Used like

EXPECT_THAT(testRVecs, Pointwise(RVecEq(tolerance), referenceRVecs));

static bool gmx::test::shouldDoComparison ( const ComparisonConditions  comparisonConditions,
const bool  referenceIsEmpty,
const bool  testIsEmpty 
)
static

Return whether the comparisonConditions and emptiness of reference and test frames means that a comparison should be attempted.

This allows the framework to determine whether it is an error if a comparison cannot be made.

static void gmx::test::splitAndAddValues ( OptionsAssigner *  assigner,
const std::string &  optionValues,
TestEnums  type 
)
static

Helper function to split string with multiple values into option settings.

Parameters
[in]assignerOptions assigner to use.
[in]optionValuesString to split that contains option values.
[in]typeDetermine which input time we are using.
static void gmx::test::throwIfNonEmptyAndOnlyWhitespace ( const std::string &  s,
const char *  id 
)
static

Throw a TestException if the caller tries to write particular refdata that can't work.

If the string to write is non-empty and has only whitespace, TinyXML2 can't read it correctly, so throw an exception for this case, so that we can't accidentally use it and run into mysterious problems.

Todo:
Eliminate this limitation of TinyXML2. See e.g. https://github.com/leethomason/tinyxml2/issues/432

Variable Documentation

const char* const gmx::test::anySupported[]
Initial value:
= { "spc2-traj.trr",
"spc2-traj.xtc", "spc2-traj.g96" }

Names here work for stuff that has no specific requirements.

PDB and GRO format are not tested here because they also require atoms information that is incompatible with the other output formats.

bool gmx::test::g_bDeleteFilesAfterTest = true
static

Controls whether TestFileManager should delete temporary files after the test finishes.

const double gmx::test::g_coords[]
Initial value:
= { 0.62, 0.70, 0.68, 0.80, 0.93, 0.87, 1.16, 1.14, 0.95, 0.89, 0.91,
0.86, 0.88, 0.79, 0.75, 0.82, 0.74, 0.70, 0.68, 0.71, 0.73 }

Database of 21 test coordinates that represent a trajectory */.

const char* const gmx::test::setAtomsSupported[]
Initial value:
= {
"spc2-traj.gro",
"spc2-traj.pdb",
}

Names here work for setAtoms module.

const char* const gmx::test::setForceSupported[]
Initial value:
= {
"spc2-traj.trr",
}

Names here work for setForce module.

const char* const gmx::test::setForceUnSupported[]
Initial value:
= { "spc2-traj.xtc", "spc2-traj.pdb", "spc2-traj.gro",
"spc2-traj.g96" }

Names here don't work for setForce module.

const char* const gmx::test::setPrecisionSupported[]
Initial value:
= {
"spc2-traj.xtc",
}

Names here work for setPrecision module.

const char* const gmx::test::setPrecisionUnSupported[]
Initial value:
= { "spc2-traj.trr", "spc2-traj.pdb", "spc2-traj.gro",
"spc2-traj.g96" }

Names here don't work for setPrecision module.

const char* const gmx::test::setVelocitySupported[]
Initial value:
= {
"spc2-traj.gro",
"spc2-traj.trr",
}

Names here work for setVelocity module.

const char* const gmx::test::trajectoryFileNames[]
Initial value:
= { "spc2-traj.trr",
"spc2-traj.xtc", "spc2-traj.pdb",
"spc2-traj.gro", "spc2-traj.g96" }

Character array of different file names to test.