2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 2016,2017,2018,2019, by the GROMACS development team, led by
5 * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
6 * and including many others, as listed in the AUTHORS file in the
7 * top-level source directory and at http://www.gromacs.org.
9 * GROMACS is free software; you can redistribute it and/or
10 * modify it under the terms of the GNU Lesser General Public License
11 * as published by the Free Software Foundation; either version 2.1
12 * of the License, or (at your option) any later version.
14 * GROMACS is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
17 * Lesser General Public License for more details.
19 * You should have received a copy of the GNU Lesser General Public
20 * License along with GROMACS; if not, see
21 * http://www.gnu.org/licenses, or write to the Free Software Foundation,
22 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
24 * If you want to redistribute modifications to GROMACS, please
25 * consider that scientific software is very special. Version
26 * control is crucial - bugs must be traceable. We will be happy to
27 * consider code for inclusion in the official distribution, but
28 * derived work must not be called official GROMACS. Details are found
29 * in the README & COPYING files - if they are missing, get the
30 * official version at http://www.gromacs.org.
32 * To help us fund GROMACS development, we humbly ask that you cite
33 * the research papers on the package. Check out http://www.gromacs.org.
37 * Implements common routines for PME tests.
39 * \author Aleksei Iupinov <a.yupinov@gmail.com>
40 * \ingroup module_ewald
44 #include "pmetestcommon.h"
50 #include "gromacs/domdec/domdec.h"
51 #include "gromacs/ewald/pme-gather.h"
52 #include "gromacs/ewald/pme-gpu-internal.h"
53 #include "gromacs/ewald/pme-grid.h"
54 #include "gromacs/ewald/pme-internal.h"
55 #include "gromacs/ewald/pme-solve.h"
56 #include "gromacs/ewald/pme-spread.h"
57 #include "gromacs/fft/parallel_3dfft.h"
58 #include "gromacs/gpu_utils/gpu_utils.h"
59 #include "gromacs/math/invertmatrix.h"
60 #include "gromacs/mdtypes/commrec.h"
61 #include "gromacs/pbcutil/pbc.h"
62 #include "gromacs/topology/topology.h"
63 #include "gromacs/utility/exceptions.h"
64 #include "gromacs/utility/gmxassert.h"
65 #include "gromacs/utility/logger.h"
66 #include "gromacs/utility/stringutil.h"
68 #include "testutils/testasserts.h"
75 bool pmeSupportsInputForMode(const gmx_hw_info_t &hwinfo,
76 const t_inputrec *inputRec,
88 implemented = (pme_gpu_supports_build(nullptr) &&
89 pme_gpu_supports_hardware(hwinfo, nullptr) &&
90 pme_gpu_supports_input(*inputRec, mtop, nullptr));
94 GMX_THROW(InternalError("Test not implemented for this mode"));
99 uint64_t getSplineModuliDoublePrecisionUlps(int splineOrder)
101 /* Arbitrary ulp tolerance for sine/cosine implementation. It's
102 * hard to know what to pick without testing lots of
103 * implementations. */
104 const uint64_t sineUlps = 10;
105 return 4 * (splineOrder - 2) + 2 * sineUlps * splineOrder;
108 //! PME initialization - internal
109 static PmeSafePointer pmeInitInternal(const t_inputrec *inputRec,
111 const gmx_device_info_t *gpuInfo,
112 PmeGpuProgramHandle pmeGpuProgram,
114 const Matrix3x3 &box,
115 real ewaldCoeff_q = 1.0f,
116 real ewaldCoeff_lj = 1.0f
119 const MDLogger dummyLogger;
120 const auto runMode = (mode == CodePath::CPU) ? PmeRunMode::CPU : PmeRunMode::Mixed;
121 t_commrec dummyCommrec = {0};
122 NumPmeDomains numPmeDomains = { 1, 1 };
123 gmx_pme_t *pmeDataRaw = gmx_pme_init(&dummyCommrec, numPmeDomains, inputRec, atomCount, false, false, true,
124 ewaldCoeff_q, ewaldCoeff_lj, 1, runMode, nullptr, gpuInfo, pmeGpuProgram, dummyLogger);
125 PmeSafePointer pme(pmeDataRaw); // taking ownership
127 // TODO get rid of this with proper matrix type
129 for (int i = 0; i < DIM; i++)
131 for (int j = 0; j < DIM; j++)
133 boxTemp[i][j] = box[i * DIM + j];
136 const char *boxError = check_box(-1, boxTemp);
137 GMX_RELEASE_ASSERT(boxError == nullptr, boxError);
142 invertBoxMatrix(boxTemp, pme->recipbox);
146 pme_gpu_set_testing(pme->gpu, true);
147 pme_gpu_update_input_box(pme->gpu, boxTemp);
151 GMX_THROW(InternalError("Test not implemented for this mode"));
157 //! Simple PME initialization based on input, no atom data
158 PmeSafePointer pmeInitEmpty(const t_inputrec *inputRec,
160 const gmx_device_info_t *gpuInfo,
161 PmeGpuProgramHandle pmeGpuProgram,
162 const Matrix3x3 &box,
167 return pmeInitInternal(inputRec, mode, gpuInfo, pmeGpuProgram, 0, box, ewaldCoeff_q, ewaldCoeff_lj);
168 // hiding the fact that PME actually needs to know the number of atoms in advance
171 //! PME initialization with atom data
172 PmeSafePointer pmeInitAtoms(const t_inputrec *inputRec,
174 const gmx_device_info_t *gpuInfo,
175 PmeGpuProgramHandle pmeGpuProgram,
176 const CoordinatesVector &coordinates,
177 const ChargesVector &charges,
181 const index atomCount = coordinates.size();
182 GMX_RELEASE_ASSERT(atomCount == charges.size(), "Mismatch in atom data");
183 PmeSafePointer pmeSafe = pmeInitInternal(inputRec, mode, gpuInfo, pmeGpuProgram, atomCount, box);
184 pme_atomcomm_t *atc = nullptr;
189 atc = &(pmeSafe->atc[0]);
190 atc->x = const_cast<rvec *>(as_rvec_array(coordinates.data()));
191 atc->coefficient = const_cast<real *>(charges.data());
192 /* With decomposition there would be more boilerplate atc code here, e.g. do_redist_pos_coeffs */
196 gmx_pme_reinit_atoms(pmeSafe.get(), atomCount, charges.data());
197 pme_gpu_copy_input_coordinates(pmeSafe->gpu, as_rvec_array(coordinates.data()));
201 GMX_THROW(InternalError("Test not implemented for this mode"));
207 //! Getting local PME real grid pointer for test I/O
208 static real *pmeGetRealGridInternal(const gmx_pme_t *pme)
210 const size_t gridIndex = 0;
211 return pme->fftgrid[gridIndex];
214 //! Getting local PME real grid dimensions
215 static void pmeGetRealGridSizesInternal(const gmx_pme_t *pme,
217 IVec &gridSize, //NOLINT(google-runtime-references)
218 IVec &paddedGridSize) //NOLINT(google-runtime-references)
220 const size_t gridIndex = 0;
221 IVec gridOffsetUnused;
225 gmx_parallel_3dfft_real_limits(pme->pfft_setup[gridIndex], gridSize, gridOffsetUnused, paddedGridSize);
229 pme_gpu_get_real_grid_sizes(pme->gpu, &gridSize, &paddedGridSize);
233 GMX_THROW(InternalError("Test not implemented for this mode"));
237 //! Getting local PME complex grid pointer for test I/O
238 static t_complex *pmeGetComplexGridInternal(const gmx_pme_t *pme)
240 const size_t gridIndex = 0;
241 return pme->cfftgrid[gridIndex];
244 //! Getting local PME complex grid dimensions
245 static void pmeGetComplexGridSizesInternal(const gmx_pme_t *pme,
246 IVec &gridSize, //NOLINT(google-runtime-references)
247 IVec &paddedGridSize) //NOLINT(google-runtime-references)
249 const size_t gridIndex = 0;
250 IVec gridOffsetUnused, complexOrderUnused;
251 gmx_parallel_3dfft_complex_limits(pme->pfft_setup[gridIndex], complexOrderUnused, gridSize, gridOffsetUnused, paddedGridSize); //TODO: what about YZX ordering?
254 //! Getting the PME grid memory buffer and its sizes - template definition
255 template<typename ValueType> static void pmeGetGridAndSizesInternal(const gmx_pme_t * /*unused*/, CodePath /*unused*/, ValueType * & /*unused*/, IVec & /*unused*/, IVec & /*unused*/) //NOLINT(google-runtime-references)
257 GMX_THROW(InternalError("Deleted function call"));
258 // explicitly deleting general template does not compile in clang/icc, see https://llvm.org/bugs/show_bug.cgi?id=17537
261 //! Getting the PME real grid memory buffer and its sizes
262 template<> void pmeGetGridAndSizesInternal<real>(const gmx_pme_t *pme, CodePath mode, real * &grid, IVec &gridSize, IVec &paddedGridSize)
264 grid = pmeGetRealGridInternal(pme);
265 pmeGetRealGridSizesInternal(pme, mode, gridSize, paddedGridSize);
268 //! Getting the PME complex grid memory buffer and its sizes
269 template<> void pmeGetGridAndSizesInternal<t_complex>(const gmx_pme_t *pme, CodePath /*unused*/, t_complex * &grid, IVec &gridSize, IVec &paddedGridSize)
271 grid = pmeGetComplexGridInternal(pme);
272 pmeGetComplexGridSizesInternal(pme, gridSize, paddedGridSize);
275 //! PME spline calculation and charge spreading
276 void pmePerformSplineAndSpread(gmx_pme_t *pme, CodePath mode, // TODO const qualifiers elsewhere
277 bool computeSplines, bool spreadCharges)
279 GMX_RELEASE_ASSERT(pme != nullptr, "PME data is not initialized");
280 pme_atomcomm_t *atc = &(pme->atc[0]);
281 const size_t gridIndex = 0;
282 const bool computeSplinesForZeroCharges = true;
283 real *fftgrid = spreadCharges ? pme->fftgrid[gridIndex] : nullptr;
284 real *pmegrid = pme->pmegrid[gridIndex].grid.grid;
289 spread_on_grid(pme, atc, &pme->pmegrid[gridIndex], computeSplines, spreadCharges,
290 fftgrid, computeSplinesForZeroCharges, gridIndex);
291 if (spreadCharges && !pme->bUseThreads)
293 wrap_periodic_pmegrid(pme, pmegrid);
294 copy_pmegrid_to_fftgrid(pme, pmegrid, fftgrid, gridIndex);
299 pme_gpu_spread(pme->gpu, gridIndex, fftgrid, computeSplines, spreadCharges);
303 GMX_THROW(InternalError("Test not implemented for this mode"));
307 //! Getting the internal spline data buffer pointer
308 static real *pmeGetSplineDataInternal(const gmx_pme_t *pme, PmeSplineDataType type, int dimIndex)
310 GMX_ASSERT((0 <= dimIndex) && (dimIndex < DIM), "Invalid dimension index");
311 const pme_atomcomm_t *atc = &(pme->atc[0]);
312 const size_t threadIndex = 0;
313 real *splineBuffer = nullptr;
316 case PmeSplineDataType::Values:
317 splineBuffer = atc->spline[threadIndex].theta[dimIndex];
320 case PmeSplineDataType::Derivatives:
321 splineBuffer = atc->spline[threadIndex].dtheta[dimIndex];
325 GMX_THROW(InternalError("Unknown spline data type"));
331 void pmePerformSolve(const gmx_pme_t *pme, CodePath mode,
332 PmeSolveAlgorithm method, real cellVolume,
333 GridOrdering gridOrdering, bool computeEnergyAndVirial)
335 t_complex *h_grid = pmeGetComplexGridInternal(pme);
336 const bool useLorentzBerthelot = false;
337 const size_t threadIndex = 0;
341 if (gridOrdering != GridOrdering::YZX)
343 GMX_THROW(InternalError("Test not implemented for this mode"));
347 case PmeSolveAlgorithm::Coulomb:
348 solve_pme_yzx(pme, h_grid, cellVolume,
349 computeEnergyAndVirial, pme->nthread, threadIndex);
352 case PmeSolveAlgorithm::LennardJones:
353 solve_pme_lj_yzx(pme, &h_grid, useLorentzBerthelot,
354 cellVolume, computeEnergyAndVirial, pme->nthread, threadIndex);
358 GMX_THROW(InternalError("Test not implemented for this mode"));
365 case PmeSolveAlgorithm::Coulomb:
366 pme_gpu_solve(pme->gpu, h_grid, gridOrdering, computeEnergyAndVirial);
370 GMX_THROW(InternalError("Test not implemented for this mode"));
375 GMX_THROW(InternalError("Test not implemented for this mode"));
379 //! PME force gathering
380 void pmePerformGather(gmx_pme_t *pme, CodePath mode,
381 PmeForceOutputHandling inputTreatment, ForcesVector &forces)
383 pme_atomcomm_t *atc = &(pme->atc[0]);
384 const index atomCount = atc->n;
385 GMX_RELEASE_ASSERT(forces.size() == atomCount, "Invalid force buffer size");
386 const bool forceReductionWithInput = (inputTreatment == PmeForceOutputHandling::ReduceWithInput);
387 const real scale = 1.0;
388 const size_t threadIndex = 0;
389 const size_t gridIndex = 0;
390 real *pmegrid = pme->pmegrid[gridIndex].grid.grid;
391 real *fftgrid = pme->fftgrid[gridIndex];
396 atc->f = as_rvec_array(forces.begin());
397 if (atc->nthread == 1)
399 // something which is normally done in serial spline computation (make_thread_local_ind())
400 atc->spline[threadIndex].n = atomCount;
402 copy_fftgrid_to_pmegrid(pme, fftgrid, pmegrid, gridIndex, pme->nthread, threadIndex);
403 unwrap_periodic_pmegrid(pme, pmegrid);
404 gather_f_bsplines(pme, pmegrid, !forceReductionWithInput, atc, &atc->spline[threadIndex], scale);
409 // Variable initialization needs a non-switch scope
410 PmeOutput output = pme_gpu_getOutput(*pme, GMX_PME_CALC_F);
411 GMX_ASSERT(forces.size() == output.forces_.size(), "Size of force buffers did not match");
412 if (forceReductionWithInput)
414 std::copy(std::begin(forces), std::end(forces), std::begin(output.forces_));
416 pme_gpu_gather(pme->gpu, inputTreatment, reinterpret_cast<float *>(fftgrid));
417 std::copy(std::begin(output.forces_), std::end(output.forces_), std::begin(forces));
422 GMX_THROW(InternalError("Test not implemented for this mode"));
426 //! PME test finalization before fetching the outputs
427 void pmeFinalizeTest(const gmx_pme_t *pme, CodePath mode)
435 pme_gpu_synchronize(pme->gpu);
439 GMX_THROW(InternalError("Test not implemented for this mode"));
443 //! Setting atom spline values/derivatives to be used in spread/gather
444 void pmeSetSplineData(const gmx_pme_t *pme, CodePath mode,
445 const SplineParamsDimVector &splineValues, PmeSplineDataType type, int dimIndex)
447 const pme_atomcomm_t *atc = &(pme->atc[0]);
448 const index atomCount = atc->n;
449 const index pmeOrder = pme->pme_order;
450 const index dimSize = pmeOrder * atomCount;
451 GMX_RELEASE_ASSERT(dimSize == splineValues.size(), "Mismatch in spline data");
452 real *splineBuffer = pmeGetSplineDataInternal(pme, type, dimIndex);
457 std::copy(splineValues.begin(), splineValues.end(), splineBuffer);
461 std::copy(splineValues.begin(), splineValues.end(), splineBuffer);
462 pme_gpu_transform_spline_atom_data(pme->gpu, atc, type, dimIndex, PmeLayoutTransform::HostToGpu);
466 GMX_THROW(InternalError("Test not implemented for this mode"));
470 //! Setting gridline indices to be used in spread/gather
471 void pmeSetGridLineIndices(const gmx_pme_t *pme, CodePath mode,
472 const GridLineIndicesVector &gridLineIndices)
474 const pme_atomcomm_t *atc = &(pme->atc[0]);
475 const index atomCount = atc->n;
476 GMX_RELEASE_ASSERT(atomCount == gridLineIndices.size(), "Mismatch in gridline indices size");
478 IVec paddedGridSizeUnused, gridSize(0, 0, 0);
479 pmeGetRealGridSizesInternal(pme, mode, gridSize, paddedGridSizeUnused);
481 for (const auto &index : gridLineIndices)
483 for (int i = 0; i < DIM; i++)
485 GMX_RELEASE_ASSERT((0 <= index[i]) && (index[i] < gridSize[i]), "Invalid gridline index");
492 memcpy(pme->gpu->staging.h_gridlineIndices, gridLineIndices.data(), atomCount * sizeof(gridLineIndices[0]));
496 // incompatible IVec and ivec assignment?
497 //std::copy(gridLineIndices.begin(), gridLineIndices.end(), atc->idx);
498 memcpy(atc->idx, gridLineIndices.data(), atomCount * sizeof(gridLineIndices[0]));
502 GMX_THROW(InternalError("Test not implemented for this mode"));
506 //! Getting plain index into the complex 3d grid
507 inline size_t pmeGetGridPlainIndexInternal(const IVec &index, const IVec &paddedGridSize, GridOrdering gridOrdering)
510 switch (gridOrdering)
512 case GridOrdering::YZX:
513 result = (index[YY] * paddedGridSize[ZZ] + index[ZZ]) * paddedGridSize[XX] + index[XX];
516 case GridOrdering::XYZ:
517 result = (index[XX] * paddedGridSize[YY] + index[YY]) * paddedGridSize[ZZ] + index[ZZ];
521 GMX_THROW(InternalError("Test not implemented for this mode"));
526 //! Setting real or complex grid
527 template<typename ValueType>
528 static void pmeSetGridInternal(const gmx_pme_t *pme, CodePath mode,
529 GridOrdering gridOrdering,
530 const SparseGridValuesInput<ValueType> &gridValues)
532 IVec gridSize(0, 0, 0), paddedGridSize(0, 0, 0);
534 pmeGetGridAndSizesInternal<ValueType>(pme, mode, grid, gridSize, paddedGridSize);
538 case CodePath::GPU: // intentional absence of break, the grid will be copied from the host buffer in testing mode
540 std::memset(grid, 0, paddedGridSize[XX] * paddedGridSize[YY] * paddedGridSize[ZZ] * sizeof(ValueType));
541 for (const auto &gridValue : gridValues)
543 for (int i = 0; i < DIM; i++)
545 GMX_RELEASE_ASSERT((0 <= gridValue.first[i]) && (gridValue.first[i] < gridSize[i]), "Invalid grid value index");
547 const size_t gridValueIndex = pmeGetGridPlainIndexInternal(gridValue.first, paddedGridSize, gridOrdering);
548 grid[gridValueIndex] = gridValue.second;
553 GMX_THROW(InternalError("Test not implemented for this mode"));
557 //! Setting real grid to be used in gather
558 void pmeSetRealGrid(const gmx_pme_t *pme, CodePath mode,
559 const SparseRealGridValuesInput &gridValues)
561 pmeSetGridInternal<real>(pme, mode, GridOrdering::XYZ, gridValues);
564 //! Setting complex grid to be used in solve
565 void pmeSetComplexGrid(const gmx_pme_t *pme, CodePath mode,
566 GridOrdering gridOrdering,
567 const SparseComplexGridValuesInput &gridValues)
569 pmeSetGridInternal<t_complex>(pme, mode, gridOrdering, gridValues);
572 //! Getting the single dimension's spline values or derivatives
573 SplineParamsDimVector pmeGetSplineData(const gmx_pme_t *pme, CodePath mode,
574 PmeSplineDataType type, int dimIndex)
576 GMX_RELEASE_ASSERT(pme != nullptr, "PME data is not initialized");
577 const pme_atomcomm_t *atc = &(pme->atc[0]);
578 const size_t atomCount = atc->n;
579 const size_t pmeOrder = pme->pme_order;
580 const size_t dimSize = pmeOrder * atomCount;
582 real *sourceBuffer = pmeGetSplineDataInternal(pme, type, dimIndex);
583 SplineParamsDimVector result;
587 pme_gpu_transform_spline_atom_data(pme->gpu, atc, type, dimIndex, PmeLayoutTransform::GpuToHost);
588 result = arrayRefFromArray(sourceBuffer, dimSize);
592 result = arrayRefFromArray(sourceBuffer, dimSize);
596 GMX_THROW(InternalError("Test not implemented for this mode"));
601 //! Getting the gridline indices
602 GridLineIndicesVector pmeGetGridlineIndices(const gmx_pme_t *pme, CodePath mode)
604 GMX_RELEASE_ASSERT(pme != nullptr, "PME data is not initialized");
605 const pme_atomcomm_t *atc = &(pme->atc[0]);
606 const size_t atomCount = atc->n;
608 GridLineIndicesVector gridLineIndices;
612 gridLineIndices = arrayRefFromArray(reinterpret_cast<IVec *>(pme->gpu->staging.h_gridlineIndices), atomCount);
616 gridLineIndices = arrayRefFromArray(reinterpret_cast<IVec *>(atc->idx), atomCount);
620 GMX_THROW(InternalError("Test not implemented for this mode"));
622 return gridLineIndices;
625 //! Getting real or complex grid - only non zero values
626 template<typename ValueType>
627 static SparseGridValuesOutput<ValueType> pmeGetGridInternal(const gmx_pme_t *pme, CodePath mode, GridOrdering gridOrdering)
629 IVec gridSize(0, 0, 0), paddedGridSize(0, 0, 0);
631 pmeGetGridAndSizesInternal<ValueType>(pme, mode, grid, gridSize, paddedGridSize);
632 SparseGridValuesOutput<ValueType> gridValues;
635 case CodePath::GPU: // intentional absence of break
638 for (int ix = 0; ix < gridSize[XX]; ix++)
640 for (int iy = 0; iy < gridSize[YY]; iy++)
642 for (int iz = 0; iz < gridSize[ZZ]; iz++)
644 IVec temp(ix, iy, iz);
645 const size_t gridValueIndex = pmeGetGridPlainIndexInternal(temp, paddedGridSize, gridOrdering);
646 const ValueType value = grid[gridValueIndex];
647 if (value != ValueType {})
649 auto key = formatString("Cell %d %d %d", ix, iy, iz);
650 gridValues[key] = value;
658 GMX_THROW(InternalError("Test not implemented for this mode"));
663 //! Getting the real grid (spreading output of pmePerformSplineAndSpread())
664 SparseRealGridValuesOutput pmeGetRealGrid(const gmx_pme_t *pme, CodePath mode)
666 return pmeGetGridInternal<real>(pme, mode, GridOrdering::XYZ);
669 //! Getting the complex grid output of pmePerformSolve()
670 SparseComplexGridValuesOutput pmeGetComplexGrid(const gmx_pme_t *pme, CodePath mode,
671 GridOrdering gridOrdering)
673 return pmeGetGridInternal<t_complex>(pme, mode, gridOrdering);
676 //! Getting the reciprocal energy and virial
677 PmeOutput pmeGetReciprocalEnergyAndVirial(const gmx_pme_t *pme, CodePath mode,
678 PmeSolveAlgorithm method)
686 case PmeSolveAlgorithm::Coulomb:
687 get_pme_ener_vir_q(pme->solve_work, pme->nthread, &output);
690 case PmeSolveAlgorithm::LennardJones:
691 get_pme_ener_vir_lj(pme->solve_work, pme->nthread, &output);
695 GMX_THROW(InternalError("Test not implemented for this mode"));
701 case PmeSolveAlgorithm::Coulomb:
702 output = pme_gpu_getEnergyAndVirial(*pme);
706 GMX_THROW(InternalError("Test not implemented for this mode"));
711 GMX_THROW(InternalError("Test not implemented for this mode"));