// Copyright 2002 - 2008, 2010, 2011 National Technology Engineering // Solutions of Sandia, LLC (NTESS). Under the terms of Contract // DE-NA0003525 with NTESS, the U.S. Government retains certain rights // in this software. // // Redistribution and use in source and binary forms, with or without // modification, are permitted provided that the following conditions are // met: // // * Redistributions of source code must retain the above copyright // notice, this list of conditions and the following disclaimer. // // * Redistributions in binary form must reproduce the above // copyright notice, this list of conditions and the following // disclaimer in the documentation and/or other materials provided // with the distribution. // // * Neither the name of NTESS nor the names of its contributors // may be used to endorse or promote products derived from this // software without specific prior written permission. // // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS // "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT // LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR // A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT // OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, // SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT // LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, // DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY // THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. // #include "gtest/gtest.h" #include "stk_util/parallel/Parallel.hpp" // for parallel_machine_rank, parallel_machine_... #include "stk_util/parallel/ParallelReduce.hpp" // for all_reduce_max, all_reduce_maxloc, all_r... #include "stk_util/stk_config.h" // for STK_HAS_MPI #include // for uint64_t, uint32_t #include // for numeric_limits #include // for vector #if defined ( STK_HAS_MPI ) TEST(ParallelComm, AllReduceLoc) { stk::ParallelMachine comm = MPI_COMM_WORLD; int myProcId = stk::parallel_machine_rank(comm); int numProcs = stk::parallel_machine_size(comm); int nvalues=5; uint64_t limit_32bit_integer = std::numeric_limits::max(); std::vector values(nvalues); std::vector locations(nvalues); std::vector globalValues(nvalues); std::vector globalLocations(nvalues); for(int n = 0; n < nvalues; n++) { values[n] = myProcId*numProcs+n; locations[n] = limit_32bit_integer + myProcId*numProcs*numProcs+n; //Want to test when outside 32-bit range for index. } stk::all_reduce_maxloc(comm, values.data(), locations.data(), globalValues.data(), globalLocations.data(), nvalues); for(int n = 0; n < nvalues; n++) { EXPECT_EQ((numProcs-1)*numProcs+n, globalValues[n]); EXPECT_EQ(limit_32bit_integer + (numProcs-1)*numProcs*numProcs+n, globalLocations[n]); } stk::all_reduce_minloc(comm, values.data(), locations.data(), globalValues.data(), globalLocations.data(), nvalues); for(int n = 0; n < nvalues; n++) { EXPECT_EQ(n, globalValues[n]); EXPECT_EQ(limit_32bit_integer+n, globalLocations[n]); } } TEST(ParallelComm, AllReduce) { stk::ParallelMachine comm = MPI_COMM_WORLD; int myProcId = stk::parallel_machine_rank(comm); int numProcs = stk::parallel_machine_size(comm); int nvalues=5; std::vector values(nvalues); std::vector globalValues(nvalues); for(int n = 0; n < nvalues; n++) { values[n] = myProcId*numProcs+n; } stk::all_reduce_max(comm, values.data(), globalValues.data(), nvalues); for(int n = 0; n < nvalues; n++) { EXPECT_EQ((numProcs-1)*numProcs+n, globalValues[n]); } stk::all_reduce_min(comm, values.data(), globalValues.data(), nvalues); for(int n = 0; n < nvalues; n++) { EXPECT_EQ(n, globalValues[n]); } int alpha = 0; for(int n = 0; n < numProcs; ++n) { alpha += n*numProcs; } stk::all_reduce_sum(comm, values.data(), globalValues.data(), nvalues); for(int n = 0; n < nvalues; n++) { EXPECT_EQ(n*numProcs + alpha, globalValues[n]); } } TEST(ParallelComm, GetGlobal) { stk::ParallelMachine comm = MPI_COMM_WORLD; int myProcId = stk::parallel_machine_rank(comm); int numProcs = stk::parallel_machine_size(comm); double localValue = myProcId * numProcs; double globalMax = stk::get_global_max(comm, localValue); EXPECT_EQ(globalMax, (numProcs-1)*numProcs); double globalMin = stk::get_global_min(comm, localValue); EXPECT_EQ(globalMin, 0); int expectedSum = 0; for(int n = 0; n < numProcs; ++n) { expectedSum += n*numProcs; } double globalSum = stk::get_global_sum(comm, localValue); EXPECT_EQ(globalSum, expectedSum); } TEST(ParallelComm, uint64_reduce_min) { // Test coverage of openmpi-4.1.4 bug with unsigned integer min reduction stk::ParallelMachine comm = MPI_COMM_WORLD; int myProcId = stk::parallel_machine_rank(comm); const uint64_t goldValue = 10; uint64_t id = (myProcId == 0) ? goldValue : std::numeric_limits::max(); uint64_t gid; stk::all_reduce_min(comm, &id, &gid, 1u); EXPECT_EQ(goldValue, gid); } #endif