/* ************************************************************************ * Copyright (C) 2020-2024 Advanced Micro Devices, Inc. All rights reserved. * * Permission is hereby granted, free of charge, to any person obtaining a copy * of this software and associated documentation files (the "Software"), to deal * in the Software without restriction, including without limitation the rights * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell cop- * ies of the Software, and to permit persons to whom the Software is furnished * to do so, subject to the following conditions: * * The above copyright notice and this permission notice shall be included in all * copies or substantial portions of the Software. * * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IM- * PLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS * FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR * COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER * IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNE- * CTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. * * * ************************************************************************ */ #include "clientcommon.hpp" template void getrf_checkBadArgs(const hipsolverHandle_t handle, const hipsolverDnParams_t params, const I m, const I n, Td dA, const I lda, const I stA, Id dIpiv, const I stP, Td dWork, const SIZE dlwork, Th hWork, const SIZE hlwork, INTd dinfo, const int bc) { // handle EXPECT_ROCBLAS_STATUS(hipsolver_getrf(API, false, nullptr, params, m, n, dA, lda, stA, dIpiv, stP, dWork, dlwork, hWork, hlwork, dinfo, bc), HIPSOLVER_STATUS_NOT_INITIALIZED); // values // N/A #if defined(__HIP_PLATFORM_HCC__) || defined(__HIP_PLATFORM_AMD__) // pointers if constexpr(!std::is_same::value) EXPECT_ROCBLAS_STATUS(hipsolver_getrf(API, false, handle, (hipsolverDnParams_t) nullptr, m, n, dA, lda, stA, dIpiv, stP, dWork, dlwork, hWork, hlwork, dinfo, bc), HIPSOLVER_STATUS_INVALID_VALUE); EXPECT_ROCBLAS_STATUS(hipsolver_getrf(API, false, handle, params, m, n, (Td) nullptr, lda, stA, dIpiv, stP, dWork, dlwork, hWork, hlwork, dinfo, bc), HIPSOLVER_STATUS_INVALID_VALUE); EXPECT_ROCBLAS_STATUS(hipsolver_getrf(API, false, handle, params, m, n, dA, lda, stA, dIpiv, stP, dWork, dlwork, hWork, hlwork, (INTd) nullptr, bc), HIPSOLVER_STATUS_INVALID_VALUE); #endif } template void testing_getrf_bad_arg() { // safe arguments hipsolver_local_handle handle; hipsolver_local_params params; I m = 1; I n = 1; I lda = 1; I stA = 1; I stP = 1; int bc = 1; if(BATCHED) { // // memory allocations // device_batch_vector dA(1, 1, 1); // device_strided_batch_vector dIpiv(1, 1, 1, 1); // device_strided_batch_vector dInfo(1, 1, 1, 1); // CHECK_HIP_ERROR(dA.memcheck()); // CHECK_HIP_ERROR(dIpiv.memcheck()); // CHECK_HIP_ERROR(dInfo.memcheck()); // SIZE size_dW, size_hW; // hipsolver_getrf_bufferSize(API, handle, params, m, n, dA.data(), lda, &size_dW, &size_hW); // host_strided_batch_vector hWork(size_hW, 1, size_hW, 1); // device_strided_batch_vector dWork(size_dW, 1, size_dW, 1); // if(size_dW) // CHECK_HIP_ERROR(dWork.memcheck()); // // check bad arguments // getrf_checkBadArgs( // handle, params, m, n, dA.data(), lda, stA, dIpiv.data(), stP, dWork.data(), size_dW, hWork.data(), size_hW, dInfo.data(), bc); } else { // memory allocations device_strided_batch_vector dA(1, 1, 1, 1); device_strided_batch_vector dIpiv(1, 1, 1, 1); device_strided_batch_vector dInfo(1, 1, 1, 1); CHECK_HIP_ERROR(dA.memcheck()); CHECK_HIP_ERROR(dIpiv.memcheck()); CHECK_HIP_ERROR(dInfo.memcheck()); SIZE size_dW, size_hW; hipsolver_getrf_bufferSize(API, handle, params, m, n, dA.data(), lda, &size_dW, &size_hW); host_strided_batch_vector hWork(size_hW, 1, size_hW, 1); device_strided_batch_vector dWork(size_dW, 1, size_dW, 1); if(size_dW) CHECK_HIP_ERROR(dWork.memcheck()); // check bad arguments getrf_checkBadArgs(handle, params, m, n, dA.data(), lda, stA, dIpiv.data(), stP, dWork.data(), size_dW, hWork.data(), size_hW, dInfo.data(), bc); } } template void getrf_initData(const hipsolverHandle_t handle, const hipsolverDnParams_t params, const I m, const I n, Td& dA, const I lda, const I stA, Id& dIpiv, const I stP, INTd& dInfo, const int bc, Th& hA, Ih& hIpiv, INTh& hInfo) { if(CPU) { T tmp; rocblas_init(hA, true); for(int b = 0; b < bc; ++b) { // scale A to avoid singularities for(I i = 0; i < m; i++) { for(I j = 0; j < n; j++) { if(i == j) hA[b][i + j * lda] += 400; else hA[b][i + j * lda] -= 4; } } if(!NPVT) { // shuffle rows to test pivoting // always the same permuation for debugging purposes for(I i = 0; i < m / 2; i++) { for(I j = 0; j < n; j++) { tmp = hA[b][i + j * lda]; hA[b][i + j * lda] = hA[b][m - 1 - i + j * lda]; hA[b][m - 1 - i + j * lda] = tmp; } } } } } if(GPU) { // now copy data to the GPU CHECK_HIP_ERROR(dA.transfer_from(hA)); } } template void getrf_getError(const hipsolverHandle_t handle, const hipsolverDnParams_t params, const I m, const I n, Td& dA, const I lda, const I stA, Id& dIpiv, const I stP, Td& dWork, const SIZE dlwork, Th& hWork, const SIZE hlwork, INTd& dInfo, const int bc, Th& hA, Th& hARes, INTh& hIpiv, Ih& hIpivRes, INTh& hInfo, INTh& hInfoRes, double* max_err) { // input data initialization getrf_initData( handle, params, m, n, dA, lda, stA, dIpiv, stP, dInfo, bc, hA, hIpiv, hInfo); // execute computations // GPU lapack CHECK_ROCBLAS_ERROR(hipsolver_getrf(API, NPVT, handle, params, m, n, dA.data(), lda, stA, dIpiv.data(), stP, dWork.data(), dlwork, hWork.data(), hlwork, dInfo.data(), bc)); CHECK_HIP_ERROR(hARes.transfer_from(dA)); CHECK_HIP_ERROR(hIpivRes.transfer_from(dIpiv)); CHECK_HIP_ERROR(hInfoRes.transfer_from(dInfo)); // CPU lapack for(int b = 0; b < bc; ++b) cpu_getrf(m, n, hA[b], lda, hIpiv[b], hInfo[b]); // expecting original matrix to be non-singular // error is ||hA - hARes|| / ||hA|| (ideally ||LU - Lres Ures|| / ||LU||) // (THIS DOES NOT ACCOUNT FOR NUMERICAL REPRODUCIBILITY ISSUES. // IT MIGHT BE REVISITED IN THE FUTURE) // using frobenius norm double err; *max_err = 0; for(int b = 0; b < bc; ++b) { err = norm_error('F', m, n, lda, hA[b], hARes[b]); *max_err = err > *max_err ? err : *max_err; // also check pivoting (count the number of incorrect pivots) if(!NPVT) { err = 0; for(I i = 0; i < min(m, n); ++i) { EXPECT_EQ(hIpiv[b][i], hIpivRes[b][i]) << "where b = " << b << ", i = " << i; if(hIpiv[b][i] != hIpivRes[b][i]) err++; } *max_err = err > *max_err ? err : *max_err; } } // also check info for singularities err = 0; for(int b = 0; b < bc; ++b) { EXPECT_EQ(hInfo[b][0], hInfoRes[b][0]) << "where b = " << b; if(hInfo[b][0] != hInfoRes[b][0]) err++; } *max_err += err; } template void getrf_getPerfData(const hipsolverHandle_t handle, const hipsolverDnParams_t params, const I m, const I n, Td& dA, const I lda, const I stA, Id& dIpiv, const I stP, Td& dWork, const SIZE dlwork, Th& hWork, const SIZE hlwork, INTd& dInfo, const int bc, Th& hA, INTh& hIpiv, INTh& hInfo, double* gpu_time_used, double* cpu_time_used, const int hot_calls, const bool perf) { if(!perf) { getrf_initData( handle, params, m, n, dA, lda, stA, dIpiv, stP, dInfo, bc, hA, hIpiv, hInfo); // cpu-lapack performance (only if not in perf mode) *cpu_time_used = get_time_us_no_sync(); for(int b = 0; b < bc; ++b) cpu_getrf(m, n, hA[b], lda, hIpiv[b], hInfo[b]); *cpu_time_used = get_time_us_no_sync() - *cpu_time_used; } getrf_initData( handle, params, m, n, dA, lda, stA, dIpiv, stP, dInfo, bc, hA, hIpiv, hInfo); // cold calls for(int iter = 0; iter < 2; iter++) { getrf_initData( handle, params, m, n, dA, lda, stA, dIpiv, stP, dInfo, bc, hA, hIpiv, hInfo); CHECK_ROCBLAS_ERROR(hipsolver_getrf(API, NPVT, handle, params, m, n, dA.data(), lda, stA, dIpiv.data(), stP, dWork.data(), dlwork, hWork.data(), hlwork, dInfo.data(), bc)); } // gpu-lapack performance hipStream_t stream; CHECK_ROCBLAS_ERROR(hipsolverGetStream(handle, &stream)); double start; for(int iter = 0; iter < hot_calls; iter++) { getrf_initData( handle, params, m, n, dA, lda, stA, dIpiv, stP, dInfo, bc, hA, hIpiv, hInfo); start = get_time_us_sync(stream); hipsolver_getrf(API, NPVT, handle, params, m, n, dA.data(), lda, stA, dIpiv.data(), stP, dWork.data(), dlwork, hWork.data(), hlwork, dInfo.data(), bc); *gpu_time_used += get_time_us_sync(stream) - start; } *gpu_time_used /= hot_calls; } template void testing_getrf(Arguments& argus) { // get arguments hipsolver_local_handle handle; hipsolver_local_params params; I m = argus.get("m"); I n = argus.get("n", m); I lda = argus.get("lda", m); I stA = argus.get("strideA", lda * n); I stP = argus.get("strideP", min(m, n)); int bc = argus.batch_count; int hot_calls = argus.iters; I stARes = (argus.unit_check || argus.norm_check) ? stA : 0; I stPRes = (argus.unit_check || argus.norm_check) ? stP : 0; // check non-supported values // N/A // determine sizes size_t size_A = size_t(lda) * n; size_t size_P = size_t(min(m, n)); double max_error = 0, gpu_time_used = 0, cpu_time_used = 0; size_t size_ARes = (argus.unit_check || argus.norm_check) ? size_A : 0; size_t size_PRes = (argus.unit_check || argus.norm_check) ? size_P : 0; // check invalid sizes bool invalid_size = (m <= 0 || n <= 0 || lda < m || bc <= 0); if(invalid_size) { #if defined(__HIP_PLATFORM_HCC__) || defined(__HIP_PLATFORM_AMD__) if(BATCHED) { // EXPECT_ROCBLAS_STATUS(hipsolver_getrf(API, // NPVT, // handle, // params, // m, // n, // (T* const*)nullptr, // lda, // stA, // (I*)nullptr, // stP, // (T*)nullptr, // 0, // (T*)nullptr, // 0, // (int*)nullptr, // bc), // HIPSOLVER_STATUS_INVALID_VALUE); } else { EXPECT_ROCBLAS_STATUS(hipsolver_getrf(API, NPVT, handle, params, m, n, (T*)nullptr, lda, stA, (I*)nullptr, stP, (T*)nullptr, 0, (T*)nullptr, 0, (int*)nullptr, bc), HIPSOLVER_STATUS_INVALID_VALUE); } #endif if(argus.timing) rocsolver_bench_inform(inform_invalid_size); return; } // memory size query is necessary SIZE size_dW, size_hW; hipsolver_getrf_bufferSize(API, handle, params, m, n, (T*)nullptr, lda, &size_dW, &size_hW); if(argus.mem_query) { rocsolver_bench_inform(inform_mem_query, size_dW); return; } if(BATCHED) { // // memory allocations // host_batch_vector hA(size_A, 1, bc); // host_batch_vector hARes(size_ARes, 1, bc); // host_strided_batch_vector hIpiv(size_P, 1, stP, bc); // host_strided_batch_vector hIpivRes(size_PRes, 1, stPRes, bc); // host_strided_batch_vector hInfo(1, 1, 1, bc); // host_strided_batch_vector hInfoRes(1, 1, 1, bc); // host_strided_batch_vector hWork(size_hW, 1, size_hW, 1); // size_hW accounts for bc // device_batch_vector dA(size_A, 1, bc); // device_strided_batch_vector dIpiv(size_P, 1, stP, bc); // device_strided_batch_vector dInfo(1, 1, 1, bc); // device_strided_batch_vector dWork(size_dW, 1, size_dW, 1); // size_dW accounts for bc // if(size_A) // CHECK_HIP_ERROR(dA.memcheck()); // CHECK_HIP_ERROR(dInfo.memcheck()); // if(size_P) // CHECK_HIP_ERROR(dIpiv.memcheck()); // if(size_dW) // CHECK_HIP_ERROR(dWork.memcheck()); // // check computations // if(argus.unit_check || argus.norm_check) // getrf_getError(handle, // params, // m, // n, // dA, // lda, // stA, // dIpiv, // stP, // dWork, // size_dW, // hWork, // size_hW, // dInfo, // bc, // hA, // hARes, // hIpiv, // hIpivRes, // hInfo, // hInfoRes, // &max_error); // // collect performance data // if(argus.timing) // getrf_getPerfData(handle, // params, // m, // n, // dA, // lda, // stA, // dIpiv, // stP, // dWork, // size_dW, // hWork, // size_hW, // dInfo, // bc, // hA, // hIpiv, // hInfo, // &gpu_time_used, // &cpu_time_used, // hot_calls, // argus.perf); } else { // memory allocations host_strided_batch_vector hA(size_A, 1, stA, bc); host_strided_batch_vector hARes(size_ARes, 1, stARes, bc); host_strided_batch_vector hIpiv(size_P, 1, stP, bc); host_strided_batch_vector hIpivRes(size_PRes, 1, stPRes, bc); host_strided_batch_vector hInfo(1, 1, 1, bc); host_strided_batch_vector hInfoRes(1, 1, 1, bc); host_strided_batch_vector hWork(size_hW, 1, size_hW, 1); // size_hW accounts for bc device_strided_batch_vector dA(size_A, 1, stA, bc); device_strided_batch_vector dIpiv(size_P, 1, stP, bc); device_strided_batch_vector dInfo(1, 1, 1, bc); device_strided_batch_vector dWork(size_dW, 1, size_dW, 1); // size_dW accounts for bc if(size_A) CHECK_HIP_ERROR(dA.memcheck()); CHECK_HIP_ERROR(dInfo.memcheck()); if(size_P) CHECK_HIP_ERROR(dIpiv.memcheck()); if(size_dW) CHECK_HIP_ERROR(dWork.memcheck()); // check computations if(argus.unit_check || argus.norm_check) getrf_getError(handle, params, m, n, dA, lda, stA, dIpiv, stP, dWork, size_dW, hWork, size_hW, dInfo, bc, hA, hARes, hIpiv, hIpivRes, hInfo, hInfoRes, &max_error); // collect performance data if(argus.timing) getrf_getPerfData(handle, params, m, n, dA, lda, stA, dIpiv, stP, dWork, size_dW, hWork, size_hW, dInfo, bc, hA, hIpiv, hInfo, &gpu_time_used, &cpu_time_used, hot_calls, argus.perf); } // validate results for rocsolver-test // using min(m,n) * machine_precision as tolerance if(argus.unit_check) ROCSOLVER_TEST_CHECK(T, max_error, min(m, n)); // output results for rocsolver-bench if(argus.timing) { if(!argus.perf) { std::cerr << "\n============================================\n"; std::cerr << "Arguments:\n"; std::cerr << "============================================\n"; if(BATCHED) { rocsolver_bench_output("m", "n", "lda", "strideP", "batch_c"); rocsolver_bench_output(m, n, lda, stP, bc); } else if(STRIDED) { rocsolver_bench_output("m", "n", "lda", "strideA", "strideP", "batch_c"); rocsolver_bench_output(m, n, lda, stA, stP, bc); } else { rocsolver_bench_output("m", "n", "lda"); rocsolver_bench_output(m, n, lda); } std::cerr << "\n============================================\n"; std::cerr << "Results:\n"; std::cerr << "============================================\n"; if(argus.norm_check) { rocsolver_bench_output("cpu_time", "gpu_time", "error"); rocsolver_bench_output(cpu_time_used, gpu_time_used, max_error); } else { rocsolver_bench_output("cpu_time", "gpu_time"); rocsolver_bench_output(cpu_time_used, gpu_time_used); } std::cerr << std::endl; } else { if(argus.norm_check) rocsolver_bench_output(gpu_time_used, max_error); else rocsolver_bench_output(gpu_time_used); } } // ensure all arguments were consumed argus.validate_consumed(); }