/* ************************************************************************ * Copyright (C) 2025-2026 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. * * * ************************************************************************ */ #pragma once #include "clientcommon.hpp" template void sytrs_checkBadArgs(const hipsolverHandle_t handle, const hipsolverFillMode_t uplo, const I n, const I nrhs, Td dA, const I lda, Id dIpiv, Td dB, const I ldb, Ud dWork, const SIZE dlwork, Ud hWork, const SIZE hlwork, INTd dInfo, const int bc) { // handle EXPECT_ROCBLAS_STATUS(hipsolver_sytrs(API, nullptr, uplo, n, nrhs, dA, lda, dIpiv, dB, ldb, dWork, dlwork, hWork, hlwork, dInfo, bc), HIPSOLVER_STATUS_NOT_INITIALIZED); // values EXPECT_ROCBLAS_STATUS(hipsolver_sytrs(API, handle, hipsolverFillMode_t(-1), n, nrhs, dA, lda, dIpiv, dB, ldb, dWork, dlwork, hWork, hlwork, dInfo, bc), HIPSOLVER_STATUS_INVALID_ENUM); #if defined(__HIP_PLATFORM_HCC__) || defined(__HIP_PLATFORM_AMD__) // pointers EXPECT_ROCBLAS_STATUS(hipsolver_sytrs(API, handle, uplo, n, nrhs, (Td) nullptr, lda, dIpiv, dB, ldb, dWork, dlwork, hWork, hlwork, dInfo, bc), HIPSOLVER_STATUS_INVALID_VALUE); EXPECT_ROCBLAS_STATUS(hipsolver_sytrs(API, handle, uplo, n, nrhs, dA, lda, (Id) nullptr, dB, ldb, dWork, dlwork, hWork, hlwork, dInfo, bc), HIPSOLVER_STATUS_INVALID_VALUE); EXPECT_ROCBLAS_STATUS(hipsolver_sytrs(API, handle, uplo, n, nrhs, dA, lda, dIpiv, (Td) nullptr, ldb, dWork, dlwork, hWork, hlwork, dInfo, bc), HIPSOLVER_STATUS_INVALID_VALUE); EXPECT_ROCBLAS_STATUS(hipsolver_sytrs(API, handle, uplo, n, nrhs, dA, lda, dIpiv, dB, ldb, dWork, dlwork, hWork, hlwork, (INTd) nullptr, bc), HIPSOLVER_STATUS_INVALID_VALUE); #endif } template void testing_sytrs_bad_arg() { // safe arguments hipsolver_local_handle handle; hipsolverFillMode_t uplo = HIPSOLVER_FILL_MODE_UPPER; I n = 1; I nrhs = 1; I lda = 1; I ldb = 1; int bc = 1; if(BATCHED) { // // memory allocations // (no batched DnX API for sytrs) } else { // memory allocations device_strided_batch_vector dA(1, 1, 1, 1); device_strided_batch_vector dB(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(dB.memcheck()); CHECK_HIP_ERROR(dIpiv.memcheck()); CHECK_HIP_ERROR(dInfo.memcheck()); SIZE bytes_dW, bytes_hW; hipsolver_sytrs_bufferSize(API, handle, uplo, n, nrhs, dA.data(), lda, dIpiv.data(), dB.data(), ldb, &bytes_dW, &bytes_hW); SIZE size_dW = (bytes_dW + sizeof(T) - 1) / sizeof(T); SIZE size_hW = (bytes_hW + sizeof(T) - 1) / sizeof(T); 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 sytrs_checkBadArgs(handle, uplo, n, nrhs, dA.data(), lda, dIpiv.data(), dB.data(), ldb, dWork.data(), bytes_dW, hWork.data(), bytes_hW, dInfo.data(), bc); } } template void sytrs_initData(const hipsolverHandle_t handle, const hipsolverFillMode_t uplo, const I n, const I nrhs, Td& dA, const I lda, Id& dIpiv, Td& dB, const I ldb, const int bc, Th& hA, Ih& hIpiv, INTh& hIpiv_cpu, Th& hB) { if(CPU) { T tmp; rocblas_init(hA, true); rocblas_init(hB, true); for(int b = 0; b < bc; ++b) { // scale A to avoid singularities for(I i = 0; i < n; i++) { for(I j = 0; j < n; j++) { if(i == j) hA[b][i + j * lda] += 400; else hA[b][i + j * lda] -= 4; } } // shuffle rows to test pivoting // always the same permutation for debugging purposes for(I i = 0; i < n / 2; i++) { for(I j = 0; j < n; j++) { std::swap(hA[b][i + j * lda], hA[b][n - 1 - i + j * lda]); } } // do the LDL^T factorization of matrix A w/ the reference LAPACK routine int sytrf_lwork = lda * n; std::vector hWork(sytrf_lwork); int info = 0; cpu_sytrf(uplo, n, hA[b], lda, hIpiv_cpu[b], hWork.data(), sytrf_lwork, &info); // copy ipiv from int to int64_t for(I i = 0; i < n; i++) hIpiv[b][i] = hIpiv_cpu[b][i]; } } if(GPU) { // now copy pivoting indices and matrices to the GPU CHECK_HIP_ERROR(dA.transfer_from(hA)); CHECK_HIP_ERROR(dB.transfer_from(hB)); CHECK_HIP_ERROR(dIpiv.transfer_from(hIpiv)); } } template void sytrs_getError(const hipsolverHandle_t handle, const hipsolverFillMode_t uplo, const I n, const I nrhs, Td& dA, const I lda, Id& dIpiv, Td& dB, const I ldb, void* dWork, const SIZE dlwork, void* hWork, const SIZE hlwork, INTd& dInfo, const int bc, Th& hA, Ih& hIpiv, INTh& hIpiv_cpu, Th& hB, Th& hBRes, INTh& hInfo, INTh& hInfoRes, double* max_err) { // input data initialization sytrs_initData( handle, uplo, n, nrhs, dA, lda, dIpiv, dB, ldb, bc, hA, hIpiv, hIpiv_cpu, hB); // execute computations // GPU lapack CHECK_ROCBLAS_ERROR(hipsolver_sytrs(API, handle, uplo, n, nrhs, dA.data(), lda, dIpiv.data(), dB.data(), ldb, dWork, dlwork, hWork, hlwork, dInfo.data(), bc)); CHECK_HIP_ERROR(hBRes.transfer_from(dB)); CHECK_HIP_ERROR(hInfoRes.transfer_from(dInfo)); // CPU lapack for(int b = 0; b < bc; ++b) { cpu_sytrs(uplo, n, nrhs, hA[b], lda, hIpiv_cpu[b], hB[b], ldb, hInfo[b]); } // error is ||hB - hBRes|| / ||hB|| // (THIS DOES NOT ACCOUNT FOR NUMERICAL REPRODUCIBILITY ISSUES. // IT MIGHT BE REVISITED IN THE FUTURE) // using vector-induced infinity norm double err; *max_err = 0; for(int b = 0; b < bc; ++b) { err = norm_error('I', n, nrhs, ldb, hB[b], hBRes[b]); *max_err = err > *max_err ? err : *max_err; } // check info 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 sytrs_getPerfData(const hipsolverHandle_t handle, const hipsolverFillMode_t uplo, const I n, const I nrhs, Td& dA, const I lda, Id& dIpiv, Td& dB, const I ldb, void* dWork, const SIZE dlwork, void* hWork, const SIZE hlwork, INTd& dInfo, const int bc, Th& hA, Ih& hIpiv, INTh& hIpiv_cpu, Th& hB, INTh& hInfo, double* gpu_time_used, double* cpu_time_used, const int hot_calls, const bool perf) { if(!perf) { sytrs_initData( handle, uplo, n, nrhs, dA, lda, dIpiv, dB, ldb, bc, hA, hIpiv, hIpiv_cpu, hB); // 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_sytrs(uplo, n, nrhs, hA[b], lda, hIpiv_cpu[b], hB[b], ldb, hInfo[b]); } *cpu_time_used = get_time_us_no_sync() - *cpu_time_used; } sytrs_initData( handle, uplo, n, nrhs, dA, lda, dIpiv, dB, ldb, bc, hA, hIpiv, hIpiv_cpu, hB); // cold calls for(int iter = 0; iter < 2; iter++) { sytrs_initData( handle, uplo, n, nrhs, dA, lda, dIpiv, dB, ldb, bc, hA, hIpiv, hIpiv_cpu, hB); CHECK_ROCBLAS_ERROR(hipsolver_sytrs(API, handle, uplo, n, nrhs, dA.data(), lda, dIpiv.data(), dB.data(), ldb, dWork, dlwork, hWork, 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++) { sytrs_initData( handle, uplo, n, nrhs, dA, lda, dIpiv, dB, ldb, bc, hA, hIpiv, hIpiv_cpu, hB); start = get_time_us_sync(stream); hipsolver_sytrs(API, handle, uplo, n, nrhs, dA.data(), lda, dIpiv.data(), dB.data(), ldb, dWork, dlwork, hWork, hlwork, dInfo.data(), bc); *gpu_time_used += get_time_us_sync(stream) - start; } *gpu_time_used /= hot_calls; } template void testing_sytrs(Arguments& argus) { // get arguments hipsolver_local_handle handle; char uploC = argus.get("uplo"); I n = argus.get("n"); I nrhs = argus.get("nrhs", n); I lda = argus.get("lda", n); I ldb = argus.get("ldb", n); hipsolverFillMode_t uplo = char2hipsolver_fill(uploC); int bc = argus.batch_count; int hot_calls = argus.iters; // check non-supported values if(uplo != HIPSOLVER_FILL_MODE_UPPER && uplo != HIPSOLVER_FILL_MODE_LOWER) { if(BATCHED) { // (no batched DnX API for sytrs) } else { EXPECT_ROCBLAS_STATUS(hipsolver_sytrs(API, handle, uplo, n, nrhs, (T*)nullptr, lda, (I*)nullptr, (T*)nullptr, ldb, (void*)nullptr, (SIZE)0, (void*)nullptr, (SIZE)0, (int*)nullptr, bc), HIPSOLVER_STATUS_INVALID_ENUM); } if(argus.timing) rocsolver_bench_inform(inform_invalid_args); return; } // determine sizes size_t size_A = size_t(lda) * n; size_t size_B = size_t(ldb) * nrhs; size_t size_P = size_t(n); double max_error = 0, gpu_time_used = 0, cpu_time_used = 0; size_t size_BRes = (argus.unit_check || argus.norm_check) ? size_B : 0; // check invalid sizes bool invalid_size = (n < 0 || nrhs < 0 || lda < n || ldb < n || bc < 0); if(invalid_size) { if(BATCHED) { // (no batched DnX API for sytrs) } else { EXPECT_ROCBLAS_STATUS(hipsolver_sytrs(API, handle, uplo, n, nrhs, (T*)nullptr, lda, (I*)nullptr, (T*)nullptr, ldb, (void*)nullptr, (SIZE)0, (void*)nullptr, (SIZE)0, (int*)nullptr, bc), HIPSOLVER_STATUS_INVALID_VALUE); } if(argus.timing) rocsolver_bench_inform(inform_invalid_size); return; } // memory size query is necessary SIZE bytes_dW, bytes_hW; hipsolver_sytrs_bufferSize(API, handle, uplo, n, nrhs, (T*)nullptr, lda, (I*)nullptr, (T*)nullptr, ldb, &bytes_dW, &bytes_hW); SIZE size_dW = (bytes_dW + sizeof(T) - 1) / sizeof(T); SIZE size_hW = (bytes_hW + sizeof(T) - 1) / sizeof(T); if(argus.mem_query) { rocsolver_bench_inform(inform_mem_query, bytes_dW); return; } if(BATCHED) { // (no batched DnX API for sytrs) } else { // memory allocations host_strided_batch_vector hA(size_A, 1, size_A, bc); host_strided_batch_vector hB(size_B, 1, size_B, bc); host_strided_batch_vector hBRes(size_BRes, 1, size_BRes, bc); host_strided_batch_vector hIpiv(size_P, 1, size_P, bc); host_strided_batch_vector hIpiv_cpu(size_P, 1, size_P, 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); device_strided_batch_vector dA(size_A, 1, size_A, bc); device_strided_batch_vector dB(size_B, 1, size_B, bc); device_strided_batch_vector dIpiv(size_P, 1, size_P, bc); device_strided_batch_vector dInfo(1, 1, 1, bc); device_strided_batch_vector dWork(size_dW, 1, size_dW, 1); if(size_A) CHECK_HIP_ERROR(dA.memcheck()); if(size_B) CHECK_HIP_ERROR(dB.memcheck()); if(size_P) CHECK_HIP_ERROR(dIpiv.memcheck()); CHECK_HIP_ERROR(dInfo.memcheck()); if(size_dW) CHECK_HIP_ERROR(dWork.memcheck()); // check computations if(argus.unit_check || argus.norm_check) sytrs_getError(handle, uplo, n, nrhs, dA, lda, dIpiv, dB, ldb, dWork.data(), bytes_dW, hWork.data(), bytes_hW, dInfo, bc, hA, hIpiv, hIpiv_cpu, hB, hBRes, hInfo, hInfoRes, &max_error); // collect performance data if(argus.timing) sytrs_getPerfData(handle, uplo, n, nrhs, dA, lda, dIpiv, dB, ldb, dWork.data(), bytes_dW, hWork.data(), bytes_hW, dInfo, bc, hA, hIpiv, hIpiv_cpu, hB, hInfo, &gpu_time_used, &cpu_time_used, hot_calls, argus.perf); } // validate results for rocsolver-test // using n * machine_precision as tolerance if(argus.unit_check) ROCSOLVER_TEST_CHECK(T, max_error, n); // output results for rocsolver-bench if(argus.timing) { if(!argus.perf) { std::cerr << "\n============================================\n"; std::cerr << "Arguments:\n"; std::cerr << "============================================\n"; rocsolver_bench_output("uplo", "n", "nrhs", "lda", "ldb"); rocsolver_bench_output(uploC, n, nrhs, lda, ldb); 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(); }