|
| 1 | +//==---- svm_gather_scatter.cpp - DPC++ ESIMD svm gather/scatter test -----==// |
| 2 | +// |
| 3 | +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
| 4 | +// See https://llvm.org/LICENSE.txt for license information. |
| 5 | +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
| 6 | +// |
| 7 | +//===----------------------------------------------------------------------===// |
| 8 | +// REQUIRES: gpu |
| 9 | +// UNSUPPORTED: cuda || hip |
| 10 | +// RUN: %clangxx -fsycl %s -o %t.out |
| 11 | +// RUN: %HOST_RUN_PLACEHOLDER %t.out |
| 12 | +// RUN: %GPU_RUN_PLACEHOLDER %t.out |
| 13 | + |
| 14 | +// Regression test for SVM gather/scatter API. |
| 15 | + |
| 16 | +#include "../esimd_test_utils.hpp" |
| 17 | + |
| 18 | +#include <CL/sycl.hpp> |
| 19 | +#include <CL/sycl/builtins_esimd.hpp> |
| 20 | +#include <algorithm> |
| 21 | +#include <array> |
| 22 | +#include <iostream> |
| 23 | + |
| 24 | +#include <sycl/ext/intel/experimental/esimd.hpp> |
| 25 | + |
| 26 | +using namespace cl::sycl; |
| 27 | +using namespace sycl::ext::intel::experimental; |
| 28 | +using namespace sycl::ext::intel::experimental::esimd; |
| 29 | + |
| 30 | +template <typename T, int N> bool test(queue &Q) { |
| 31 | + std::cout << " Running " << typeid(T).name() << " test, N=" << N << "...\n"; |
| 32 | + |
| 33 | + struct Deleter { |
| 34 | + queue Q; |
| 35 | + void operator()(T *Ptr) { |
| 36 | + if (Ptr) { |
| 37 | + sycl::free(Ptr, Q); |
| 38 | + } |
| 39 | + } |
| 40 | + }; |
| 41 | + |
| 42 | + std::unique_ptr<T, Deleter> Ptr1(sycl::malloc_shared<T>(N, Q), Deleter{Q}); |
| 43 | + std::unique_ptr<T, Deleter> Ptr2(sycl::malloc_shared<T>(N, Q), Deleter{Q}); |
| 44 | + |
| 45 | + T *Src = Ptr1.get(); |
| 46 | + T *Dst = Ptr2.get(); |
| 47 | + |
| 48 | + for (int I = 0; I < N; ++I) { |
| 49 | + Src[I] = I + 1; |
| 50 | + Dst[I] = 0; |
| 51 | + } |
| 52 | + |
| 53 | + try { |
| 54 | + Q.submit([&](handler &CGH) { |
| 55 | + CGH.parallel_for(sycl::range<1>{1}, [=](id<1>) SYCL_ESIMD_KERNEL { |
| 56 | + simd<uint32_t, N> Offsets(0u, sizeof(T)); |
| 57 | + scatter<T, N>(Dst, Offsets, gather<T, N>(Src, Offsets)); |
| 58 | + }); |
| 59 | + }).wait(); |
| 60 | + } catch (cl::sycl::exception const &E) { |
| 61 | + std::cout << "ERROR. SYCL exception caught: " << E.what() << std::endl; |
| 62 | + return false; |
| 63 | + } |
| 64 | + |
| 65 | + unsigned NumErrs = 0; |
| 66 | + for (int I = 0; I < N; ++I) |
| 67 | + if (Dst[I] != Src[I]) |
| 68 | + if (++NumErrs <= 10) |
| 69 | + std::cout << "failed at " << I << ": " << Dst[I] |
| 70 | + << " (Dst) != " << Src[I] << " (Src)\n"; |
| 71 | + |
| 72 | + std::cout << (NumErrs == 0 ? " Passed\n" : " FAILED\n"); |
| 73 | + return NumErrs == 0; |
| 74 | +} |
| 75 | + |
| 76 | +int main(void) { |
| 77 | + queue Q(esimd_test::ESIMDSelector{}, esimd_test::createExceptionHandler()); |
| 78 | + auto Dev = Q.get_device(); |
| 79 | + std::cout << "Running on " << Dev.get_info<info::device::name>() << "\n"; |
| 80 | + |
| 81 | + bool Pass = true; |
| 82 | + |
| 83 | + Pass &= test<int8_t, 8>(Q); |
| 84 | + Pass &= test<int8_t, 16>(Q); |
| 85 | + Pass &= test<int8_t, 32>(Q); |
| 86 | + |
| 87 | + Pass &= test<int16_t, 8>(Q); |
| 88 | + Pass &= test<int16_t, 16>(Q); |
| 89 | + Pass &= test<int16_t, 32>(Q); |
| 90 | + |
| 91 | + Pass &= test<int32_t, 8>(Q); |
| 92 | + Pass &= test<int32_t, 16>(Q); |
| 93 | + Pass &= test<int32_t, 32>(Q); |
| 94 | + |
| 95 | + std::cout << (Pass ? "Test Passed\n" : "Test FAILED\n"); |
| 96 | + return Pass ? 0 : 1; |
| 97 | +} |
0 commit comments