File size: 3,536 Bytes
8ae5fc5 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 |
#include <unittest/unittest.h>
#include <thrust/generate.h>
#include <thrust/execution_policy.h>
template<typename T>
struct return_value
{
T val;
return_value(void){}
return_value(T v):val(v){}
__host__ __device__
T operator()(void){ return val; }
};
#ifdef THRUST_TEST_DEVICE_SIDE
template<typename ExecutionPolicy, typename Iterator, typename Function>
__global__
void generate_kernel(ExecutionPolicy exec, Iterator first, Iterator last, Function f)
{
thrust::generate(exec, first, last, f);
}
template<typename T, typename ExecutionPolicy>
void TestGenerateDevice(ExecutionPolicy exec, const size_t n)
{
thrust::host_vector<T> h_result(n);
thrust::device_vector<T> d_result(n);
T value = 13;
return_value<T> f(value);
thrust::generate(h_result.begin(), h_result.end(), f);
generate_kernel<<<1,1>>>(exec, d_result.begin(), d_result.end(), f);
{
cudaError_t const err = cudaDeviceSynchronize();
ASSERT_EQUAL(cudaSuccess, err);
}
ASSERT_EQUAL(h_result, d_result);
}
template<typename T>
void TestGenerateDeviceSeq(const size_t n)
{
TestGenerateDevice<T>(thrust::seq, n);
}
DECLARE_VARIABLE_UNITTEST(TestGenerateDeviceSeq);
template<typename T>
void TestGenerateDeviceDevice(const size_t n)
{
TestGenerateDevice<T>(thrust::device, n);
}
DECLARE_VARIABLE_UNITTEST(TestGenerateDeviceDevice);
#endif
void TestGenerateCudaStreams()
{
thrust::device_vector<int> result(5);
int value = 13;
return_value<int> f(value);
cudaStream_t s;
cudaStreamCreate(&s);
thrust::generate(thrust::cuda::par.on(s), result.begin(), result.end(), f);
cudaStreamSynchronize(s);
ASSERT_EQUAL(result[0], value);
ASSERT_EQUAL(result[1], value);
ASSERT_EQUAL(result[2], value);
ASSERT_EQUAL(result[3], value);
ASSERT_EQUAL(result[4], value);
cudaStreamDestroy(s);
}
DECLARE_UNITTEST(TestGenerateCudaStreams);
#ifdef THRUST_TEST_DEVICE_SIDE
template<typename ExecutionPolicy, typename Iterator, typename Size, typename Function>
__global__
void generate_n_kernel(ExecutionPolicy exec, Iterator first, Size n, Function f)
{
thrust::generate_n(exec, first, n, f);
}
template<typename T, typename ExecutionPolicy>
void TestGenerateNDevice(ExecutionPolicy exec, const size_t n)
{
thrust::host_vector<T> h_result(n);
thrust::device_vector<T> d_result(n);
T value = 13;
return_value<T> f(value);
thrust::generate_n(h_result.begin(), h_result.size(), f);
generate_n_kernel<<<1,1>>>(exec, d_result.begin(), d_result.size(), f);
{
cudaError_t const err = cudaDeviceSynchronize();
ASSERT_EQUAL(cudaSuccess, err);
}
ASSERT_EQUAL(h_result, d_result);
}
template<typename T>
void TestGenerateNDeviceSeq(const size_t n)
{
TestGenerateNDevice<T>(thrust::seq, n);
}
DECLARE_VARIABLE_UNITTEST(TestGenerateNDeviceSeq);
template<typename T>
void TestGenerateNDeviceDevice(const size_t n)
{
TestGenerateNDevice<T>(thrust::device, n);
}
DECLARE_VARIABLE_UNITTEST(TestGenerateNDeviceDevice);
#endif
void TestGenerateNCudaStreams()
{
thrust::device_vector<int> result(5);
int value = 13;
return_value<int> f(value);
cudaStream_t s;
cudaStreamCreate(&s);
thrust::generate_n(thrust::cuda::par.on(s), result.begin(), result.size(), f);
cudaStreamSynchronize(s);
ASSERT_EQUAL(result[0], value);
ASSERT_EQUAL(result[1], value);
ASSERT_EQUAL(result[2], value);
ASSERT_EQUAL(result[3], value);
ASSERT_EQUAL(result[4], value);
cudaStreamDestroy(s);
}
DECLARE_UNITTEST(TestGenerateNCudaStreams);
|