GitHub-Proxy/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2025-08-16 21:55:58 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fixed existing test.

Browse Source
This commit is contained in:
Luke Iwanski 2016-11-17 17:46:55 +00:00
parent c5130dedbe
commit 7878756dea
1 changed files with 44 additions and 38 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

34
unsupported/test/cxx11_tensor_builtins_sycl.cpp
View File

@ -25,8 +25,7 @@ using Eigen::SyclDevice;
using Eigen::Tensor; using Eigen::Tensor;
using Eigen::TensorMap; using Eigen::TensorMap;
namespace std namespace std {
{
template <typename T> T rsqrt(T x) { return 1 / std::sqrt(x); } template <typename T> T rsqrt(T x) { return 1 / std::sqrt(x); }
template <typename T> T square(T x) { return x * x; } template <typename T> T square(T x) { return x * x; }
template <typename T> T cube(T x) { return x * x * x; } template <typename T> T cube(T x) { return x * x * x; }
@ -35,18 +34,25 @@ namespace std
#define TEST_UNARY_BUILTINS_FOR_SCALAR(FUNC, SCALAR) \ #define TEST_UNARY_BUILTINS_FOR_SCALAR(FUNC, SCALAR) \
{ \ { \
Tensor<SCALAR, 3> in1(tensorRange); \ Tensor<SCALAR, 3> in(tensorRange); \
Tensor<SCALAR, 3> out1(tensorRange); \ Tensor<SCALAR, 3> out(tensorRange); \
in1 = in1.random(); \ in = in.random(); \
SCALAR* gpu_data1 = static_cast<SCALAR*>(sycl_device.allocate(in1.size()*sizeof(SCALAR))); \ SCALAR *gpu_data = static_cast<SCALAR *>( \
TensorMap<Tensor<SCALAR, 3>> gpu1(gpu_data1, tensorRange); \ sycl_device.allocate(in.size() * sizeof(SCALAR))); \
sycl_device.memcpyHostToDevice(gpu_data1, in1.data(),(in1.size())*sizeof(SCALAR)); \ SCALAR *gpu_data_out = static_cast<SCALAR *>( \
gpu1.device(sycl_device) = gpu1.FUNC(); \ sycl_device.allocate(out.size() * sizeof(SCALAR))); \
sycl_device.memcpyDeviceToHost(out1.data(), gpu_data1,(out1.size())*sizeof(SCALAR)); \ TensorMap<Tensor<SCALAR, 3>> gpu(gpu_data, tensorRange); \
for (int i = 0; i < in1.size(); ++i) { \ TensorMap<Tensor<SCALAR, 3>> gpu_out(gpu_data_out, tensorRange); \
VERIFY_IS_APPROX(out1(i), std::FUNC(in1(i))); \ sycl_device.memcpyHostToDevice(gpu_data, in.data(), \
(in.size()) * sizeof(SCALAR)); \
gpu_out.device(sycl_device) = gpu.FUNC(); \
sycl_device.memcpyDeviceToHost(out.data(), gpu_data_out, \
(out.size()) * sizeof(SCALAR)); \
for (int i = 0; i < in.size(); ++i) { \
VERIFY_IS_APPROX(out(i), std::FUNC(in(i))); \
} \ } \
sycl_device.deallocate(gpu_data1); \ sycl_device.deallocate(gpu_data); \
sycl_device.deallocate(gpu_data_out); \
} }
#define TEST_UNARY_BUILTINS(SCALAR) \ #define TEST_UNARY_BUILTINS(SCALAR) \
@ -73,8 +79,8 @@ static void test_builtin_unary_sycl(const Eigen::SyclDevice &sycl_device){
TEST_UNARY_BUILTINS(float) TEST_UNARY_BUILTINS(float)
TEST_UNARY_BUILTINS(double) TEST_UNARY_BUILTINS(double)
}
}
void test_cxx11_tensor_builtins_sycl() { void test_cxx11_tensor_builtins_sycl() {
cl::sycl::gpu_selector s; cl::sycl::gpu_selector s;