devtools/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2026-04-10 11:34:33 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix bugs exposed by enabling GPU asserts.

Browse Source
This commit is contained in:
Antonio Sánchez
2023-01-27 21:43:00 +00:00
committed by Rasmus Munk Larsen
parent ab8725d947
commit 17ae83a966
5 changed files with 19 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
unsupported/test/cxx11_tensor_argmax_gpu.cu
View File

@@ -23,8 +23,8 @@ template <int Layout>
void test_gpu_simple_argmax()
{
Tensor<double, 3, Layout> in(Eigen::array<DenseIndex, 3>(72,53,97));
Tensor<DenseIndex, 1, Layout> out_max(Eigen::array<DenseIndex, 1>(1));
Tensor<DenseIndex, 1, Layout> out_min(Eigen::array<DenseIndex, 1>(1));
Tensor<DenseIndex, 0, Layout> out_max;
Tensor<DenseIndex, 0, Layout> out_min;
in.setRandom();
in *= in.constant(100.0);
in(0, 0, 0) = -1000.0;
@@ -46,8 +46,8 @@ void test_gpu_simple_argmax()
Eigen::GpuDevice gpu_device(&stream);
Eigen::TensorMap<Eigen::Tensor<double, 3, Layout>, Aligned > gpu_in(d_in, Eigen::array<DenseIndex, 3>(72,53,97));
Eigen::TensorMap<Eigen::Tensor<DenseIndex, 1, Layout>, Aligned > gpu_out_max(d_out_max, Eigen::array<DenseIndex, 1>(1));
Eigen::TensorMap<Eigen::Tensor<DenseIndex, 1, Layout>, Aligned > gpu_out_min(d_out_min, Eigen::array<DenseIndex, 1>(1));
Eigen::TensorMap<Eigen::Tensor<DenseIndex, 0, Layout>, Aligned > gpu_out_max(d_out_max);
Eigen::TensorMap<Eigen::Tensor<DenseIndex, 0, Layout>, Aligned > gpu_out_min(d_out_min);
gpu_out_max.device(gpu_device) = gpu_in.argmax();
gpu_out_min.device(gpu_device) = gpu_in.argmin();
@@ -56,8 +56,8 @@ void test_gpu_simple_argmax()
assert(gpuMemcpyAsync(out_min.data(), d_out_min, out_bytes, gpuMemcpyDeviceToHost, gpu_device.stream()) == gpuSuccess);
assert(gpuStreamSynchronize(gpu_device.stream()) == gpuSuccess);
VERIFY_IS_EQUAL(out_max(Eigen::array<DenseIndex, 1>(0)), 72*53*97 - 1);
VERIFY_IS_EQUAL(out_min(Eigen::array<DenseIndex, 1>(0)), 0);
VERIFY_IS_EQUAL(out_max(), 72*53*97 - 1);
VERIFY_IS_EQUAL(out_min(), 0);
gpuFree(d_in);
gpuFree(d_out_max);

6
unsupported/test/cxx11_tensor_gpu.cu
View File

@@ -1100,9 +1100,9 @@ void test_gpu_erfc(const Scalar stddev)
template <typename Scalar>
void test_gpu_ndtri()
{
Tensor<Scalar, 1> in_x(8);
Tensor<Scalar, 1> out(8);
Tensor<Scalar, 1> expected_out(8);
Tensor<Scalar, 1> in_x(9);
Tensor<Scalar, 1> out(9);
Tensor<Scalar, 1> expected_out(9);
out.setZero();
in_x(0) = Scalar(1);