GitHub-Proxy/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2025-05-02 08:44:12 +08:00
Code Issues Packages Projects Releases Wiki Activity

Use numext::ceil instead of std::ceil

Browse Source
This commit is contained in:
Benoit Steiner 2016-04-19 14:33:30 -07:00
parent a278414d1b
commit 884c075058
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
unsupported/Eigen/CXX11/src/Tensor/TensorReductionCuda.h
View File

@ -136,7 +136,7 @@ struct FullReducer<Self, Op, GpuDevice, Vectorizable> {
const Index num_coeffs = array_prod(self.m_impl.dimensions()); const Index num_coeffs = array_prod(self.m_impl.dimensions());
const int block_size = 256; const int block_size = 256;
const int num_per_thread = 128; const int num_per_thread = 128;
const int num_blocks = std::ceil(static_cast<float>(num_coeffs) / (block_size * num_per_thread)); const int num_blocks = numext::ceil(static_cast<float>(num_coeffs) / (block_size * num_per_thread));
if (num_blocks > 1) { if (num_blocks > 1) {
// We initialize the outputs outside the reduction kernel when we can't be sure that there // We initialize the outputs outside the reduction kernel when we can't be sure that there