GitHub-Proxy/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2025-08-12 19:59:05 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fixed a boundary condition bug in the outer reduction kernel

Browse Source
This commit is contained in:
Benoit Steiner 2016-01-14 09:29:48 -08:00
parent 9f013a9d86
commit 8fe2532e70
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

@ -241,7 +241,7 @@ __global__ void OuterReductionKernel(Reducer reducer, const Self input, Index nu
} }
// Do the reduction. // Do the reduction.
const Index max_iter = divup<Index>(num_coeffs_to_reduce, NumPerThread) * num_preserved_coeffs; const Index max_iter = num_preserved_coeffs * numext::maxi<Index>(1, (num_coeffs_to_reduce - NumPerThread + 1));
for (Index i = thread_id; i < max_iter; i += num_threads) { for (Index i = thread_id; i < max_iter; i += num_threads) {
const Index input_col = i % num_preserved_coeffs; const Index input_col = i % num_preserved_coeffs;
const Index input_row = (i / num_preserved_coeffs) * NumPerThread; const Index input_row = (i / num_preserved_coeffs) * NumPerThread;