GitHub-Proxy/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2025-06-04 18:54:00 +08:00
Code Issues Packages Projects Releases Wiki Activity

Removed unnecessary thread synchronization

Browse Source
This commit is contained in:
Benoit Steiner 2016-05-13 10:49:38 -07:00
parent 7aa3557d31
commit c4fc8b70ec
1 changed files with 2 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
unsupported/Eigen/CXX11/src/Tensor/TensorReductionCuda.h
View File

@ -177,7 +177,7 @@ static __global__ void FullReductionKernelHalfFloat(Reducer reducer, const Self
}
half2 accum = reducer.template initializePacket<half2>();
Index max_iter = numext::mini<Index>((num_coeffs - first_index) / 2, NumPerThread*BlockSize / 2);
const Index max_iter = numext::mini<Index>((num_coeffs - first_index) / 2, NumPerThread*BlockSize / 2);
for (Index i = 0; i < max_iter; i += BlockSize) {
const Index index = first_index + 2*i;
eigen_assert(index + 1 < num_coeffs);
@ -357,8 +357,6 @@ __global__ void InnerReductionKernel(Reducer reducer, const Self input, Index nu
atomicReduce(&(output[row]), reduced_val, reducer);
}
}
__syncthreads();
}
}