mirror/eigen
2
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2024-12-27 07:29:52 +08:00
Code Issues Packages Projects Releases Wiki Activity

Backed out changeset 01197e4452

Browse Source
This commit is contained in:
Rasmus Munk Larsen 2018-09-13 16:20:57 -07:00
parent b3f4c067d9
commit 0db590d22d
2 changed files with 6 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

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

@ -1205,6 +1205,7 @@ struct TensorEvaluator<const TensorReductionOp<Op, Dims, ArgType, MakePointer_>,
const Index max_coeff_count,
const DSizes<Index, NumInputDims>& input_slice_sizes,
DSizes<Index, NumInputDims>* target_input_block_sizes) const {
typedef typename internal::packet_traits<Scalar>::type Packet;
typedef internal::BlockReducer<Self, Op> BlockReducer;
// TODO(andydavis) Compute reducer overhead correctly for the case where
// we are preserving the inner most dimension, and a single reducer
@ -1232,6 +1233,7 @@ struct TensorEvaluator<const TensorReductionOp<Op, Dims, ArgType, MakePointer_>,
} else if (!first_preserved_dim_allocated) {
// TODO(andydavis) Include output block size in this L1 working set
// calculation.
const Index allocated = max_coeff_count - coeff_to_allocate;
const Index alloc_size = numext::maxi(
static_cast<Index>(1), coeff_to_allocate / reducer_overhead);
(*target_input_block_sizes)[dim] =

8
unsupported/test/cxx11_tensor_executor.cpp
View File

@ -27,7 +27,7 @@ static array<Index, NumDims> RandomDims(int min_dim = 1, int max_dim = 20) {
dims[i] = internal::random<int>(min_dim, max_dim);
}
return dims;
}
};
template <typename T, int NumDims, typename Device, bool Vectorizable,
bool Tileable, int Layout>
@ -117,7 +117,7 @@ static void test_execute_broadcasting(Device d)
for (Index i = 0; i < dst.dimensions().TotalSize(); ++i) {
VERIFY_IS_EQUAL(dst.coeff(i), golden.coeff(i));
}
}
};
template <typename T, int NumDims, typename Device, bool Vectorizable,
bool Tileable, int Layout>
@ -155,7 +155,7 @@ static void test_execute_chipping_rvalue(Device d) {
TEST_CHIPPING(5)
#undef TEST_CHIPPING
}
};
template <typename T, int NumDims, typename Device, bool Vectorizable,
bool Tileable, int Layout>
@ -207,7 +207,7 @@ static void test_execute_chipping_lvalue(Device d) {
TEST_CHIPPING(5)
#undef TEST_CHIPPING
}
};
template <typename T, int NumDims, typename Device, bool Vectorizable,
bool Tileable, int Layout>