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

Added a benchmark to measure the performance of full reductions of 16 bit floats

Browse Source
This commit is contained in:
Benoit Steiner 2016-05-05 14:15:11 -07:00
parent 28d5572658
commit f81e413180
2 changed files with 2 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
bench/tensors/tensor_benchmarks.h
View File

@ -368,7 +368,7 @@ template <typename Device, typename T> class BenchmarkSuite {
const TensorMap<Tensor<T, 2, 0, TensorIndex>, Eigen::Aligned> B( const TensorMap<Tensor<T, 2, 0, TensorIndex>, Eigen::Aligned> B(
b_, input_size); b_, input_size);
Eigen::array<TensorIndex, 0> output_size; Eigen::array<TensorIndex, 0> output_size;
TensorMap<Tensor<float, 0, 0, TensorIndex>, Eigen::Aligned> C( TensorMap<Tensor<T, 0, 0, TensorIndex>, Eigen::Aligned> C(
c_, output_size); c_, output_size);
StartBenchmarkTiming(); StartBenchmarkTiming();

1
bench/tensors/tensor_benchmarks_fp16_gpu.cu
View File

@ -33,6 +33,7 @@ BM_FuncGPU(algebraicFunc);
BM_FuncGPU(transcendentalFunc); BM_FuncGPU(transcendentalFunc);
BM_FuncGPU(rowReduction); BM_FuncGPU(rowReduction);
BM_FuncGPU(colReduction); BM_FuncGPU(colReduction);
BM_FuncGPU(fullReduction);
// Contractions // Contractions