summaryrefslogtreecommitdiff
path: root/src/tuning/kernels/transpose_fast.hpp
diff options
context:
space:
mode:
Diffstat (limited to 'src/tuning/kernels/transpose_fast.hpp')
-rw-r--r--src/tuning/kernels/transpose_fast.hpp9
1 files changed, 9 insertions, 0 deletions
diff --git a/src/tuning/kernels/transpose_fast.hpp b/src/tuning/kernels/transpose_fast.hpp
index e8917ad2..c6e3f98d 100644
--- a/src/tuning/kernels/transpose_fast.hpp
+++ b/src/tuning/kernels/transpose_fast.hpp
@@ -79,6 +79,15 @@ TunerSettings TransposeGetTunerSettings(const int, const Arguments<T> &args) {
template <typename T>
void TransposeTestValidArguments(const int, const Arguments<T> &) { }
std::vector<Constraint> TransposeSetConstraints(const int) { return {}; }
+template <typename T>
+LocalMemSizeInfo TransposeComputeLocalMemSize(const int) {
+ return {
+ [] (std::vector<size_t> v) -> size_t {
+ return GetBytes(PrecisionValue<T>()) * v[1] * (v[1] * v[0]) * (v[0] + v[2]);
+ },
+ {"TRA_DIM", "TRA_WPT", "TRA_PAD"}
+ };
+}
// Sets the kernel's arguments
template <typename T>