Skip to content

Commit

Permalink
Inlining these helper functions leads to better perf
Browse files Browse the repository at this point in the history
  • Loading branch information
@ghugo83
ghugo83 committed Mar 29, 2021
1 parent 36f8c3d commit b4c3682
Showing 1 changed file with 28 additions and 28 deletions.
56 changes: 28 additions & 28 deletions src/alpaka/AlpakaCore/alpakaWorkDivHelper.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -153,10 +153,10 @@ namespace cms {
* Indexes are local to the BLOCK.
*/
template <typename T_Acc, typename Func>
ALPAKA_FN_ACC void for_each_element_in_thread_1D_index_in_block(const T_Acc& acc,
const uint32_t maxNumberOfElements,
const uint32_t elementIdxShift,
Func&& func) {
ALPAKA_FN_ACC ALPAKA_FN_INLINE void for_each_element_in_thread_1D_index_in_block(const T_Acc& acc,
const uint32_t maxNumberOfElements,
const uint32_t elementIdxShift,
Func&& func) {
const auto& [firstElementIdx, endElementIdx] = cms::alpakatools::element_index_range_in_block_truncated(
acc, Vec1::all(maxNumberOfElements), Vec1::all(elementIdxShift));

Expand All @@ -169,9 +169,9 @@ namespace cms {
* Overload for elementIdxShift = 0
*/
template <typename T_Acc, typename Func>
ALPAKA_FN_ACC void for_each_element_in_thread_1D_index_in_block(const T_Acc& acc,
const uint32_t maxNumberOfElements,
Func&& func) {
ALPAKA_FN_ACC ALPAKA_FN_INLINE void for_each_element_in_thread_1D_index_in_block(const T_Acc& acc,
const uint32_t maxNumberOfElements,
Func&& func) {
const uint32_t elementIdxShift = 0;
cms::alpakatools::for_each_element_in_thread_1D_index_in_block(
acc, maxNumberOfElements, elementIdxShift, std::forward<Func>(func));
Expand All @@ -183,10 +183,10 @@ namespace cms {
* Indexes are expressed in GRID 'frame-of-reference'.
*/
template <typename T_Acc, typename Func>
ALPAKA_FN_ACC void for_each_element_in_thread_1D_index_in_grid(const T_Acc& acc,
const uint32_t maxNumberOfElements,
uint32_t elementIdxShift,
Func&& func) {
ALPAKA_FN_ACC ALPAKA_FN_INLINE void for_each_element_in_thread_1D_index_in_grid(const T_Acc& acc,
const uint32_t maxNumberOfElements,
uint32_t elementIdxShift,
Func&& func) {
// Take into account the block index in grid to compute the element indices.
const uint32_t blockIdxInGrid(alpaka::idx::getIdx<alpaka::Grid, alpaka::Blocks>(acc)[0u]);
const uint32_t blockDimension(alpaka::workdiv::getWorkDiv<alpaka::Block, alpaka::Elems>(acc)[0u]);
Expand All @@ -199,9 +199,9 @@ namespace cms {
* Overload for elementIdxShift = 0
*/
template <typename T_Acc, typename Func>
ALPAKA_FN_ACC void for_each_element_in_thread_1D_index_in_grid(const T_Acc& acc,
const uint32_t maxNumberOfElements,
Func&& func) {
ALPAKA_FN_ACC ALPAKA_FN_INLINE void for_each_element_in_thread_1D_index_in_grid(const T_Acc& acc,
const uint32_t maxNumberOfElements,
Func&& func) {
const uint32_t elementIdxShift = 0;
cms::alpakatools::for_each_element_in_thread_1D_index_in_grid(
acc, maxNumberOfElements, elementIdxShift, std::forward<Func>(func));
Expand All @@ -218,10 +218,10 @@ namespace cms {
* Indexes are local to the BLOCK.
*/
template <typename T_Acc, typename Func>
ALPAKA_FN_ACC void for_each_element_1D_block_stride(const T_Acc& acc,
const uint32_t maxNumberOfElements,
const uint32_t elementIdxShift,
Func&& func) {
ALPAKA_FN_ACC ALPAKA_FN_INLINE void for_each_element_1D_block_stride(const T_Acc& acc,
const uint32_t maxNumberOfElements,
const uint32_t elementIdxShift,
Func&& func) {
// Get thread / element indices in block.
const auto& [firstElementIdxNoStride, endElementIdxNoStride] =
cms::alpakatools::element_index_range_in_block(acc, Vec1::all(elementIdxShift));
Expand All @@ -244,9 +244,9 @@ namespace cms {
* Overload for elementIdxShift = 0
*/
template <typename T_Acc, typename Func>
ALPAKA_FN_ACC void for_each_element_1D_block_stride(const T_Acc& acc,
const uint32_t maxNumberOfElements,
Func&& func) {
ALPAKA_FN_ACC ALPAKA_FN_INLINE void for_each_element_1D_block_stride(const T_Acc& acc,
const uint32_t maxNumberOfElements,
Func&& func) {
const uint32_t elementIdxShift = 0;
cms::alpakatools::for_each_element_1D_block_stride(
acc, maxNumberOfElements, elementIdxShift, std::forward<Func>(func));
Expand All @@ -259,10 +259,10 @@ namespace cms {
* Indexes are local to the GRID.
*/
template <typename T_Acc, typename Func>
ALPAKA_FN_ACC void for_each_element_1D_grid_stride(const T_Acc& acc,
const uint32_t maxNumberOfElements,
const uint32_t elementIdxShift,
Func&& func) {
ALPAKA_FN_ACC ALPAKA_FN_INLINE void for_each_element_1D_grid_stride(const T_Acc& acc,
const uint32_t maxNumberOfElements,
const uint32_t elementIdxShift,
Func&& func) {
Vec1 elementIdxShiftVec = Vec1::all(elementIdxShift);

// Get thread / element indices in block.
Expand All @@ -287,9 +287,9 @@ namespace cms {
* Overload for elementIdxShift = 0
*/
template <typename T_Acc, typename Func>
ALPAKA_FN_ACC void for_each_element_1D_grid_stride(const T_Acc& acc,
const uint32_t maxNumberOfElements,
Func&& func) {
ALPAKA_FN_ACC ALPAKA_FN_INLINE void for_each_element_1D_grid_stride(const T_Acc& acc,
const uint32_t maxNumberOfElements,
Func&& func) {
const uint32_t elementIdxShift = 0;
cms::alpakatools::for_each_element_1D_grid_stride(
acc, maxNumberOfElements, elementIdxShift, std::forward<Func>(func));
Expand Down

0 comments on commit b4c3682

Please sign in to comment.