Skip to content

Commit ee7d6fb

Browse files
Stylistic changes in accumulators.hpp
Use explicit type instead of auto for readability, save common sub-expression to a temporary variable. Move std::array variable declaration closer to its first use.
1 parent 9b3341a commit ee7d6fb

File tree

1 file changed

+9
-8
lines changed

1 file changed

+9
-8
lines changed

dpctl/tensor/libtensor/include/kernels/accumulators.hpp

Lines changed: 9 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -215,16 +215,17 @@ inclusive_scan_base_step(sycl::queue &exec_q,
215215
const size_t gid = it.get_global_id(0);
216216
const size_t lid = it.get_local_id(0);
217217

218-
const size_t iter_gid = gid / (acc_groups * wg_size);
219-
const size_t chunk_gid = gid - (iter_gid * acc_groups * wg_size);
218+
const size_t reduce_chunks = acc_groups * wg_size;
219+
const size_t iter_gid = gid / reduce_chunks;
220+
const size_t chunk_gid = gid - (iter_gid * reduce_chunks);
220221

221-
std::array<outputT, n_wi> local_iscan;
222-
223-
size_t i = chunk_gid * n_wi;
222+
const size_t i = chunk_gid * n_wi;
224223
const auto &iter_offsets = iter_indexer(iter_gid);
225224
const auto &inp_iter_offset = iter_offsets.get_first_offset();
226225
const auto &out_iter_offset = iter_offsets.get_second_offset();
227226

227+
std::array<outputT, n_wi> local_iscan;
228+
228229
#pragma unroll
229230
for (nwiT m_wi = 0; m_wi < n_wi; ++m_wi) {
230231
const size_t i_m_wi = i + m_wi;
@@ -324,7 +325,7 @@ sycl::event inclusive_scan_iter_1d(sycl::queue &exec_q,
324325
std::vector<sycl::event> &host_tasks,
325326
const std::vector<sycl::event> &depends = {})
326327
{
327-
ScanOpT scan_op = ScanOpT();
328+
ScanOpT scan_op{};
328329
constexpr outputT identity = su_ns::Identity<ScanOpT, outputT>::value;
329330

330331
constexpr size_t _iter_nelems = 1;
@@ -352,9 +353,9 @@ sycl::event inclusive_scan_iter_1d(sycl::queue &exec_q,
352353
size_t n_groups_ = n_groups;
353354
size_t temp_size = 0;
354355
while (n_groups_ > 1) {
355-
const auto this_size = (n_groups_ - 1);
356+
const size_t this_size = (n_groups_ - 1);
356357
temp_size += this_size;
357-
n_groups_ = ceiling_quotient<size_t>(this_size, chunk_size);
358+
n_groups_ = ceiling_quotient(this_size, chunk_size);
358359
}
359360

360361
// allocate

0 commit comments

Comments
 (0)