Skip to content

Fix for incorrect result in reduction over axis=0 #1392

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
oleksandr-pavlyk merged 2 commits into from
Sep 6, 2023
Merged

Fix for incorrect result in reduction over axis=0 #1392

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
36bc04a
Restricted use of reduce_over_axis0 special kernels
oleksandr-pavlyk Sep 6, 2023
8d8ef0b
Add test based on gh-1391
oleksandr-pavlyk Sep 6, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 3 additions & 1 deletion dpctl/tensor/libtensor/source/sum_reductions.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -218,7 +218,9 @@ std::pair<sycl::event, sycl::event> py_sum_over_axis(
return std::make_pair(keep_args_event, sum_over_axis_contig_ev);
}
}
else if (is_src_f_contig & is_dst_c_contig) {
else if (is_src_f_contig &&
((is_dst_c_contig && dst_nd == 1) || dst.is_f_contiguous()))
{
auto fn = sum_over_axis0_contig_atomic_dispatch_table[src_typeid]
[dst_typeid];
if (fn != nullptr) {
Expand Down
15 changes: 15 additions & 0 deletions dpctl/tests/test_tensor_sum.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -172,3 +172,18 @@ def test_largish_reduction(arg_dtype, n):

assert dpt.all(dpt.equal(y1, y2))
assert dpt.all(dpt.equal(y1, n * m))


def test_axis0_bug():
"gh-1391"
get_queue_or_skip()

sh = (1, 2, 3)
a = dpt.arange(sh[0] * sh[1] * sh[2], dtype="i4")
a = dpt.reshape(a, sh)
aT = dpt.permute_dims(a, (2, 1, 0))

s = dpt.sum(aT, axis=2)
expected = dpt.asarray([[0, 3], [1, 4], [2, 5]])

assert dpt.all(s == expected)