Skip to content

[KUNLUNXIN] open threshold in flaggems benchmark for "index index_add instance_norm kron linspace nll_loss" #1143

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
nianqi-tian merged 2 commits into from
Dec 3, 2025
Merged

[KUNLUNXIN] open threshold in flaggems benchmark for "index index_add instance_norm kron linspace nll_loss" #1143

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
d6890be
[KUNLUNXIN] open threshold in flaggems benchmark for "index index_add…
Dec 2, 2025
3b87ea5
[KUNLUNXIN] open threshold in flaggems benchmark for "index index_add…
Dec 2, 2025
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
3 changes: 0 additions & 3 deletions benchmark/test_select_and_slice_perf.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -337,7 +337,6 @@ def index_add_gbps(bench_fn_args, latency):
return io_amount * 1e-9 / (latency * 1e-3)


@pytest.mark.skipif(vendor_name == "kunlunxin", reason="RESULT TODOFIX")
@pytest.mark.index_add
def test_index_add_perf():
def index_add_input_fn(shape, dtype, device):
Expand All @@ -361,7 +360,6 @@ def index_add_input_fn(shape, dtype, device):
bench.run()


@pytest.mark.skipif(vendor_name == "kunlunxin", reason="RESULT TODOFIX")
@pytest.mark.index_add_
def test_index_add__perf():
def index_add__input_fn(shape, dtype, device):
Expand Down Expand Up @@ -537,7 +535,6 @@ def index_input_fn(shapes, dtype, device):
yield inp, indices


@pytest.mark.skipif(vendor_name == "kunlunxin", reason="RESULT TODOFIX")
@pytest.mark.index
def test_index_acc_perf():
gems_op = flag_gems.index
Expand Down
6 changes: 5 additions & 1 deletion benchmark/test_special_perf.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,7 @@
GenericBenchmark2DOnly,
GenericBenchmarkExcluse1D,
GenericBenchmarkExcluse3D,
SkipVersion,
generate_tensor_input,
vendor_name,
)
Expand Down Expand Up @@ -389,7 +390,10 @@ def diagonal_backward_input_fn(shape, dtype, device):
bench.run()


@pytest.mark.skipif(vendor_name == "kunlunxin", reason="RESULT TODOFIX")
@pytest.mark.skipif(
vendor_name == "kunlunxin" and SkipVersion("torch", "<2.5"),
reason="only support torch >= 2.5.",
)
@pytest.mark.skipif(vendor_name == "cambricon", reason="TODOFIX")
@pytest.mark.kron
def test_perf_kron():
Expand Down
14 changes: 10 additions & 4 deletions benchmark/test_tensor_constructor_perf.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,7 @@
from benchmark.performance_utils import (
Config,
GenericBenchmark,
SkipVersion,
generate_tensor_input,
unary_input_fn,
vendor_name,
Expand Down Expand Up @@ -149,10 +150,15 @@ def _2D_input_fn(shape, dtype, device):
],
)
def test_tensor_constructor_benchmark(op_name, torch_op, input_fn):
if vendor_name == "kunlunxin" and op_name in [
"linspace",
]:
pytest.skip("RUNTIME TODOFIX.")
if (
vendor_name == "kunlunxin"
and SkipVersion("torch", "<2.5")
and op_name
in [
"linspace",
]
):
pytest.skip("only support torch >= 2.5.")
bench = GenericBenchmark(input_fn=input_fn, op_name=op_name, torch_op=torch_op)
bench.run()

Expand Down
Loading