Skip to content

[xpu][test] Port test_float8_utils.py and test_hqq_affine.py UT files to intel XPU #3424

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

Open
zxd1997066 wants to merge 3 commits into
base: main
Choose a base branch
from
+12 −6
Open

[xpu][test] Port test_float8_utils.py and test_hqq_affine.py UT files to intel XPU #3424

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
33d0629
add test/float8/test_float8_utils.py
zxd1997066 Dec 3, 2025
b1a6b82
test/hqq/test_hqq_affine.py
zxd1997066 Dec 3, 2025
6c2ca4b
test/hqq/test_hqq_affine.py
zxd1997066 Dec 3, 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
9 changes: 6 additions & 3 deletions test/float8/test_float8_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,11 +10,14 @@

from torchao.float8.float8_utils import _round_scale_down_to_power_of_2
from torchao.testing.utils import skip_if_rocm
from torchao.utils import get_current_accelerator_device

_DEVICE = get_current_accelerator_device()


# source for notable single-precision cases:
# https://en.wikipedia.org/wiki/Single-precision_floating-point_format
@unittest.skipIf(not torch.cuda.is_available(), "CUDA not available")
@unittest.skipIf(not torch.accelerator.is_available(), "GPU not available")
@pytest.mark.parametrize(
"test_case",
[
Expand All @@ -38,8 +41,8 @@ def test_round_scale_down_to_power_of_2_valid_inputs(
):
test_case_name, input, expected_result = test_case
input_tensor, expected_tensor = (
torch.tensor(input, dtype=torch.float32).cuda(),
torch.tensor(expected_result, dtype=torch.float32).cuda(),
torch.tensor(input, dtype=torch.float32).to(_DEVICE),
torch.tensor(expected_result, dtype=torch.float32).to(_DEVICE),
)
result = _round_scale_down_to_power_of_2(input_tensor)

Expand Down
9 changes: 6 additions & 3 deletions test/hqq/test_hqq_affine.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,12 +14,14 @@
ZeroPointDomain,
quantize_,
)
from torchao.testing.utils import skip_if_rocm
from torchao.testing.utils import skip_if_rocm, skip_if_xpu
from torchao.utils import get_current_accelerator_device

cuda_available = torch.cuda.is_available()
cuda_available = torch.accelerator.is_available()
_DEVICE = get_current_accelerator_device()

# Parameters
device = "cuda:0"
device = f"{_DEVICE}:0"
compute_dtype = torch.bfloat16
group_size = 64
mapping_type = MappingType.ASYMMETRIC
Expand Down Expand Up @@ -114,6 +116,7 @@ def test_hqq_plain_5bit(self):
)

@skip_if_rocm("ROCm enablement in progress")
@skip_if_xpu("XPU enablement in progress")
def test_hqq_plain_4bit(self):
self._test_hqq(
dtype=torch.uint4,
Expand Down