Skip to content
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

Fix inductor config BC change #382

Merged
merged 6 commits into from
Jun 16, 2024
Merged

Fix inductor config BC change #382

Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
e5e7166
Fix inductor config BC change
msaroufim Jun 15, 2024
9600963
yolo
msaroufim Jun 15, 2024
ad72c8c
yolo
msaroufim Jun 15, 2024
9737bce
yolo
msaroufim Jun 15, 2024
b549939
add
msaroufim Jun 16, 2024
1e0b478
Update test_integration.py
msaroufim Jun 16, 2024
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
12 changes: 8 additions & 4 deletions test/integration/test_integration.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -873,9 +873,11 @@ def test_weight_only_quant_force_mixed_mm(self, device, dtype):
if dtype == torch.bfloat16 and torch.cuda.get_device_capability() < (8, 0):
self.skipTest("test requires SM capability of at least (8, 0).")
from torch._inductor import config
mixed_mm_key, mixed_mm_val = ("mixed_mm_choice", "triton") if TORCH_VERSION_AFTER_2_4 else ("force_mixed_mm", True)

with config.patch({
"epilogue_fusion": True,
"force_mixed_mm": True
mixed_mm_key: mixed_mm_val,
}):
for x_shape in [[2, 4], [5, 5, 5, 4], [1, 4, 4]]:
torch._dynamo.reset()
Expand All @@ -887,7 +889,7 @@ def test_weight_only_quant_force_mixed_mm(self, device, dtype):
m_c = torch.compile(m, mode="max-autotune")
y_wo, (code,) = run_and_get_code(m_c, x)
sqnr = compute_error(y_ref, y_wo)
self.assertGreaterEqual(sqnr, 42.75)
self.assertGreaterEqual(sqnr, 42.50)
if device == "cuda":
self.assertTrue("mixed_mm" in code, f"got code: {code}")

Expand All @@ -900,9 +902,11 @@ def test_weight_only_quant_use_mixed_mm(self, device, dtype):
self.skipTest("test requires SM capability of at least (8, 0).")
torch.manual_seed(0)
from torch._inductor import config
mixed_mm_key, mixed_mm_val = ("mixed_mm_choice", "triton") if TORCH_VERSION_AFTER_2_4 else ("force_mixed_mm", True)

with config.patch({
"epilogue_fusion": False,
"force_mixed_mm": True
mixed_mm_key: mixed_mm_val,
}):
for x_shape in [[2, 4], [5, 5, 5, 4], [1, 4, 4]]:
torch._dynamo.reset()
Expand All @@ -913,7 +917,7 @@ def test_weight_only_quant_use_mixed_mm(self, device, dtype):
m_c = torch.compile(m, mode="max-autotune")
y_wo, (code,) = run_and_get_code(m_c, x)
sqnr = compute_error(y_ref, y_wo)
self.assertGreater(sqnr, 43.0)
self.assertGreater(sqnr, 42.75)


class TestSaveLoadMeta(unittest.TestCase):
Expand Down
Loading