Skip to content

[Issue 192] Tail split support for dynamic matmul #227

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
merged 2 commits into from
Oct 21, 2024
Merged

[Issue 192] Tail split support for dynamic matmul #227

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
b680f65
[Issue 192] Tail split support for dynamic matmul
tzj-fxz Oct 21, 2024
eac121b
[Issue 192] Tail split support for dynamic matmul & add test case
tzj-fxz Oct 21, 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
2 changes: 1 addition & 1 deletion 3rdparty/tvm
Open in desktop
Submodule tvm updated from 8d73b0 to b63fba
13 changes: 10 additions & 3 deletions testing/python/tilelang/test_tilelang_dyanmic_symbolic.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -372,13 +372,18 @@ def assert_tl_matmul_block_all_dynamic_correctness(
)
mod, params = TL.lower(program)

A = torch.rand(M, K, device="cuda", dtype=getattr(torch, in_dtype))
B = torch.rand(N, K, device="cuda", dtype=getattr(torch, in_dtype))
if trans_A:
A = torch.rand(K, M, device="cuda", dtype=getattr(torch, in_dtype))
else:
A = torch.rand(M, K, device="cuda", dtype=getattr(torch, in_dtype))
if trans_B:
B = torch.rand(N, K, device="cuda", dtype=getattr(torch, in_dtype))
else:
B = torch.rand(K, N, device="cuda", dtype=getattr(torch, in_dtype))
C = torch.zeros(M, N, device="cuda", dtype=getattr(torch, out_dtype))

mod = TL.Profiler(mod, params, [], TL.TensorSupplyType.Integer)
mod(A, B, C)
print(mod.mod.imported_modules[0].get_source())

def ref_program(A, B):
import torch
Expand Down Expand Up @@ -419,6 +424,8 @@ def test_assert_tl_matmul_block_all_dynamic():
"float16", 64, 64, 32)
assert_tl_matmul_block_all_dynamic_correctness(36, 128, 128, False, False, "float16", "float16",
"float16", 64, 64, 32)
assert_tl_matmul_block_all_dynamic_correctness(36, 115, 103, False, False, "float16", "float16",
"float16", 64, 64, 32)


if __name__ == "__main__":
Expand Down
Loading