Skip to content

[DLight] Skip GEMV rules when more than one vector #17052

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 1 commit into from
May 31, 2024
Merged

[DLight] Skip GEMV rules when more than one vector #17052

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
6 changes: 3 additions & 3 deletions python/tvm/dlight/gpu/gemv.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -206,8 +206,7 @@ def apply( # pylint: disable=too-many-locals,too-many-branches,too-many-return-
if is_inner_reduction is None:
return None
elif is_inner_reduction:
self.sch_inner_reduction(sch, target, block, vector_input_buffers, epilogue)
return sch
return self.sch_inner_reduction(sch, target, block, vector_input_buffers, epilogue)
elif target.kind.name == "opencl" and "android" in str(target.host):
ret = self.sch_outer_reduction(sch, target, block, vector_input_buffers, epilogue)
if ret is None:
Expand Down Expand Up @@ -313,7 +312,8 @@ def apply(

# load vector into shared memory, shape should be the whole vector
if LOAD_V_SHARED:
assert len(vector_input_buffers) == 1
if len(vector_input_buffers) != 1:
return None
V_shared = sch.cache_read(rf, read_buffer_index=0, storage_scope="shared")
sch.compute_at(V_shared, tr, preserve_unit_loops=True)
l = sch.get_loops(block=V_shared)[-1]
Expand Down