Skip to content

[FlashInfer] Update include path and interface #18317

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
Sep 15, 2025
Merged

[FlashInfer] Update include path and interface #18317

Show file tree
Hide file tree
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
13 changes: 9 additions & 4 deletions python/tvm/relax/backend/cuda/flashinfer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -141,8 +141,8 @@ def get_object_file_path(src: Path) -> Path:
)
include_paths += [
Path(tvm_home).resolve() / "include",
Path(tvm_home).resolve() / "ffi" / "include",
Path(tvm_home).resolve() / "ffi" / "3rdparty" / "dlpack" / "include",
Path(tvm_home).resolve() / "3rdparty" / "tvm-ffi" / "include",
Path(tvm_home).resolve() / "3rdparty" / "tvm-ffi" / "3rdparty" / "dlpack" / "include",
Path(tvm_home).resolve() / "3rdparty" / "dmlc-core" / "include",
]
else:
Expand All @@ -160,8 +160,13 @@ def get_object_file_path(src: Path) -> Path:
# The package is installed from source.
include_paths += [
tvm_package_path.parent.parent / "include",
tvm_package_path.parent.parent / "ffi" / "include",
tvm_package_path.parent.parent / "ffi" / "3rdparty" / "dlpack" / "include",
tvm_package_path.parent.parent / "3rdparty" / "tvm-ffi" / "include",
tvm_package_path.parent.parent
/ "3rdparty"
/ "tvm-ffi"
/ "3rdparty"
/ "dlpack"
/ "include",
tvm_package_path.parent.parent / "3rdparty" / "dmlc-core" / "include",
]
else:
Expand Down
6 changes: 4 additions & 2 deletions src/runtime/vm/attn_backend.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -176,7 +176,8 @@ class FlashInferPagedPrefillFunc : public PagedPrefillFunc {
plan_func_(float_workspace_buffer, int_workspace_buffer, page_locked_int_workspace_buffer,
qo_indptr->as_tensor(), page_indptr->as_tensor(), IntTuple(std::move(kv_len)),
total_qo_len, batch_size, num_qo_heads, num_kv_heads, page_size,
/*enable_cuda_graph=*/false, qk_head_dim, v_head_dim, causal, copy_stream)
/*enable_cuda_graph=*/false, qk_head_dim, v_head_dim, causal,
/*window_left=*/-1, copy_stream)
.cast<IntTuple>();
} else if (attn_kind == AttnKind::kMLA) {
plan_info_vec =
Expand Down Expand Up @@ -280,7 +281,8 @@ class FlashInferRaggedPrefillFunc : public RaggedPrefillFunc {
plan_func_(float_workspace_buffer, int_workspace_buffer, page_locked_int_workspace_buffer,
qo_indptr->as_tensor(), kv_indptr->as_tensor(), IntTuple(std::move(kv_len)),
total_qo_len, batch_size, num_qo_heads, num_kv_heads, /*page_size=*/1,
/*enable_cuda_graph=*/false, qk_head_dim, v_head_dim, causal, copy_stream)
/*enable_cuda_graph=*/false, qk_head_dim, v_head_dim, causal,
/*window_left=*/-1, copy_stream)
.cast<IntTuple>();
}

Expand Down
Loading