Skip to content

link bmm, mm, view_copy, slice_copy, split_with_sizes_copy to jarvis #10436

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 3 commits into from
Apr 28, 2025
Merged

link bmm, mm, view_copy, slice_copy, split_with_sizes_copy to jarvis #10436

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
2aea989
Adding bmm, mm, view_copy, slice_copy, split_with_sizes_copy optimiza…
cad-audio Apr 23, 2025
02817c0
link bmm, mm, view_copy, slice_copy, split_with_sizes_copy to jarvis
zonglinpeng Apr 24, 2025
95f8b7a
Merge branch 'main' into export-D73514531
zonglinpeng Apr 28, 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
10 changes: 5 additions & 5 deletions backends/cadence/hifi/operators/op_bmm.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,8 +16,8 @@ using exec_aten::ScalarType;
using executorch::runtime::KernelRuntimeContext;
using executorch::runtime::kTensorDimensionLimit;
using executorch::runtime::resize_tensor;
using executorch::runtime::tensors_have_same_dim_order;
using executorch::runtime::tensor_is_default_dim_order;
using executorch::runtime::tensors_have_same_dim_order;
using torch::executor::check_bmm_args;
using torch::executor::Error;
using torch::executor::get_bmm_out_target_size;
Expand Down Expand Up @@ -78,16 +78,16 @@ Tensor& bmm_out(
WORD32 out_stride = p;

WORD32* __restrict__ tmp =
(WORD32* __restrict__)kernels::allocate_temp_memory(
ctx, (batch_size * m * p) * sizeof(float));
(WORD32* __restrict__)kernels::allocate_temp_memory(
ctx, (batch_size * m * p) * sizeof(float));

ET_KERNEL_CHECK(ctx, tmp != nullptr, MemoryAllocationFailed, out);

tmp[batch_size * m * p] = {0};

WORD32* __restrict__ p_o =
(WORD32* __restrict__)kernels::allocate_temp_memory(
ctx, (batch_size * m * p) * sizeof(WORD32));
(WORD32* __restrict__)kernels::allocate_temp_memory(
ctx, (batch_size * m * p) * sizeof(WORD32));

ET_KERNEL_CHECK(ctx, p_o != nullptr, MemoryAllocationFailed, out);

Expand Down
4 changes: 2 additions & 2 deletions backends/cadence/hifi/operators/op_mm.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -76,8 +76,8 @@ Tensor& mm_out(
WORD32 out_stride = p;

WORD32* __restrict__ p_o =
(WORD32* __restrict__)kernels::allocate_temp_memory(
ctx, (n * p) * sizeof(WORD32));
(WORD32* __restrict__)kernels::allocate_temp_memory(
ctx, (n * p) * sizeof(WORD32));

WORD32 p_inp_shape[2];
p_inp_shape[0] = n;
Expand Down
5 changes: 5 additions & 0 deletions backends/cadence/hifi/operators/targets.bzl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -34,6 +34,11 @@ def define_operator(name: str, deps: list[str] | None = None) -> None:
OPERATORS = [
"add",
"atan2",
"bmm",
"mm",
"slice_copy",
"split_with_sizes_copy",
"view_copy",
"cat",
"clamp",
"dequantize_per_tensor",
Expand Down
Loading