Skip to content

[ET-VK] Fix Build Errors in Vulkan Backend #13170

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
Aug 7, 2025
Merged

[ET-VK] Fix Build Errors in Vulkan Backend #13170

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
28 changes: 8 additions & 20 deletions backends/vulkan/op_registry.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -384,26 +384,14 @@ def check_reduce_node(node: torch.fx.Node) -> bool:
memory_layout = utils.get_node_memory_layout(node)

# If we have memory layout information, check if any dimension in dim_list corresponds to a packed dimension
if memory_layout is not None:
for dim in dim_list:
# For WIDTH_PACKED layout, dimension 3 (W) is packed
# For HEIGHT_PACKED layout, dimension 2 (H) is packed
# For CHANNELS_PACKED layout, dimension 1 (C) is packed
if (
(
memory_layout == VkMemoryLayout.TENSOR_WIDTH_PACKED
and dim == 3
)
or (
memory_layout == VkMemoryLayout.TENSOR_HEIGHT_PACKED
and dim == 2
)
or (
memory_layout == VkMemoryLayout.TENSOR_CHANNELS_PACKED
and dim == 1
)
):
return False
if (
memory_layout is not None
and memory_layout != VkMemoryLayout.DEFAULT_LAYOUT
):
# For now only default layout is supported for 2D reduction.
# Because we can't determine if the input is NCHW or NHWC here,
# assume the reduction dimension is packed so we cannot support it.
return False
except (AssertionError, KeyError, AttributeError):
# If we can't get memory layout information, we'll assume the dims aren't packed
pass
Expand Down
8 changes: 4 additions & 4 deletions backends/vulkan/runtime/graph/ops/impl/Reduce.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -37,19 +37,19 @@ void resize_reduce2d_node(
ComputeGraph* graph,
const std::vector<ArgGroup>& args,
const std::vector<ValueRef>& resize_args) {
vTensorPtr out = graph->get_tensor(args[0].refs[0]);
vTensorPtr in = graph->get_tensor(args[1].refs[0]);
const ValueRef out = args.at(0).refs.at(0);
const ValueRef in = args.at(1).refs.at(0);

// Extract the dimensions to reduce over
const std::vector<int64_t> dims_list =
graph->extract_int_or_symint_list(resize_args.at(0));
int32_t reduce_dim1_nchw = dims_list[0];
int32_t reduce_dim2_nchw = dims_list[1];

std::vector<int64_t> new_sizes = in->sizes();
std::vector<int64_t> new_sizes = graph->sizes_of(in);
new_sizes.at(normalize(reduce_dim1_nchw, new_sizes.size())) = 1;
new_sizes.at(normalize(reduce_dim2_nchw, new_sizes.size())) = 1;
out->virtual_resize(new_sizes);
graph->virtual_resize(out, new_sizes);
}

utils::uvec3 reduce_global_wg_size(
Expand Down
Loading