Skip to content

[DML EP] Prefer MatMulInteger over MatMulIntegerToFloat in case of #22469

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 4 commits into from
Nov 7, 2024
Merged

[DML EP] Prefer MatMulInteger over MatMulIntegerToFloat in case of #22469

Changes from 1 commit
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
86e34fb
[DML EP] Prefer MatMulInteger over MatMulIntegerToFloat in case of
Oct 16, 2024
5945b72
Add large resource constraint
Nov 5, 2024
a150024
clean up
Nov 5, 2024
db6c4a6
Address Review and lint
Nov 5, 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
9 changes: 9 additions & 0 deletions onnxruntime/core/optimizer/matmul_integer_to_float.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -114,6 +114,15 @@ Status MatMulIntegerToFloatFusion::ApplyImpl(Graph& graph, bool& modified, int g
continue;
}

const Node* p_dynamicquantize_node = graph_utils::FirstParentByType(*p_matmulinteger_node, "DynamicQuantizeLinear");

// Check MatMulInteger Nodes' input is coming from DynamicQuantizeLinear
// For larger tensors DynamicQuantizeLinear -> MatMulInteger is used to be resource efficient
// And we have better MatMulInteger Metacommand coverage in DML
if (is_dml_ep && p_dynamicquantize_node) {
continue;
}

// Find bias node
Node* p_add_node = nullptr;
if (optimizer_utils::CheckOutputEdges(graph, mul_node, 1)) {
Expand Down
Loading