Skip to content

Commit

Permalink
disable gemm f16 on CPU (#19744)
Browse files Browse the repository at this point in the history 
### Description
<!-- Describe your changes. -->
Temporarily disable fp16 gemm on CPU because it usually needs a
following Cast which offsets the gain. Need more fp16 operators
implementation and performance tuning.

Also fix a fusion error of LayerNormalization.
### Motivation and Context
<!-- - Why is this change required? What problem does it solve?
- If it fixes an open issue, please link to the issue here. -->
  • Loading branch information
@yufenglee
yufenglee authored Mar 1, 2024
1 parent 5672cde commit 22176a5
Show file tree
Hide file tree
Showing 4 changed files with 19 additions and 23 deletions.
5 changes: 4 additions & 1 deletion .vscode/settings.json
View file
Original file line number Diff line number Diff line change
Expand Up @@ -21,5 +21,8 @@
"cpplint.filters": [
"-build/include_subdir",
"-runtime/references"
]
],
"files.associations": {
"span": "cpp"
}
}
14 changes: 14 additions & 0 deletions onnxruntime/core/optimizer/layer_norm_fusion.cc
View file
Original file line number Diff line number Diff line change
Expand Up @@ -447,6 +447,13 @@ Status LayerNormFusion::ApplyImpl(Graph& graph, bool& modified, int graph_level,

NodeArg* x_input = has_leading_cast ? graph.GetNode(p_reduce_mean_input_node->Index())->MutableInputDefs()[0]
: reduce_mean_node.MutableInputDefs()[0];

// CPU doesn't support fp16
if (reduce_mean_node.GetExecutionProviderType() == kCpuExecutionProvider &&
x_input->TypeAsProto()->tensor_type().elem_type() == ONNX_NAMESPACE::TensorProto_DataType_FLOAT16) {
continue;
}

InlinedVector<NodeArg*> layer_norm_input_defs{x_input, scale, bias};
Node& layer_norm_node = graph.AddNode(graph.GenerateNodeName("LayerNormalization"),
"LayerNormalization",
Expand Down Expand Up @@ -689,6 +696,13 @@ Status SimplifiedLayerNormFusion::ApplyImpl(Graph& graph, bool& modified, int gr

NodeArg* x_input = has_leading_cast ? graph.GetNode(p_pow_input_node->Index())->MutableInputDefs()[0]
: pow_node.MutableInputDefs()[0];

// CPU doesn't support fp16
if (reduce_mean_node.GetExecutionProviderType() == kCpuExecutionProvider &&
x_input->TypeAsProto()->tensor_type().elem_type() == ONNX_NAMESPACE::TensorProto_DataType_FLOAT16) {
continue;
}

InlinedVector<NodeArg*> layer_norm_input_defs{x_input, scale};
Node& layer_norm_node =
graph.AddNode(graph.GenerateNodeName("SimplifiedLayerNormalization"), "SimplifiedLayerNormalization",
Expand Down
21 changes: 0 additions & 21 deletions onnxruntime/core/providers/cpu/cpu_execution_provider.cc
View file
Original file line number Diff line number Diff line change
Expand Up @@ -143,9 +143,6 @@ class ONNX_OPERATOR_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 7, Aco
class ONNX_OPERATOR_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 7, Atan);
class ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 7, 8, float, Gemm);
class ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 7, 8, double, Gemm);
#ifdef MLAS_F16VEC_INTRINSICS_SUPPORTED
class ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 7, 8, MLFloat16, Gemm);
#endif
class ONNX_OPERATOR_VERSIONED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 1, 10, Hardmax);
class ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 1, 10, float, LogSoftmax);
class ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 1, 10, double, LogSoftmax);
Expand Down Expand Up @@ -335,9 +332,6 @@ class ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOn
class ONNX_OPERATOR_VERSIONED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 9, 10, Flatten);
class ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 9, 10, float, Gemm);
class ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 9, 10, double, Gemm);
#ifdef MLAS_F16VEC_INTRINSICS_SUPPORTED
class ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 9, 10, MLFloat16, Gemm);
#endif
class ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 9, 12, float, MatMul);
class ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 9, 12, double, MatMul);
class ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 9, 12, int32_t, MatMul);
Expand Down Expand Up @@ -497,9 +491,6 @@ class ONNX_OPERATOR_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 11, Sp
class ONNX_OPERATOR_VERSIONED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 11, 12, ScatterND);
class ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 11, 12, float, Gemm);
class ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 11, 12, double, Gemm);
#ifdef MLAS_F16VEC_INTRINSICS_SUPPORTED
class ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 11, 12, MLFloat16, Gemm);
#endif
class ONNX_OPERATOR_VERSIONED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 11, 12, GatherElements);
class ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 11, uint8_t, BitShift);
class ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 11, uint32_t, BitShift);
Expand Down Expand Up @@ -606,9 +597,6 @@ class ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain,
class ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 13, string, Expand);
class ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 13, float, Gemm);
class ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 13, double, Gemm);
#ifdef MLAS_F16VEC_INTRINSICS_SUPPORTED
class ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 13, MLFloat16, Gemm);
#endif
class ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 13, float, MatMul);
class ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 13, double, MatMul);
class ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 13, int32_t, MatMul);
Expand Down Expand Up @@ -2617,15 +2605,6 @@ Status RegisterFp16Kernels(KernelRegistry& kernel_registry) {
MLFloat16, LeakyRelu)>,
BuildKernelCreateInfo<ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 16, MLFloat16,
LeakyRelu)>,
BuildKernelCreateInfo<ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 7, 8,
MLFloat16, Gemm)>,
BuildKernelCreateInfo<ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 9, 10,
MLFloat16, Gemm)>,
BuildKernelCreateInfo<ONNX_OPERATOR_VERSIONED_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 11, 12,
MLFloat16, Gemm)>,

BuildKernelCreateInfo<ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 13, MLFloat16,
Gemm)>,
};

for (auto& function_table_entry : function_table) {
Expand Down
2 changes: 1 addition & 1 deletion onnxruntime/test/providers/cpu/math/gemm_test.cc
View file
Original file line number Diff line number Diff line change
Expand Up @@ -277,7 +277,7 @@ class GemmOpTypedTests : public ::testing::Test {
// On CPUs without fp16 instructions the tests will output a warning:
// "registered execution providers CPUExecutionProvider were unable to run the model"
// , then they will still pass.
using GemmOpTypedTestsTypes = ::testing::Types<float, double, MLFloat16>;
using GemmOpTypedTestsTypes = ::testing::Types<float, double>;
TYPED_TEST_SUITE(GemmOpTypedTests, GemmOpTypedTestsTypes);

TYPED_TEST(GemmOpTypedTests, TestGemmScalarBroadcast) {
Expand Down

0 comments on commit 22176a5

Please sign in to comment.