From bc322c1cdaf9938e2cf3bd8725465f6725db19b8 Mon Sep 17 00:00:00 2001
From: Patrice Vignola <vignola.patrice@gmail.com>
Date: Fri, 6 Oct 2023 23:00:18 -0700
Subject: [PATCH] Add check in case CPU inputs changed

---
 .../src/DmlRuntimeFusedGraphKernel.cpp        | 44 +++++++++++++++----
 1 file changed, 35 insertions(+), 9 deletions(-)

diff --git a/onnxruntime/core/providers/dml/DmlExecutionProvider/src/DmlRuntimeFusedGraphKernel.cpp b/onnxruntime/core/providers/dml/DmlExecutionProvider/src/DmlRuntimeFusedGraphKernel.cpp
index 2c9d27b0546f6..fd26a90be87e7 100644
--- a/onnxruntime/core/providers/dml/DmlExecutionProvider/src/DmlRuntimeFusedGraphKernel.cpp
+++ b/onnxruntime/core/providers/dml/DmlExecutionProvider/src/DmlRuntimeFusedGraphKernel.cpp
@@ -99,36 +99,62 @@ namespace Dml
         {
             ORT_THROW_HR_IF(E_UNEXPECTED, m_subgraphInputs.size() != kernelContext->InputCount());
 
-            bool recompiledNeeded = m_compiledExecutionPlanOperator == nullptr;
+            bool recompileNeeded = m_compiledExecutionPlanOperator == nullptr;
 
             for (int inputIndex = 0; inputIndex < kernelContext->InputCount(); ++inputIndex)
             {
                 const auto& input = kernelContext->RequiredInput<onnxruntime::Tensor>(inputIndex);
                 const std::string& inputName = m_subgraphInputs[inputIndex]->Name();
-                auto iter = m_inferredInputShapes.find(inputName);
+                auto shapeIter = m_inferredInputShapes.find(inputName);
 
-                if (iter == m_inferredInputShapes.end())
+                if (shapeIter == m_inferredInputShapes.end())
                 {
                     m_inferredInputShapes[inputName] = input.Shape();
-                    recompiledNeeded = true;
+                    recompileNeeded = true;
                 }
-                else if (iter->second != input.Shape())
+                else if (shapeIter->second != input.Shape())
                 {
-                    iter->second = input.Shape();
-                    recompiledNeeded = true;
+                    shapeIter->second = input.Shape();
+                    recompileNeeded = true;
                 }
 
                 // If we have CPU inputs that are not initializers (i.e. they were computed at runtime), add them to the initializer list
                 if (input.Location().device.Type() == OrtDevice::CPU)
                 {
-                    // TODO (pavignol): Force recompile if CPU data changed
                     auto inputProto = onnxruntime::utils::TensorToTensorProto(input, inputName);
+
+                    // We can only avoid recompiling the graph when all CPU inputs are identical
+                    auto initializerIter = m_isInitializerTransferable.find(inputName);
+
+                    if (initializerIter != m_isInitializerTransferable.end())
+                    {
+                        if (initializerIter->second.first->raw_data().length() == inputProto.raw_data().length())
+                        {
+                            for (int i = 0; i < inputProto.raw_data().length(); ++i)
+                            {
+                                if (initializerIter->second.first->raw_data()[i] != inputProto.raw_data()[i])
+                                {
+                                    recompileNeeded = true;
+                                    break;
+                                }
+                            }
+                        }
+                        else
+                        {
+                            recompileNeeded = true;
+                        }
+                    }
+                    else
+                    {
+                        recompileNeeded = true;
+                    }
+
                     m_ownedCpuInputs.push_back(std::make_unique<ONNX_NAMESPACE::TensorProto>(std::move(inputProto)));
                     m_isInitializerTransferable[inputName] = std::make_pair(m_ownedCpuInputs.back().get(), false);
                 }
             }
 
-            if (recompiledNeeded)
+            if (recompileNeeded)
             {
                 // Go through all the node args and replace their shapes with the real ones
                 for (auto& nodeArg : m_intermediateNodeArgs)