fixed the global hoisiting of dequantization op to preserve i8 weights

compiler/src/iree/compiler/Codegen/Dialect/Codegen/IR/IREECodegenAttrs.td

@@ -71,7 +71,7 @@ def Linalg_TransformDialectCodegen
 def Custom
     : I32EnumAttrCase<"Custom", 1001>;
 
-def EXSLERATEV2_Conv2D 
+def EXSLERATEV2_Conv2D
     : I32EnumAttrCase<"Exsleratev2Conv2D", 1002>;
 
 def EXSLERATEV2_Pool

compiler/src/iree/compiler/Dialect/Encoding/Utils/CMakeLists.txt

@@ -28,7 +28,6 @@ iree_cc_library(
     MLIRLinalgUtils
     iree::compiler::Dialect::Encoding::IR
     iree::compiler::Dialect::Util::IR
-    iree::compiler::Dialect::LinalgExt::Utils
   PUBLIC
 )
 

compiler/src/iree/compiler/Dialect/LinalgExt/Utils/MatchUtils.cpp

@@ -516,7 +516,7 @@ isEXSLTiledConvolutionInterfaceImpl(Operation *op) {
   if (hascomplexExpr(indexingMaps.back())) {
     return detail::MatchEXSLTiledConvolutionResult::OutputDimsNotParallel;
   }
-  
+
   auto genericOp = dyn_cast<mlir::linalg::GenericOp>(linalgOp.getOperation());
   Block &body = genericOp.getRegion().front();
   bool hasmul = false;

compiler/src/iree/compiler/Dialect/Util/Analysis/Constant/ConstExpr.cpp

@@ -6,6 +6,8 @@
 
 #include "iree/compiler/Dialect/Util/Analysis/Constant/ConstExpr.h"
 
+#include "iree/compiler/Dialect/LinalgExt/IR/LinalgExtDialect.h"
+#include "iree/compiler/Dialect/LinalgExt/Utils/Utils.h"
 #include "iree/compiler/Dialect/Util/Analysis/Constant/OpOracle.h"
 #include "iree/compiler/Dialect/Util/Analysis/Explorer.h"
 #include "iree/compiler/Dialect/Util/IR/UtilOps.h"
@@ -502,6 +504,28 @@ void ConstExprHoistingPolicy::makeDecision(
     return;
   }
 
+  // Check if this operation or any of its producers is a bit-extend operation
+  // (e.g., i8→f32 dequantization). Don't hoist these to keep quantized weights
+  // compact and avoid pulling dequantization into global initializers.
+  if (IREE::LinalgExt::isBitExtendOp(info->getOperation())) {
+    decision->disableHoist();
+    return;
+  }
+
+  // Also check if any producer is a bit-extend op - this catches the case
+  // where we're trying to hoist the result of: i8 -> sitofp -> mulf(scale)
+  for (auto *producer : info->producers) {
+    if (IREE::LinalgExt::isBitExtendOp(producer->getOperation())) {
+      LLVM_DEBUG({
+        llvm::dbgs()
+            << "[ConstExprPolicy] Detected bit-extend producer, "
+            << "disabling hoisting to keep quantized weights compact\n";
+      });
+      decision->disableHoist();
+      return;
+    }
+  }
+
   // Otherwise, we can conditionally enable hoisting (based on cost model, etc).
   // TODO: Implement further conditions.
   decision->enableHoist();

compiler/src/iree/compiler/DispatchCreation/SetEncoding.cpp

@@ -220,7 +220,7 @@ static LogicalResult setDataTilingEncodings(RewriterBase &rewriter,
     encodedInputOperands.push_back(setEncodingWrapper(
         linalgOp.getDpsInputs()[1], IREE::Encoding::CONV_RHS));
     encodedInitOperand = setEncodingWrapper(linalgOp.getDpsInits()[0],
-                                             IREE::Encoding::CONV_RESULT);
+                                            IREE::Encoding::CONV_RESULT);
   } else {
     encodedInputOperands.push_back(setEncodingWrapper(
         linalgOp.getDpsInputs()[0], IREE::Encoding::SCALED_MATMUL_LHS));

compiler/src/iree/compiler/GlobalOptimization/CMakeLists.txt

@@ -50,7 +50,6 @@ iree_cc_library(
     "EraseUnusedLinalgOperands.cpp"
     "ExpandTensorShapes.cpp"
     "FuseDequantizationMatmul.cpp"
-    "GetPaddingValue.cpp"
     "GeneralizeLinalgNamedOps.cpp"
     "GlobalLoopInvariantCodeMotion.cpp"
     "InferNumericNarrowing.cpp"

compiler/src/iree/compiler/GlobalOptimization/GeneralizeLinalgNamedOps.cpp

@@ -97,7 +97,6 @@ void GeneralizeLinalgNamedOpsPass::runOnOperation() {
       }
     }
 
-
     FailureOr<linalg::GenericOp> generalizedOp =
         linalg::generalizeNamedOp(rewriter, linalgOp);
     if (failed(generalizedOp)) {

runtime/src/iree/schemas/CMakeLists.txt

@@ -126,22 +126,6 @@ flatbuffer_c_library(
   PUBLIC
 )
 
-flatbuffer_c_library(
-  NAME
-    exsleratev2_executable_def_c_fbs
-  SRCS
-    "exsleratev2_executable_def.fbs"
-  FLATCC_ARGS
-    "--reader"
-    "--builder"
-    "--verifier"
-    "--json"
-  INCLUDES
-    "executable_debug_info.fbs"
-  PUBLIC
-)
-
-
 add_custom_command(OUTPUT cpu_data_headers_filegroup.stamp
     COMMAND ${CMAKE_COMMAND} -E touch cpu_data_headers_filegroup.stamp
   DEPENDS