better MaterializeMergeSchedule.cpp that emits much more compact IR

add support for other constant-time arith ops in codegen

src/PIM/Pass/PimCodegen/MaterializeHostConstantsPass.cpp

@@ -2,6 +2,7 @@
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/Dialect/SCF/IR/SCF.h"
 #include "mlir/IR/Builders.h"
+#include "mlir/IR/Dominance.h"
 #include "mlir/IR/PatternMatch.h"
 #include "mlir/Pass/Pass.h"
 
@@ -37,6 +38,7 @@ static void materializeHostConstantsInCore(CoreOpTy coreOp,
                                            OperationFolder& constantFolder,
                                            bool& hasFailure) {
   DenseMap<Value, DenseMap<int64_t, DenseMap<Type, Value>>> materializedValues;
+  DominanceInfo dominance(coreOp);
   SmallVector<Operation*> ops;
   coreOp.getBody().front().walk([&](Operation* op) {
     if (!isa<pim::PimHaltOp, scf::YieldOp>(op))
@@ -70,7 +72,7 @@ static void materializeHostConstantsInCore(CoreOpTy coreOp,
       auto& cachedByOffset = materializedValues[resolvedAddress->base];
       auto& cachedByType = cachedByOffset[resolvedAddress->byteOffset];
       auto cachedValue = cachedByType.find(originalType);
-      if (cachedValue != cachedByType.end()) {
+      if (cachedValue != cachedByType.end() && dominance.properlyDominates(cachedValue->second, op)) {
         operand.set(cachedValue->second);
         continue;
       }