[mlir][Vector] Make VectorToSCF deterministic

mlir/lib/Conversion/VectorToSCF/VectorToSCF.cpp

LogicalResult VectorTransferRewriter<TransferReadOp>::matchAndRewrite(
   auto lbs = vectorBoundsCapture.getLbs();
   auto ubs = vectorBoundsCapture.getUbs();
   SmallVector<Value, 8> steps;
   steps.reserve(vectorBoundsCapture.getSteps().size());
   for (auto step : vectorBoundsCapture.getSteps())
     steps.push_back(std_constant_index(step));
 
   // 2. Emit alloc-copy-load-dealloc.
+  MLIRContext *ctx = op->getContext();
   Value tmp = setAllocAtFunctionEntry(tmpMemRefType(transfer), transfer);
   StdIndexedValue local(tmp);
-  Value vec = vector_type_cast(tmp);
   loopNestBuilder(lbs, ubs, steps, [&](ValueRange loopIvs) {
     auto ivs = llvm::to_vector<8>(loopIvs);
     // Swap the ivs which will reorder memory accesses.
     if (coalescedIdx >= 0)
       std::swap(ivs.back(), ivs[coalescedIdx]);
     // Computes clippedScalarAccessExprs in the loop nest scope (ivs exist).
     SmallVector<Value, 8> indices = clip(transfer, memRefBoundsCapture, ivs);
     ArrayRef<Value> indicesRef(indices), ivsRef(ivs);
-    Value pos =
-        std_index_cast(IntegerType::get(32, op->getContext()), ivsRef.back());
-    Value vector = vector_insert_element(remote(indicesRef),
-                                         local(ivsRef.drop_back()), pos);
+    Value pos = std_index_cast(IntegerType::get(32, ctx), ivsRef.back());
+    Value scal = remote(indicesRef);
+    Value vector = vector_insert_element(scal, local(ivsRef.drop_back()), pos);
     local(ivsRef.drop_back()) = vector;
   });
-  Value vectorValue = std_load(vec);
+  Value vectorValue = std_load(vector_type_cast(tmp));
 
   // 3. Propagate.
   rewriter.replaceOp(op, vectorValue);
   return success();
 }
 
 /// Lowers TransferWriteOp into a combination of:
 ///   1. local memory allocation;

mlir/test/Conversion/VectorToSCF/vector-to-loops.mlir

func @materialize_read(%M: index, %N: index, %O: index, %P: index) {
   // CHECK-NEXT:                {{.*}} = affine.apply #[[$ADD]](%[[I3]], %[[I6]])
   // CHECK-NEXT:                {{.*}} = affine.apply #[[$SUB]]()[%{{.*}}]
   // CHECK-NEXT:                {{.*}} = cmpi "slt", {{.*}} : index
   // CHECK-NEXT:                {{.*}} = select
   // CHECK-NEXT:                {{.*}} = cmpi "slt", {{.*}}, %[[C0]] : index
   // CHECK-NEXT:                %[[L3:.*]] = select
   // CHECK-NEXT:                %[[VIDX:.*]] = index_cast %[[I4]]
   //
-  //  CHECK-DAG:                %[[SCAL:.*]] = load %{{.*}}[%[[L0]], %[[L1]], %[[L2]], %[[L3]]] : memref<?x?x?x?xf32>
-  //  CHECK-DAG:                %[[VEC:.*]] = load %[[ALLOC]][%[[I6]], %[[I5]]] : memref<5x4xvector<3xf32>>
+  // CHECK-NEXT:                %[[SCAL:.*]] = load %{{.*}}[%[[L0]], %[[L1]], %[[L2]], %[[L3]]] : memref<?x?x?x?xf32>
+  // CHECK-NEXT:                %[[VEC:.*]] = load %[[ALLOC]][%[[I6]], %[[I5]]] : memref<5x4xvector<3xf32>>
   // CHECK-NEXT:                %[[RVEC:.*]] = vector.insertelement %[[SCAL]], %[[VEC]][%[[VIDX]] : i32] : vector<3xf32>
   // CHECK-NEXT:                store %[[RVEC]], %[[ALLOC]][%[[I6]], %[[I5]]] : memref<5x4xvector<3xf32>>
   // CHECK-NEXT:              }
   // CHECK-NEXT:            }
   // CHECK-NEXT:          }
   // CHECK-NEXT:        }
   // CHECK-NEXT:      }
   // CHECK-NEXT:    }