produce ops in definite order

Author: fabianmcg

mlir/lib/Conversion/AMDGPUToROCDL/AMDGPUToROCDL.cpp

@@ -218,9 +218,9 @@ struct RawBufferOpLowering : public ConvertOpToLLVMPattern<GpuOp> {
     } else {
       Value maxIndex;
       for (uint32_t i = 0, e = memrefType.getRank(); i < e; ++i) {
-        Value maxThisDim = rewriter.create<LLVM::MulOp>(
-            loc, memrefDescriptor.size(rewriter, loc, i),
-            memrefDescriptor.stride(rewriter, loc, i));
+        Value size = memrefDescriptor.size(rewriter, loc, i);
+        Value stride = memrefDescriptor.stride(rewriter, loc, i);
+        Value maxThisDim = rewriter.create<LLVM::MulOp>(loc, size, stride);
         maxIndex =
             maxIndex ? rewriter.create<LLVM::UMaxOp>(loc, maxIndex, maxThisDim)
                      : maxThisDim;

mlir/test/Conversion/AMDGPUToROCDL/amdgpu-to-rocdl.mlir

@@ -39,7 +39,7 @@ func.func @gpu_gcn_raw_buffer_load_i32_strided(%buf: memref<16x16xi32, strided<[
     // CHECK: %[[ptr:.*]] = llvm.getelementptr %[[algn_ptr]][%[[offset]]] : (!llvm.ptr, i64) -> !llvm.ptr, i32
     // CHECK: %[[stride:.*]] = llvm.mlir.constant(0 : i16) : i16
     // CHECK: %[[sz_i:.*]] = llvm.extractvalue %[[descriptor]][3, 0] : !llvm.struct<(ptr, ptr, i64, array<2 x i64>, array<2 x i64>)> 
-    // CHECK: %[[stride_i:[0-9_a-zA-Z]*]] = llvm.extractvalue %[[descriptor]][4, 0] : !llvm.struct<(ptr, ptr, i64, array<2 x i64>, array<2 x i64>)> 
+    // CHECK: %[[stride_i:.*]] = llvm.extractvalue %[[descriptor]][4, 0] : !llvm.struct<(ptr, ptr, i64, array<2 x i64>, array<2 x i64>)> 
     // CHECK: %[[ext_i:.*]] = llvm.mul %[[sz_i]], %[[stride_i]] : i64
     // CHECK: %[[sz_j:.*]] = llvm.extractvalue %[[descriptor]][3, 1] : !llvm.struct<(ptr, ptr, i64, array<2 x i64>, array<2 x i64>)> 
     // CHECK: %[[stride_j:.*]] = llvm.extractvalue %[[descriptor]][4, 1] : !llvm.struct<(ptr, ptr, i64, array<2 x i64>, array<2 x i64>)>