[BACKEND][FRONTEND] Fix problems with test_matmul (#973)

1. Handle induction variable when step is negative
2. Restore async_wait that accidentally deleted
3. Add missing induction variable in prefetch
4. Add device property functions

Co-authored-by: Philippe Tillet <Phil.Tillet@gmail.com>

lib/Dialect/TritonGPU/Transforms/Combine.cpp

@@ -782,8 +782,8 @@ public:
                                                  newRetType.getEncoding()));
     a = rewriter.create<triton::gpu::ConvertLayoutOp>(a.getLoc(), newAType, a);
     b = rewriter.create<triton::gpu::ConvertLayoutOp>(b.getLoc(), newBType, b);
-    auto newDot = rewriter.create<triton::DotOp>(
-        dotOp.getLoc(), newRetType, a, b, newAcc, dotOp.allowTF32());
+    auto newDot = rewriter.create<triton::DotOp>(dotOp.getLoc(), newRetType, a,
+                                                 b, newAcc, dotOp.allowTF32());
 
     rewriter.replaceOpWithNewOp<triton::gpu::ConvertLayoutOp>(
         op, oldRetType, newDot.getResult());