Collection of changes to fix clang build. (#1200)

* Remove unused variables

* Qualify calls to make_fragment_? from templated base class.

Fixes clang build error.

* Add missing `#include <cstdio>`

* Various changes to fix clang compile errors.

* More changes to fix clang build.

Remaining issues:

- `params` initializer of `CollectiveEpilogue`.
- `ops` initializer of `Sm90VisitorImplBase`.
- `__usAtomicCAS` needs to be added to clang upstream.

* Fix remaining clang build issues.

* Qualify `cute::rank()` calls.

* Qualify some more calls that are otherwise ambiguous between `cute` and `std` namespace.

* Double-escape special registers in inline asm.

* small change

---------

Co-authored-by: Haicheng Wu <haichengw@nvidia.com>

examples/53_hopper_gemm_permute/permute_traits.hpp

@@ -78,7 +78,7 @@ reshape(Shape const& shape, TargetShape const& target_shape)
 template<class Permute, bool Transpose, class Shape, class Stride>
 constexpr auto 
 make_permute_layout(Layout<Shape,Stride> const& layout) {
-  static_assert(rank(Shape{}) == 3, "Only rank-3 layouts are supported");
+  static_assert(cute::rank(Shape{}) == 3, "Only rank-3 layouts are supported");
   if constexpr (Transpose) {
     // Deal with tensor B by transposing appropriately before and after computing the permute layout.
     // Its CuTe-canonical mode order is [N,K,L], while permute operations expect [row,col,batch]. 
@@ -135,7 +135,7 @@ using inverse_t = decltype(inverse(T{}));
 template<class Permute, bool Transpose, class Shape, class Stride>
 constexpr auto 
 make_original_layout(Layout<Shape,Stride> const& layout) {
-  static_assert(rank(Shape{}) == 3, "Only rank-3 layouts are supported");
+  static_assert(cute::rank(Shape{}) == 3, "Only rank-3 layouts are supported");
   if constexpr (Transpose) {
     // Deal with tensor B by transposing appropriately before and after computing the permute layout.
     // Its CuTe-canonical mode order is [N,K,L], while permute operations expect [row,col,batch].