v3.9 (#2185)

* v3.8 update x

* fix blackwell gg

* doc change

* doc change

* doc change

---------

Co-authored-by: yuzhai <yuzhai@nvidia.com>
Co-authored-by: Haicheng Wu <haichengw@nvidia.com>
Co-authored-by: Haicheng Wu <57973641+hwu36@users.noreply.github.com>

test/unit/cute/core/complement.cpp

@@ -29,6 +29,8 @@
  *
  **************************************************************************************************/
 
+//#define CUTLASS_DEBUG_TRACE_LEVEL 1
+
 #include "cutlass_unit_test.h"
 
 #include <cutlass/trace.h>
@@ -68,6 +70,9 @@ test_complement(Layout const& layout, CoTarget const& cotarget)
   if constexpr (is_static<decltype(stride(completed))>::value) {  // If we can apply complement again
     EXPECT_EQ(size(complement(completed)), 1);                    // There's no more codomain left over
   }
+  if constexpr (is_static<decltype(result)>::value && is_static<decltype(layout)>::value) {
+    EXPECT_TRUE(bool(complement(complement(result,cosize(layout)),cotarget) == result));
+  }
 }
 
 template <class Layout>