[BACKEND] Support MMA V3 with register operand (#2375)

MMA V3 support taking operand A from register. This helps for chained
matmul operations like in attention.
Add an optimization to use this mode when it helps and add the lowering
for it.

include/triton/Analysis/Utility.h

@@ -133,6 +133,10 @@ bool isMmaToDotShortcut(RankedTensorType &srcTy, RankedTensorType &dstTy);
 
 bool isMmaToMmaShortcut(RankedTensorType &srcTy, RankedTensorType &dstTy);
 
+// Return true if the src and dst layout match.
+bool matchMmaV3AndDotOperandLayout(RankedTensorType srcTy,
+                                   RankedTensorType dstTy);
+
 // TODO: Move utility functions that belong to ConvertLayoutOp to class
 // ConvertLayoutOpHelper in the future
 bool shouldUseDistSmem(Attribute srcLayout, Attribute dstLayout);