diff --git a/mlir/include/mlir/Dialect/LLVMIR/NVVMOps.td b/mlir/include/mlir/Dialect/LLVMIR/NVVMOps.td
--- a/mlir/include/mlir/Dialect/LLVMIR/NVVMOps.td
+++ b/mlir/include/mlir/Dialect/LLVMIR/NVVMOps.td
@@ -514,7 +514,7 @@
   let assemblyFormat = "$mask attr-dict `:` type($mask)";
 }
 
-// https://docs.nvidia.com/cuda/parallel-thread-execution/#id62
+
 def LoadCacheModifierCA : I32EnumAttrCase<"CA", 0, "ca">;
 def LoadCacheModifierCG : I32EnumAttrCase<"CG", 1, "cg">;
 def LoadCacheModifierCS : I32EnumAttrCase<"CS", 2, "cs">;
@@ -528,6 +528,11 @@
     LoadCacheModifierLU, LoadCacheModifierCV]> {
   let genSpecializedAttr = 0;
   let cppNamespace = "::mlir::NVVM";
+  let description = [{
+    Enum attribute of the different kinds of cache operators for load instructions.
+
+    [For more information, see PTX ISA](https://docs.nvidia.com/cuda/parallel-thread-execution/#id62)    
+  }];
 }
 
 def LoadCacheModifierAttr : EnumAttr<NVVM_Dialect, LoadCacheModifierKind, "load_cache_modifier">;
@@ -1436,8 +1441,8 @@
   let description = [{
     Enforce an ordering of register accesses between warpgroup level matrix 
     multiplication and other operations. 
-    See for more information:
-    https://docs.nvidia.com/cuda/parallel-thread-execution/index.html#asynchronous-warpgroup-level-matrix-instructions-wgmma-fence
+    
+    [For more information, see PTX ISA](https://docs.nvidia.com/cuda/parallel-thread-execution/index.html#asynchronous-warpgroup-level-matrix-instructions-wgmma-fence)
   }];
   let assemblyFormat = "attr-dict";
   let extraClassDefinition = [{
@@ -1451,8 +1456,8 @@
   let assemblyFormat = "attr-dict";
   let description = [{
     Commits all prior uncommitted warpgroup level matrix multiplication operations.
-    See for more information:
-    https://docs.nvidia.com/cuda/parallel-thread-execution/index.html#asynchronous-warpgroup-level-matrix-instructions-wgmma-commit-group
+    
+    [For more information, see PTX ISA](https://docs.nvidia.com/cuda/parallel-thread-execution/index.html#asynchronous-warpgroup-level-matrix-instructions-wgmma-commit-group)
   }];
   let extraClassDefinition = [{
     std::string $cppClass::getPtx() { return std::string("wgmma.commit_group.sync.aligned;"); }
@@ -1465,8 +1470,8 @@
   let assemblyFormat = "attr-dict $group";
   let description = [{
     Signal the completion of a preceding warpgroup operation.
-    See for more information:
-    https://docs.nvidia.com/cuda/parallel-thread-execution/index.html#asynchronous-warpgroup-level-matrix-instructions-wgmma-wait-group
+    
+    [For more information, see PTX ISA](https://docs.nvidia.com/cuda/parallel-thread-execution/index.html#asynchronous-warpgroup-level-matrix-instructions-wgmma-wait-group)
   }];
   let extraClassDefinition = [{
     std::string $cppClass::getPtx() { return std::string("wgmma.wait_group.sync.aligned %0;"); }
@@ -1603,8 +1608,8 @@
     |--------------|--------------|------------|--------------|---------------|
     ```
 
-    See for more information:
-    https://docs.nvidia.com/cuda/parallel-thread-execution/index.html#asynchronous-warpgroup-level-matrix-instructions
+    
+    [For more information, see PTX ISA](https://docs.nvidia.com/cuda/parallel-thread-execution/index.html#asynchronous-warpgroup-level-matrix-instructions)
   }];
   
   let hasVerifier = 1;