rename sharded_layout => blocked_layout

lib/Dialect/TritonGPU/Transforms/TritonGPUConversion.cpp

@@ -46,7 +46,7 @@ TritonGPUTypeConverter::TritonGPUTypeConverter(MLIRContext *context,
       remainingThreads /= blockTileSize[dim];
       // TODO: will we need repetition?
     }
-    Attribute encoding = triton::gpu::TritonGPUShardedEncodingAttr::get(
+    Attribute encoding = triton::gpu::TritonGPUBlockedEncodingAttr::get(
         context, threadTileSize, warpTileSize, blockTileSize, order);
     return RankedTensorType::get(shape, elementType, encoding);
   });

lib/Dialect/TritonGPU/Transforms/Verifier.cpp

@@ -50,7 +50,7 @@ private:
           if (!encoding)
             return dotOp.emitError() << name << " should have encoding";
           if (!encoding.isa<triton::gpu::TritonGPUMmaEncodingAttr>() && 
-              !encoding.isa<triton::gpu::TritonGPUShardedEncodingAttr>())
+              !encoding.isa<triton::gpu::TritonGPUBlockedEncodingAttr>())
             return dotOp.emitError() << name << " should be of distributed layout";
           if (name == 'c')
             cLayout = encoding;