MasterJH5574 commented on code in PR #11088:
URL: https://github.com/apache/tvm/pull/11088#discussion_r856845596


##########
src/meta_schedule/postproc/rewrite_tensorize.cc:
##########
@@ -0,0 +1,104 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *   http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ * KIND, either express or implied.  See the License for the
+ * specific language governing permissions and limitations
+ * under the License.
+ */
+#include <tvm/runtime/container/base.h>
+
+#include <algorithm>
+
+#include "../utils.h"
+
+namespace tvm {
+namespace meta_schedule {
+
+using tir::BlockRV;
+using tir::LoopRV;
+
+void ApplyTensorization(const tir::Schedule& sch, const String& func_name,
+                        const tir::PrimFuncNode* func, bool 
vectorize_init_loop) {
+  std::vector<std::pair<std::string, std::function<void(tir::BlockRV)>>> jobs;
+
+  tir::PostOrderVisit(func->body, [=, &jobs](const ObjectRef& obj) {
+    if (const auto* block = obj.as<tir::BlockNode>()) {
+      tir::StmtSRef block_sref = sch->GetSRef(block);
+      if (Optional<String> intrin_name =
+              tir::GetAnn<String>(block_sref, 
tir::attr::meta_schedule_auto_tensorize)) {
+        std::string block_name = 
block_sref->StmtAs<tir::BlockNode>()->name_hint;
+        if (block_name.find("init") == std::string::npos) {
+          jobs.emplace_back(block_name, [sch, intrin_name](tir::BlockRV block) 
{
+            try {
+              sch->Tensorize(block, intrin_name.value());
+            } catch (const std::exception& e) {
+              LOG(WARNING) << "Tensorize failed with error " << e.what();
+            }
+          });
+        } else if (vectorize_init_loop) {
+          jobs.emplace_back(block_name, [sch](tir::BlockRV block) {
+            Array<BlockRV> child_blocks = sch->GetChildBlocks(block);
+            ICHECK(child_blocks.size() == 1);
+            Array<LoopRV> init_loops = sch->GetLoops(child_blocks[0]);
+            ICHECK(init_loops.size() == 1);
+            sch->Vectorize(init_loops[0]);

Review Comment:
   Quite interesting.. So here the case is, on one hand we don’t want the block 
being annotated by rule `ParallelVectorizeUnroll`, but on the other hand we do 
want its init block to be vectorized after the decomposition. Am I right?
   
   Since before decomposition the block wasn’t annotated by 
`ParallelVectorizeUnroll`, the decomposed init block isn’t vectorized, which 
makes sense. In addition, the decomposed init block doesn’t have any 
information to indicate that it’s supposed to vectorized (e.g., it doesn’t have 
an “need vectorization” annotation). In this case, no matter we vectorize the 
init block loop in `RewriteReductionBlock` or `RewriteTensorize`, it’s all due 
to our human knowledge, which I don’t think is perfect.
   
   For upstreaming, it might be okay to do manual vectorization in 
`RewriteTensorize` (how does the vectorization in `RewriteTensorize` bypass the 
compact dataflow issue BTW?). But in the long term I suppose we should enhance 
the compact dataflow check to allow such vectorization. After all, such 
vectorization won’t incur any incorrectness.
   
   cc @junrushao1994 @spectrometerHBH 



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to