[GitHub] [incubator-tvm] jcf94 commented on a change in pull request #6512: [ANSOR] Auto-scheduler tutorial for GPU and necessary refactor/fix

[GitBox <gi...@apache.org>]

jcf94 commented on a change in pull request #6512:
URL: https://github.com/apache/incubator-tvm/pull/6512#discussion_r490718547



##########
File path: python/tvm/auto_scheduler/measure.py
##########
@@ -25,8 +25,8 @@
 A builder builds the executable binary files and a runner runs the binary files to
 get the measurement results. The flow of data structures is
 
-                `ProgramBuilder`                 `ProgramRunner`
-`MeasureInput` -----------------> `BuildResult` ----------------> `MeasureResult`
+  .                `ProgramBuilder`                 `ProgramRunner`

Review comment:
       The "." is for?

##########
File path: tutorials/auto_scheduler/tune_matmul_x86.py
##########
@@ -155,19 +169,22 @@ def resume_search(task, log_file):
     sch, args = auto_scheduler.auto_schedule(task, search_policy, tuning_options=tune_option)
 
 
-# resume_search(task, "matmul.json")
+#resume_search(task, "matmul.json")
 
 ######################################################################
 # .. note::
 #   We cannot run the line above because of the conflict between
 #   python's multiprocessing and tvm's thread pool.
-#   After running a tvm generated binary (L112), the python's multiprocessing
-#   library will hang forever.
-#   You have to make sure that you don't run any tvm generated binaries before
-#   calling ansor's search. To run the L156 above, you should comment out L112-114.
+#   After running a tvm generated binary the python's multiprocessing library 
+#   will hang forever. You have to make sure that you don't run any tvm 
+#   generated binaries before calling auot-scheduler's search.
+#   To run the function above, you should comment out all code in 
+#   "Check correctness and evaluate performance" section.
 #
 #   You should be careful about this problem in your applications.
 #   There are other workarounds for this problem.
 #   For example, you can start a new thread/process (with the builtin python library
 #   threading or multiprocessing) and run the tvm binaries in the new thread/process.
 #   This provides an isolation and avoids the conflict in the main thread/process.
+#   You can also use :any:`auto_scheduler.measure.LocalRPCMeasureContext` for auto-scheduler,
+#   as shown in the GPU tutorial (:ref:`auto-scheduler-conv-gpu`).

Review comment:
       This may still be confusing. I'm thinking it may be better to just delete the resume part of the CPU tutorials to keep it as simple as possible.
   
   We can add a comment saying that `LocalRPCMeasureContext` is recommended to be used in all hardware targets in the GPU tutorial.




----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org