You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@tvm.apache.org by GitBox <gi...@apache.org> on 2021/10/07 16:45:16 UTC

[GitHub] [tvm] mikepapadim commented on pull request #9103: [Core][Build] Move build module transformations and utilities to C++

mikepapadim commented on pull request #9103:
URL: https://github.com/apache/tvm/pull/9103#issuecomment-937972179


   > Thanks @mikepapadim . One thing that would be very helpful is to split the logic into two parts:
   > 
   >     * A function that configures and returns the composite Pass(aka the sequential)
   > 
   >     * A function that uses the above function to construct the composite pass and execute the optimization.
   > 
   > 
   > As part of python first principle, we want to enable ways to access and tweak optimizations in python easily. While majority of the default logic can sit in c++. It would be super helpful to have a python function that returns the composite pipeline, and possibly allow us to do further customizations (in python) if necessary and invoke them.
   
   Thanks @tqchen. So, we want to expose on the python side the ability to register additional passes for `mixed_mod`, `device_mod` and `host_mod` per pass manager when we are doing the split?


-- 
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: commits-unsubscribe@tvm.apache.org

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