Icemist opened a new pull request, #13349:
URL: https://github.com/apache/tvm/pull/13349

   
   The main items that have been added are:
   
   tvm.tir.analysis.calculate_allocated_bytes() 
   > to calculate allocated memory per memory scope
   
   tir.transform.VerifyVTCMLimit(limit)
   
   > to verify if the size of the allocated vtcm memory satisfies the limit
   
   tvm.target.hexagon().vtcm_capacity 
   
   > attribute to pass the limit
   
   tir.vtcm_capacity 
   
   > context configuration attribute to pass the limit alternatively
   
   Interaction with tuners:
   - autotvm - In the case of the rejection of all trials falls within the 
implementation
   - auto_scheduler - сan fall into an infinite loop when all variants are 
rejected, requires a limit on the number of errors in the search phase, the 
same as in the measurement phase
   - meta_scheduler - need to add a test


-- 
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