randomkang commented on PR #3145:
URL: https://github.com/apache/brpc/pull/3145#issuecomment-3664518090

   > > In fact, I still don't understand the specific function of the 5/4 
window in the latest Premiere Pro.
   > 
   > I think it would be more reasonable to add a sliding window to limit the 
number of IMM.
   > 
   > @randomkang Please try the latest PR again
   
   I run 2 tasks, and the error of "Fail to ibv_post_send: Cannot allocate 
memory" does not happpen again. 
   
   One last for 1790 mins, and fail due to "[wk-15] Thread E1217 
16:32:30.900705 106575 12820477405150 
external/brpc/src/brpc/rdma/rdma_helper.cpp:220 BlockAllocate] Fail to get 
block from memory pool".
   
   The other last for 1787mins, and fail due to "[tf-ps-7] W1217 
04:18:39.676858 24875 78292958891373 
tensorflow/contrib/pms/rpc/brpc_remote_worker.cc:383 operator()] Fail to recv 
tensorflow.RecvTensorResponse from 10.28.132.10:3642 : Aborted: 
[E10010][10.28.132.10:3642][E10010]Step 60851325530873895". I don't see the 
error message of brpc, maybe this error is caused by tensorflow.


-- 
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]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to