Project

General

Profile

Feature #1436

trace CcdCallFnAfter() causality

Added by Jim Phillips over 2 years ago. Updated 4 months ago.

Status:
In Progress
Priority:
Normal
Assignee:
Category:
Tracing
Target version:
Start date:
02/17/2017
Due date:
% Done:

0%


Description

NAMD uses CcdCallFnAfter() to poll for GPU completion, which breaks the causality chain. It should be possible to trace a chain of CcdCallFnAfter() calls back to the entry method that called the first one.

History

#1 Updated by Eric Bohm almost 2 years ago

  • Target version changed from 6.8.1 to 6.9.0

#2 Updated by Phil Miller almost 2 years ago

Ronak or Karthik, please get whatever further details are necessary, and decide if this should be addressed in the next release timeframe.

#3 Updated by Eric Bohm over 1 year ago

  • Assignee set to Karthik Senthil

#4 Updated by Karthik Senthil over 1 year ago

  • Target version changed from 6.9.0 to 6.9.1

#5 Updated by Sam White 6 months ago

  • Target version changed from 6.9.1 to 6.10.0
  • Assignee changed from Karthik Senthil to Ronak Buch

#6 Updated by Ronak Buch 4 months ago

  • Status changed from New to In Progress

I think this is doable, but the detail of functions coming from a CcdCallFnAfter may be a bit suspect (since it just calls a function directly, so it's hard for the RTS to know what's executing) and the ordering causality chain may include extraneous calls due to differing levels of delays for different invocations of CcdCallFnAfter. Regardless, getting causation to the right place may be a bit tricky without adding significantly more disambiguation code to adding stuff to the CcdCallFnAfter call. I've started toying with an implementation to see if I can get it to work correctly or not.

Also available in: Atom PDF