Think I may have just run across a bug, either in my code, or  in the API  regarding  NParticles.  

I'm working on  updating the  code for  partio4Maya, and something seems to have changed in 2015 or 2016  and  for the life of me, I now cannot  force an update to  Nparticles  any other way except  running it  per frame via  mel code frame stepping forward, and exporting a frame.  Slightly complicating things is that this seems to only be a problem with viewport 2.0   or  if the particles are pre-nucleus cached. 

I end up getting all the particles, the  counts are  correct.. but  they end up all at the origin.

to be clear..  (writing out  cache  using the api  via   my partio plugin) 

live playback  + vp1.0   =  success!
n-cache  +  vp1.0    =  failure (particles at the origin)
live playback + vp2.0 = failure (particles at the origin)
n-cache + vp2.0  =  failure(particles at the origin)

Legacy particles  are not affected by this problem at all.

Previously  I was using the   MFnParticleSystem  function   "evaluateDynamics"   and it was working
with both  particles and NParticles,  now  for whatever reason I can't use it anymore  becuase in cases where NParticles  use  lifespan of any kind,  it simply forgets to kill the particles with lifespan<age.    If I  don't use that evaluateDynamics  call,  in viewport  1.0,  it works correctly now.  but  in viewport 2.0   it fails to dirty something and  all the particle positions  end up at the origin... 

Not sure where to point my finger at this point..  it seems like  a bug  to me, but does anyone else know of any  other functions or  ways to force an update to  Nucleus  ?   I need to look at the NCache command  to see if its doing  a frame step in python/mel or not.. thats my only  other idea.

any ideas  would be welcome on this  

Quote 0 0
Things changed for sure in the particles area recently.
I didn't hit problems like yours but encountered different issues with initial states and stuff.
The old particles are somehow less affected but nParticles are more of a headache.
Maya Devs never really finished their work on them. :S

I don't know what to say really. Maybe reach to Julia or Duncan at Autodesk.
Quote 0 0
Could this be a parallel/ DG thing?

We've had loads of issues with nParticles that were 'solved' by switching to DG evaluation mode. 
Of course, it could be something completely different...
Quote 0 0