Jump to content

HQUEUE clients stalling


etudenc

Recommended Posts

Hi all,

 

I'm trying to sim an exhaust trail via a distributed cluster simulation on my small farm thru HQUEUE. The server and clients work fine... clients are assigned their respective clusters, etc. 

 

The problem is that once a client machine is finished with the "heavy lifting" it seems to stall and stop calculating its cluster. 

 

To put it another way, for a frame range of 1-100, perhaps the first 60 frames get simulated (represented by 2KB file sizes leading up to the heavier file sizes (which indicates the emitter is traveling thru the fluid box) then back down to 2KB file sizes once the density has dissipated. But then the client stops simulating and I have to cancel its job to move it on to the next. Of that 100 frames it was supposed to render, perhaps it only gets through 66 frames before stalling.

 

In the DOP network a gas resize fluid DOP is fed into the second input of the pyro solver (as normal). Does anyone know off-hand if that's a "no-no" when used in a clustering sim?

 

To get around the stalling problem I was hoping I could maybe code something in Python to run in the background and check clients' status and if they're still simming after so many minutes it would cancel their current job, but the Python API doesn't appear to include commands for Hqueue. Anyone know if this is possible?

 

Thanks for any insight you can share!

Link to comment
Share on other sites

Is this with 12.5?  There was a bug fixed in 12.5.557 that could cause problems with clustered smoke sims with dissipation:

 

 

Friday, October 18, 2013 
Houdini 12.5.557: Fixed error in Multigrid pressure solve that lead to invalid values in velocity field when dealing with 1-dimensional fields. This situation most often occurred when using clustering along with smoke dissipation.

 

 

 

 

 

Is your version older than that?

Link to comment
Share on other sites

I have found similar stalling behavior with the FLIP solver in a setup where you supply a proxy collision field in a static object. It seems to stall when you feed it an empty SDF. In particular its the pressure solve that keeps on going for ever. This occurs in 13.0.237, the previous production build.

Link to comment
Share on other sites

@johner, thanks again for the suggestion, that fixed the problem! 

 

Great! Note that I'm pretty sure this bug still exists for the OpenCL path, so in case you were considering trying it, I'd avoid it just for this type of sim.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...