Jump to content

VRay Distributed rendering


ahbtsang
 Share

Recommended Posts

Hi VRay pros, i have some difficulties in distributed rendering, in fact something to do with speed.

 

currently i have a pentium D945 to render a scene, it took 11mins7sec

my 2nd machine is P4HT @3.9GHz (overclocked) which takes 14mins to render the same scene.

 

i wanted find out how much faster the distributed rendering will be with both machines. Surprisingly, the time taken to render this scene with DR took 13mins!!!! To all pros here, is there any problems with my setting? or do i need to do something or setup something first before using DR?? both computers are connected on a LAN, not wireless. Also i would like to know will GigaLAN affect the DR speed??

 

this is a fully textured scene with lightings, and my vray version is 1.47. i greatly appreciate if any of you can offer help.

Link to comment
Share on other sites

Yes, the speed you're getting with DR is way off.

 

First of all, I have no idea if GigaLAN would affect your speed, although I'm sure it's entirely possible. If there is any software involved, perhaps you could try deleting it temporarily.

 

There's a lot of answers to know before anyone can provide an answer with any certainity. How many buckets are rendering while using one machine without DR, how many using one machine with DR, and how many using both machines with DR? Is the 2nd computers IP showing up in the buckets? How long to render using 1st machine without DR and how long to render using 1st machine only with DR?

 

All I can say is use deductive reasoning and take certain things out of the equation. For example, you say it's a fully textured scene. Create a new scene with no materials applied, say 100 teapots, and try it again. There's got to be a reason. I'm not a networking guru by any means, but I'm sure the reason many very well have more to do with the network than with VRay/Max. Maybe try disabling the LAN and reenabling. Maybe switch from a static IP to dynamic or vice versa. Try disabling GI to see if it has any bearing.

 

Perhaps if all else fails, you could try installing 1.5 and hope that the issue doesn't persist. Good luck.

Link to comment
Share on other sites

How big is the file? I've rendered things quicker on my dual Opteron 248 machine than on my farm of 4 - dual 3.2 xeons with DR activated. Sometimes it's just not worth sending over the network and overall network speed will affect DR's effectiveness.

 

I'm not a V-ray user but with FinalRender's DR feature there seems to be a happy balance with bucket size ~ in which all of the buckets are contributing, maximally, to the scene... that is, depending upon the output size of the rendering, you shouldn't have your bucket size more than what is ideal given scene, textures, poly's, etc. Let's say you have 4 cores from your P4D and your P4+HT, the bucket size should not be set to any smaller than what the slowest machine can render without being noticeably far behind the faster machines and therefore holding up the rendering. Example: I have a dual Opteron, 4 dual xeons, and a P4 3.2ghz laptop, all of which are respectively fast but the 3.2ghz laptop just can't keep up with the rest of the machines so I make sure that the bucket "turnover" rate is sufficient so that the slowest machine is compensated for by the faster machines in the DR chain. If not, the slowest machine "hangs" on the final bucket in the rendering and totally offsets the value of DR.

 

Do you have enough RAM installed on the nodes? Nothing will cripple a pipeline like a lack of ram!! Watch Task Manager when rendering using DR and that will tell you all you whether or not your node is being forced to page the rendering.

 

hope this helps : )

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...