Does this work? Our work stations are 32x Max9 with 1.5(32x), we have a test machine that is 64x Max9 with 1.5(64x), if we try to send the file through Backburner, we get the following error immediatly:
No plugin available to handle job.
Is it not possible to send a 32x VRay file to a 64x VRay render machine?
I don’t want to start anything here about it, but I just heard from our IT guy that the floating points are different between the 32x and the 64x Max formats. Which, in renderings, would produce different calculations and return images of different color values, object placements, etc. So from his standpoint they did it to ‘protect’ renderings sent to mixed renderfarms. Which, if true, makes sense. :roll:
yeah but say if your workstation is still 32bit and your entire farm is 64bit then all of the images should match. I know they may look slightly different to the workstation but if you only use the farm then it shouldnt be an issue?
Instead of just blocking it, there should at least be an option to use ALL 64bit machines or use ALL 32bit machines IMHO.
Im sure they thought of that though? well id hope so!
You are correct. I’ve made some test and put them in internet. See following: http://hapate.com/vray/ (when page opens the visible image is 64x. When you put mouse on it you’ll see 32bit).
Also, as for offtopic, I’ve tried some overclocking which resulted few funny things: http://hapate.com/vray\_oc/ (put mouse over image and you’ll see what happened when I overclocked).
WOW! Those are quite drastic changes, especially in the overclocke one. With the 32x/64x one the 64x has the nicer render times, but I like the feel of the 32x on, looks more accurate.
Due to fact that renders above were done on completely different processors, I’ve made new renders. These are both rendered with Intel’s QX6700 processor, 4GB memory and winxp64. Vray version is 1.5RC3.
32Bit version (in this case) seems to be slightly faster.
Renni, any chance of you doing some tests with say qmc + lc , qmc , lc + lc, and some with hi setting Irr/lc ? Perhaps some with different noise/adaptive amounts in the qmc solver.