I can’t find it in the docs or official changelog. Just wondering if it is ‘safe’ to always use the number of passes at 0?. It does seem to work in that it always uses the max number of cores that the machine you are using has.
This option was added in V-Ray 2.0. A note about it can be found in the vray_changelog file:
(*) V-Ray now uses auto number of light cache passes when it is specified as zero;
All V-Ray versions later than 2.0 will use the max number of cores when the LC passes value is set to zero.
Furthermore the value is not set as default for compatibility reasons.
It is not the default setting because it will cause problems if you render animations on a render farm which has machines with varying number of cores. If the different machines on the farm use different number of light cache passes, they will end up with slightly different light caches, which may lead to flickering.
The highest of the lowest (cpus on the farm) I think he means? If you are only calculating LC with one machine (like a fly-through) animation then 0 is fine, but it only counts in a moving geometry animation where you are calculating LC across multiple machines.
If the slaves have the same number of cores, then yes. Otherwise there might be differences between the strips from different machines. You can just set it to the largest core number; V-Ray 3.0 simply fixes it at 64 passes always.
my workstation is Xeon dual core (8 core 16 thread), and i have 4 slaves, each slave have 12 core 16thread, total is 96 thread, so how much should i set?
thank you
before i get your answer i have set it to 96 according to my slave thread, and send it to backburner for over night render, i hope it won’t be any problem to the result.
It’s not there. The light cache will still be computed using 64 threads. The performance is not that much worse since the missing 8 cores are hyper-threaded anyways.