OTOY Forums

Hi,
Recently i wanted to catch up with octane new releases and I've install 3.08 - 15.2.3, last version I was using was 3.05.3 so it was a bit old
but basically what i saw was that the new version is actually slower for me ... 3.05.3 rendered a scene at 15min 10sec and 3.08 renders at same settings in 16min 44sec ...
not using adaptive sampling function, just max samples set to 1000, is it normal or am I doing something wrong ?
thanks in advance for any suggestions

Increasing Parallel samples to max value cam speed up rendering a bit at the cost of used memory.

Thanks for the reply J. C., for comparision I was using the same value, that is 16, in both renders

This is a tough one to answer because of the number of variables in it. In general with GPU coding, the more you add, the slower it runs. Every additional feature can slow down the render in many different ways. The reason is that GPUs hate branching, they need you to bundle as much as you can together so each individual processor accesses memory at the same time, etc. The more complex the program becomes the harder it is to do this and everything slows down.

Now this is just my 2 cents, I have no idea if the dev team will be able to optimize more or refactor the code, or it's a bug, etc., etc.

Have you compared standalone versions like 3.07 vs 3.08? Is it the same speed difference?

thanks for sugestions, grimm and J.C, grimm what You say makes sense, maby it is the way that it suposed to be, J.C i wasn't comparing standalone times just blender plugin, i dont really use standalone but I'll try to play with it on weekend