Hi Martin @spikeyxxx,
I am not looking at absolute performance here (but the ratio difference in performance). When use portal as a tool, I wish to understand more of the performance with and without portal added. As I studied the use of portal, I use little tests like this to gain actual experiences.
Having read your comment, I disabled the Volumetric Cube and did another test. I rendered with and without portal alternatively three times and took average of the results. The data surprised me! On all three occassions the without portal is actually faster! The averages are: with portal 170.2secs and without portal 166.4sec. The without portal is faster by 3.8sec(-2.3%)!
However, with portal and same number of samples, I noticed that the quality of the image is better, noise is lesser. Therefore, there is no simple answer to use portal or not! In the beginning, I have a naive thought that the portal will help speed up render time in noticeable magnitude. It is not the case now that I find it affect the quality too.
